diff --git a/.idea/deployment.xml b/.idea/deployment.xml
index e111d3e..7378699 100644
--- a/.idea/deployment.xml
+++ b/.idea/deployment.xml
@@ -16,6 +16,34 @@
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
\ No newline at end of file
diff --git a/.idea/workspace.xml b/.idea/workspace.xml
index 403489e..2b21939 100644
--- a/.idea/workspace.xml
+++ b/.idea/workspace.xml
@@ -5,9 +5,38 @@
-
-
-
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
@@ -39,34 +68,43 @@
+
+
-
-
+
+
@@ -75,12 +113,12 @@
-
+
-
+
@@ -89,7 +127,30 @@
-
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
@@ -103,7 +164,30 @@
-
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
@@ -227,7 +311,30 @@
-
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
@@ -241,7 +348,30 @@
-
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
@@ -252,9 +382,11 @@
-
@@ -267,6 +399,9 @@
+
+ C:\Users\10561\AppData\Roaming\Subversion
+
@@ -281,7 +416,23 @@
-
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
@@ -307,7 +458,23 @@
1736125962873
-
+
+
+ 1736126943869
+
+
+
+ 1736126943869
+
+
+
+ 1736126953818
+
+
+
+ 1736126953818
+
+
@@ -329,8 +496,12 @@
+
+
+
-
+
+
\ No newline at end of file
diff --git a/002_B_Car/K210采取图片.py b/002_B_Car/K210采取图片.py
index 2d3e306..ff2607a 100644
--- a/002_B_Car/K210采取图片.py
+++ b/002_B_Car/K210采取图片.py
@@ -1,24 +1,4 @@
-'''
- image caupture tool, capture images to SD card with directory switch support
- @usage 1. Change camera and lcd configs according to your hardware like `lcd.rotation`
- 2. Prepare a SD card with SPI mode support, and format to FAT32 with MBR(msdos) partition
- 3. Copy this boot.py to your SD card root directory
- 4. Power off board, then insert SD card to board
- 5. Power on, it will automatically create a directory like `cap_images_1`,
- the next time will be `cap_images_2` to avoid rewrite
- 6. Capture images for one class,
- push `boot` button on board and release to capture one image,
- this will save images as `cap_images_1/0/0.jpg`,
- the name of image will automaitcally increase, like `0.jpg` `1.jpg` ...
- 7. Long push `boot` botton on board to switch class,
- this will create a new directory like `cap_images_1/1/`,
- and later captured images will be saved to this directory like `cap_images_1/1/0.jpg`
- 8. Power off board, pop out SD card, mount on your PC, now you get your images in your file brower
- @update https://maixhub.com/app/3
- @version v1.0.0
- @author neucrack@sipeed
- @license MIT
-'''
+
import sensor, lcd
from Maix import GPIO
diff --git a/002_B_Car/image_with_red_rectangle.jpg b/002_B_Car/image_with_red_rectangle.jpg
index ae4cceb..10ea7eb 100644
Binary files a/002_B_Car/image_with_red_rectangle.jpg and b/002_B_Car/image_with_red_rectangle.jpg differ
diff --git a/002_B_Car/model-178789.nncase/main.py b/002_B_Car/model-178789.nncase/main.py
new file mode 100644
index 0000000..95fc1fc
--- /dev/null
+++ b/002_B_Car/model-178789.nncase/main.py
@@ -0,0 +1,112 @@
+# generated by maixhub, tested on maixpy3 v0.4.8
+# copy files to TF card and plug into board and power on
+import sensor, image, lcd, time
+import KPU as kpu
+from machine import UART
+import gc, sys
+from fpioa_manager import fm
+
+input_size = (224, 224)
+labels = ['light', 'yellow', 'green', 'red']
+anchors = [0.63, 0.69, 0.81, 0.91, 0.75, 2.47, 0.94, 3.78, 0.69, 2.78]
+
+def lcd_show_except(e):
+ import uio
+ err_str = uio.StringIO()
+ sys.print_exception(e, err_str)
+ err_str = err_str.getvalue()
+ img = image.Image(size=input_size)
+ img.draw_string(0, 10, err_str, scale=1, color=(0xff,0x00,0x00))
+ lcd.display(img)
+
+class Comm:
+ def __init__(self, uart):
+ self.uart = uart
+
+ def send_detect_result(self, objects, labels):
+ msg = ""
+ for obj in objects:
+ pos = obj.rect()
+ p = obj.value()
+ idx = obj.classid()
+ label = labels[idx]
+ msg += "{}:{}:{}:{}:{}:{:.2f}:{}, ".format(pos[0], pos[1], pos[2], pos[3], idx, p, label)
+ if msg:
+ msg = msg[:-2] + "\n"
+ self.uart.write(msg.encode())
+
+def init_uart():
+ fm.register(10, fm.fpioa.UART1_TX, force=True)
+ fm.register(11, fm.fpioa.UART1_RX, force=True)
+
+ uart = UART(UART.UART1, 115200, 8, 0, 0, timeout=1000, read_buf_len=256)
+ return uart
+
+def main(anchors, labels = None, model_addr="/sd/m.kmodel", sensor_window=input_size, lcd_rotation=0, sensor_hmirror=False, sensor_vflip=False):
+ sensor.reset()
+ sensor.set_pixformat(sensor.RGB565)
+ sensor.set_framesize(sensor.QVGA)
+ sensor.set_windowing(sensor_window)
+ sensor.set_hmirror(sensor_hmirror)
+ sensor.set_vflip(sensor_vflip)
+ sensor.run(1)
+
+ lcd.init(type=1)
+ lcd.rotation(lcd_rotation)
+ lcd.clear(lcd.WHITE)
+
+ if not labels:
+ with open('labels.txt','r') as f:
+ exec(f.read())
+ if not labels:
+ print("no labels.txt")
+ img = image.Image(size=(320, 240))
+ img.draw_string(90, 110, "no labels.txt", color=(255, 0, 0), scale=2)
+ lcd.display(img)
+ return 1
+ try:
+ img = image.Image("startup.jpg")
+ lcd.display(img)
+ except Exception:
+ img = image.Image(size=(320, 240))
+ img.draw_string(90, 110, "loading model...", color=(255, 255, 255), scale=2)
+ lcd.display(img)
+
+ uart = init_uart()
+ comm = Comm(uart)
+
+ try:
+ task = None
+ task = kpu.load(model_addr)
+ kpu.init_yolo2(task, 0.7, 0.3, 5, anchors) # threshold:[0,1], nms_value: [0, 1]
+ while(True):
+ img = sensor.snapshot()
+ t = time.ticks_ms()
+ objects = kpu.run_yolo2(task, img)
+ t = time.ticks_ms() - t
+ if objects:
+ for obj in objects:
+ pos = obj.rect()
+ img.draw_rectangle(pos)
+ img.draw_string(pos[0], pos[1], "%s : %.2f" %(labels[obj.classid()], obj.value()), scale=2, color=(255, 0, 0))
+ comm.send_detect_result(objects, labels)
+ img.draw_string(0, 200, "t:%dms" %(t), scale=2, color=(255, 0, 0))
+ img.draw_string(0, 2, "Upgrade to MaixCAM to use YOLOv8", scale=1.2, color=(255, 0, 0))
+ img.draw_string(0, 30, "wiki.sipeed.com/maixcam", scale=1.2, color=(255, 0, 0))
+ lcd.display(img)
+ except Exception as e:
+ raise e
+ finally:
+ if not task is None:
+ kpu.deinit(task)
+
+
+if __name__ == "__main__":
+ try:
+ # main(anchors = anchors, labels=labels, model_addr=0x300000, lcd_rotation=0)
+ main(anchors = anchors, labels=labels, model_addr="/sd/model-178789.kmodel")
+ except Exception as e:
+ sys.print_exception(e)
+ lcd_show_except(e)
+ finally:
+ gc.collect()
diff --git a/002_B_Car/model-178789.nncase/model-178789.kmodel b/002_B_Car/model-178789.nncase/model-178789.kmodel
new file mode 100644
index 0000000..d39e967
Binary files /dev/null and b/002_B_Car/model-178789.nncase/model-178789.kmodel differ
diff --git a/002_B_Car/model-178789.nncase/report.json b/002_B_Car/model-178789.nncase/report.json
new file mode 100644
index 0000000..a9bdfc7
--- /dev/null
+++ b/002_B_Car/model-178789.nncase/report.json
@@ -0,0 +1,1748 @@
+{
+ "anchors": [
+ 0.63,
+ 0.69,
+ 0.81,
+ 0.91,
+ 0.75,
+ 2.47,
+ 0.94,
+ 3.78,
+ 0.69,
+ 2.78
+ ],
+ "inputs": {
+ "input0": [
+ 224,
+ 224,
+ 3
+ ]
+ },
+ "outputs": {
+ "output0": [
+ 7,
+ 7,
+ 45
+ ]
+ },
+ "params_quantity": 1867581,
+ "loss": [
+ [
+ 1,
+ 0.23358
+ ],
+ [
+ 2,
+ 0.08057
+ ],
+ [
+ 3,
+ 0.05845
+ ],
+ [
+ 4,
+ 0.04289
+ ],
+ [
+ 5,
+ 0.05269
+ ],
+ [
+ 6,
+ 0.19217
+ ],
+ [
+ 7,
+ 0.05192
+ ],
+ [
+ 8,
+ 0.02652
+ ],
+ [
+ 9,
+ 0.02197
+ ],
+ [
+ 10,
+ 0.0186
+ ],
+ [
+ 11,
+ 0.016
+ ],
+ [
+ 12,
+ 0.01364
+ ],
+ [
+ 13,
+ 0.01297
+ ],
+ [
+ 14,
+ 0.01098
+ ],
+ [
+ 15,
+ 0.01035
+ ],
+ [
+ 16,
+ 0.01036
+ ],
+ [
+ 17,
+ 0.01444
+ ],
+ [
+ 18,
+ 0.01098
+ ],
+ [
+ 19,
+ 0.01022
+ ],
+ [
+ 20,
+ 0.00875
+ ],
+ [
+ 21,
+ 0.01012
+ ],
+ [
+ 22,
+ 0.00778
+ ],
+ [
+ 23,
+ 0.00662
+ ],
+ [
+ 24,
+ 0.00659
+ ],
+ [
+ 25,
+ 0.00623
+ ],
+ [
+ 26,
+ 0.00555
+ ],
+ [
+ 27,
+ 0.00477
+ ],
+ [
+ 28,
+ 0.00466
+ ],
+ [
+ 29,
+ 0.00431
+ ],
+ [
+ 30,
+ 0.00406
+ ],
+ [
+ 31,
+ 0.00409
+ ],
+ [
+ 32,
+ 0.00386
+ ],
+ [
+ 33,
+ 0.00412
+ ],
+ [
+ 34,
+ 0.00399
+ ],
+ [
+ 35,
+ 0.00359
+ ],
+ [
+ 36,
+ 0.00361
+ ],
+ [
+ 37,
+ 0.00351
+ ],
+ [
+ 38,
+ 0.00352
+ ],
+ [
+ 39,
+ 0.00351
+ ],
+ [
+ 40,
+ 0.00339
+ ],
+ [
+ 41,
+ 0.00341
+ ],
+ [
+ 42,
+ 0.00333
+ ],
+ [
+ 43,
+ 0.00338
+ ],
+ [
+ 44,
+ 0.00326
+ ],
+ [
+ 45,
+ 0.00315
+ ],
+ [
+ 46,
+ 0.00333
+ ],
+ [
+ 47,
+ 0.00334
+ ],
+ [
+ 48,
+ 0.00322
+ ],
+ [
+ 49,
+ 0.0035
+ ],
+ [
+ 50,
+ 0.00339
+ ],
+ [
+ 51,
+ 0.00325
+ ],
+ [
+ 52,
+ 0.00301
+ ],
+ [
+ 53,
+ 0.00308
+ ],
+ [
+ 54,
+ 0.00299
+ ],
+ [
+ 55,
+ 0.00296
+ ],
+ [
+ 56,
+ 0.00296
+ ],
+ [
+ 57,
+ 0.00292
+ ],
+ [
+ 58,
+ 0.00298
+ ],
+ [
+ 59,
+ 0.00294
+ ],
+ [
+ 60,
+ 0.00294
+ ],
+ [
+ 61,
+ 0.00281
+ ],
+ [
+ 62,
+ 0.00274
+ ],
+ [
+ 63,
+ 0.00277
+ ],
+ [
+ 64,
+ 0.00276
+ ],
+ [
+ 65,
+ 0.00282
+ ],
+ [
+ 66,
+ 0.00278
+ ],
+ [
+ 67,
+ 0.00284
+ ],
+ [
+ 68,
+ 0.00279
+ ],
+ [
+ 69,
+ 0.00273
+ ],
+ [
+ 70,
+ 0.00275
+ ],
+ [
+ 71,
+ 0.0027
+ ],
+ [
+ 72,
+ 0.00273
+ ],
+ [
+ 73,
+ 0.00287
+ ],
+ [
+ 74,
+ 0.00268
+ ],
+ [
+ 75,
+ 0.00267
+ ],
+ [
+ 76,
+ 0.00263
+ ],
+ [
+ 77,
+ 0.00262
+ ],
+ [
+ 78,
+ 0.0026
+ ],
+ [
+ 79,
+ 0.00265
+ ],
+ [
+ 80,
+ 0.00262
+ ],
+ [
+ 81,
+ 0.00264
+ ],
+ [
+ 82,
+ 0.0026
+ ],
+ [
+ 83,
+ 0.00258
+ ],
+ [
+ 84,
+ 0.00258
+ ],
+ [
+ 85,
+ 0.00259
+ ],
+ [
+ 86,
+ 0.00255
+ ],
+ [
+ 87,
+ 0.00255
+ ],
+ [
+ 88,
+ 0.00258
+ ],
+ [
+ 89,
+ 0.00256
+ ],
+ [
+ 90,
+ 0.00255
+ ],
+ [
+ 91,
+ 0.00252
+ ],
+ [
+ 92,
+ 0.00253
+ ],
+ [
+ 93,
+ 0.00252
+ ],
+ [
+ 94,
+ 0.00254
+ ],
+ [
+ 95,
+ 0.00252
+ ],
+ [
+ 96,
+ 0.00252
+ ],
+ [
+ 97,
+ 0.00249
+ ],
+ [
+ 98,
+ 0.00252
+ ],
+ [
+ 99,
+ 0.00248
+ ],
+ [
+ 100,
+ 0.00246
+ ]
+ ],
+ "acc": [
+ [
+ 1,
+ 0.02405
+ ],
+ [
+ 2,
+ 0.03661
+ ],
+ [
+ 3,
+ 0.03926
+ ],
+ [
+ 4,
+ 0.04699
+ ],
+ [
+ 5,
+ 0.04808
+ ],
+ [
+ 6,
+ 0.1038
+ ],
+ [
+ 7,
+ 0.11539
+ ],
+ [
+ 8,
+ 0.11909
+ ],
+ [
+ 9,
+ 0.12244
+ ],
+ [
+ 10,
+ 0.12026
+ ],
+ [
+ 11,
+ 0.12135
+ ],
+ [
+ 12,
+ 0.12235
+ ],
+ [
+ 13,
+ 0.12335
+ ],
+ [
+ 14,
+ 0.12418
+ ],
+ [
+ 15,
+ 0.12554
+ ],
+ [
+ 16,
+ 0.12542
+ ],
+ [
+ 17,
+ 0.12408
+ ],
+ [
+ 18,
+ 0.13009
+ ],
+ [
+ 19,
+ 0.13263
+ ],
+ [
+ 20,
+ 0.13009
+ ],
+ [
+ 21,
+ 0.12906
+ ],
+ [
+ 22,
+ 0.13047
+ ],
+ [
+ 23,
+ 0.13081
+ ],
+ [
+ 24,
+ 0.13216
+ ],
+ [
+ 25,
+ 0.13344
+ ],
+ [
+ 26,
+ 0.13275
+ ],
+ [
+ 27,
+ 0.13285
+ ],
+ [
+ 28,
+ 0.13195
+ ],
+ [
+ 29,
+ 0.13088
+ ],
+ [
+ 30,
+ 0.13072
+ ],
+ [
+ 31,
+ 0.13138
+ ],
+ [
+ 32,
+ 0.13274
+ ],
+ [
+ 33,
+ 0.13175
+ ],
+ [
+ 34,
+ 0.13031
+ ],
+ [
+ 35,
+ 0.13059
+ ],
+ [
+ 36,
+ 0.13162
+ ],
+ [
+ 37,
+ 0.1315
+ ],
+ [
+ 38,
+ 0.13173
+ ],
+ [
+ 39,
+ 0.13269
+ ],
+ [
+ 40,
+ 0.1329
+ ],
+ [
+ 41,
+ 0.13183
+ ],
+ [
+ 42,
+ 0.13213
+ ],
+ [
+ 43,
+ 0.13152
+ ],
+ [
+ 44,
+ 0.13147
+ ],
+ [
+ 45,
+ 0.13112
+ ],
+ [
+ 46,
+ 0.13049
+ ],
+ [
+ 47,
+ 0.13143
+ ],
+ [
+ 48,
+ 0.13148
+ ],
+ [
+ 49,
+ 0.13241
+ ],
+ [
+ 50,
+ 0.13266
+ ],
+ [
+ 51,
+ 0.13297
+ ],
+ [
+ 52,
+ 0.13299
+ ],
+ [
+ 53,
+ 0.13302
+ ],
+ [
+ 54,
+ 0.1316
+ ],
+ [
+ 55,
+ 0.13264
+ ],
+ [
+ 56,
+ 0.13245
+ ],
+ [
+ 57,
+ 0.13117
+ ],
+ [
+ 58,
+ 0.13046
+ ],
+ [
+ 59,
+ 0.13088
+ ],
+ [
+ 60,
+ 0.13193
+ ],
+ [
+ 61,
+ 0.13135
+ ],
+ [
+ 62,
+ 0.13089
+ ],
+ [
+ 63,
+ 0.13007
+ ],
+ [
+ 64,
+ 0.13064
+ ],
+ [
+ 65,
+ 0.13058
+ ],
+ [
+ 66,
+ 0.1314
+ ],
+ [
+ 67,
+ 0.13059
+ ],
+ [
+ 68,
+ 0.13061
+ ],
+ [
+ 69,
+ 0.13171
+ ],
+ [
+ 70,
+ 0.13094
+ ],
+ [
+ 71,
+ 0.13069
+ ],
+ [
+ 72,
+ 0.13002
+ ],
+ [
+ 73,
+ 0.1286
+ ],
+ [
+ 74,
+ 0.12936
+ ],
+ [
+ 75,
+ 0.13031
+ ],
+ [
+ 76,
+ 0.13113
+ ],
+ [
+ 77,
+ 0.13111
+ ],
+ [
+ 78,
+ 0.13178
+ ],
+ [
+ 79,
+ 0.13127
+ ],
+ [
+ 80,
+ 0.13113
+ ],
+ [
+ 81,
+ 0.13069
+ ],
+ [
+ 82,
+ 0.13025
+ ],
+ [
+ 83,
+ 0.13079
+ ],
+ [
+ 84,
+ 0.1305
+ ],
+ [
+ 85,
+ 0.13073
+ ],
+ [
+ 86,
+ 0.1315
+ ],
+ [
+ 87,
+ 0.13092
+ ],
+ [
+ 88,
+ 0.13014
+ ],
+ [
+ 89,
+ 0.1303
+ ],
+ [
+ 90,
+ 0.13039
+ ],
+ [
+ 91,
+ 0.12953
+ ],
+ [
+ 92,
+ 0.12983
+ ],
+ [
+ 93,
+ 0.1303
+ ],
+ [
+ 94,
+ 0.13084
+ ],
+ [
+ 95,
+ 0.13011
+ ],
+ [
+ 96,
+ 0.12984
+ ],
+ [
+ 97,
+ 0.12969
+ ],
+ [
+ 98,
+ 0.12999
+ ],
+ [
+ 99,
+ 0.13057
+ ],
+ [
+ 100,
+ 0.13107
+ ]
+ ],
+ "lr": [
+ [
+ 1,
+ 0.001
+ ],
+ [
+ 2,
+ 0.00098
+ ],
+ [
+ 3,
+ 0.00095
+ ],
+ [
+ 4,
+ 0.00093
+ ],
+ [
+ 5,
+ 0.00091
+ ],
+ [
+ 6,
+ 0.00089
+ ],
+ [
+ 7,
+ 0.00087
+ ],
+ [
+ 8,
+ 0.00085
+ ],
+ [
+ 9,
+ 0.00083
+ ],
+ [
+ 10,
+ 0.00081
+ ],
+ [
+ 11,
+ 0.00079
+ ],
+ [
+ 12,
+ 0.00078
+ ],
+ [
+ 13,
+ 0.00076
+ ],
+ [
+ 14,
+ 0.00074
+ ],
+ [
+ 15,
+ 0.00072
+ ],
+ [
+ 16,
+ 0.00071
+ ],
+ [
+ 17,
+ 0.00069
+ ],
+ [
+ 18,
+ 0.00068
+ ],
+ [
+ 19,
+ 0.00066
+ ],
+ [
+ 20,
+ 0.00065
+ ],
+ [
+ 21,
+ 0.00063
+ ],
+ [
+ 22,
+ 0.00062
+ ],
+ [
+ 23,
+ 0.0006
+ ],
+ [
+ 24,
+ 0.00059
+ ],
+ [
+ 25,
+ 0.00058
+ ],
+ [
+ 26,
+ 0.00056
+ ],
+ [
+ 27,
+ 0.00055
+ ],
+ [
+ 28,
+ 0.00054
+ ],
+ [
+ 29,
+ 0.00052
+ ],
+ [
+ 30,
+ 0.00051
+ ],
+ [
+ 31,
+ 0.0005
+ ],
+ [
+ 32,
+ 0.00049
+ ],
+ [
+ 33,
+ 0.00048
+ ],
+ [
+ 34,
+ 0.00047
+ ],
+ [
+ 35,
+ 0.00046
+ ],
+ [
+ 36,
+ 0.00045
+ ],
+ [
+ 37,
+ 0.00044
+ ],
+ [
+ 38,
+ 0.00043
+ ],
+ [
+ 39,
+ 0.00042
+ ],
+ [
+ 40,
+ 0.00041
+ ],
+ [
+ 41,
+ 0.0004
+ ],
+ [
+ 42,
+ 0.00039
+ ],
+ [
+ 43,
+ 0.00038
+ ],
+ [
+ 44,
+ 0.00037
+ ],
+ [
+ 45,
+ 0.00036
+ ],
+ [
+ 46,
+ 0.00035
+ ],
+ [
+ 47,
+ 0.00035
+ ],
+ [
+ 48,
+ 0.00034
+ ],
+ [
+ 49,
+ 0.00033
+ ],
+ [
+ 50,
+ 0.00032
+ ],
+ [
+ 51,
+ 0.00032
+ ],
+ [
+ 52,
+ 0.00031
+ ],
+ [
+ 53,
+ 0.0003
+ ],
+ [
+ 54,
+ 0.0003
+ ],
+ [
+ 55,
+ 0.00029
+ ],
+ [
+ 56,
+ 0.00028
+ ],
+ [
+ 57,
+ 0.00028
+ ],
+ [
+ 58,
+ 0.00027
+ ],
+ [
+ 59,
+ 0.00026
+ ],
+ [
+ 60,
+ 0.00026
+ ],
+ [
+ 61,
+ 0.00025
+ ],
+ [
+ 62,
+ 0.00025
+ ],
+ [
+ 63,
+ 0.00024
+ ],
+ [
+ 64,
+ 0.00023
+ ],
+ [
+ 65,
+ 0.00023
+ ],
+ [
+ 66,
+ 0.00022
+ ],
+ [
+ 67,
+ 0.00022
+ ],
+ [
+ 68,
+ 0.00021
+ ],
+ [
+ 69,
+ 0.00021
+ ],
+ [
+ 70,
+ 0.0002
+ ],
+ [
+ 71,
+ 0.0002
+ ],
+ [
+ 72,
+ 0.00019
+ ],
+ [
+ 73,
+ 0.00019
+ ],
+ [
+ 74,
+ 0.00019
+ ],
+ [
+ 75,
+ 0.00018
+ ],
+ [
+ 76,
+ 0.00018
+ ],
+ [
+ 77,
+ 0.00017
+ ],
+ [
+ 78,
+ 0.00017
+ ],
+ [
+ 79,
+ 0.00017
+ ],
+ [
+ 80,
+ 0.00016
+ ],
+ [
+ 81,
+ 0.00016
+ ],
+ [
+ 82,
+ 0.00015
+ ],
+ [
+ 83,
+ 0.00015
+ ],
+ [
+ 84,
+ 0.00015
+ ],
+ [
+ 85,
+ 0.00014
+ ],
+ [
+ 86,
+ 0.00014
+ ],
+ [
+ 87,
+ 0.00014
+ ],
+ [
+ 88,
+ 0.00013
+ ],
+ [
+ 89,
+ 0.00013
+ ],
+ [
+ 90,
+ 0.00013
+ ],
+ [
+ 91,
+ 0.00013
+ ],
+ [
+ 92,
+ 0.00012
+ ],
+ [
+ 93,
+ 0.00012
+ ],
+ [
+ 94,
+ 0.00012
+ ],
+ [
+ 95,
+ 0.00011
+ ],
+ [
+ 96,
+ 0.00011
+ ],
+ [
+ 97,
+ 0.00011
+ ],
+ [
+ 98,
+ 0.00011
+ ],
+ [
+ 99,
+ 0.0001
+ ],
+ [
+ 100,
+ 0.0001
+ ]
+ ],
+ "val_acc": [
+ [
+ 10,
+ 0.82362
+ ],
+ [
+ 20,
+ 0.92272
+ ],
+ [
+ 30,
+ 0.94956
+ ],
+ [
+ 40,
+ 0.97203
+ ],
+ [
+ 50,
+ 0.94956
+ ],
+ [
+ 60,
+ 0.95218
+ ],
+ [
+ 70,
+ 0.97465
+ ],
+ [
+ 80,
+ 0.97465
+ ],
+ [
+ 90,
+ 0.97465
+ ],
+ [
+ 100,
+ 0.97465
+ ]
+ ],
+ "AP_light": [
+ [
+ 10,
+ 0.89952
+ ],
+ [
+ 20,
+ 0.89965
+ ],
+ [
+ 30,
+ 0.89965
+ ],
+ [
+ 40,
+ 0.98951
+ ],
+ [
+ 50,
+ 0.89965
+ ],
+ [
+ 60,
+ 0.89965
+ ],
+ [
+ 70,
+ 0.98951
+ ],
+ [
+ 80,
+ 0.98951
+ ],
+ [
+ 90,
+ 0.98951
+ ],
+ [
+ 100,
+ 0.98951
+ ]
+ ],
+ "AP_yellow": [
+ [
+ 10,
+ 0.79167
+ ],
+ [
+ 20,
+ 0.97306
+ ],
+ [
+ 30,
+ 1.0
+ ],
+ [
+ 40,
+ 1.0
+ ],
+ [
+ 50,
+ 1.0
+ ],
+ [
+ 60,
+ 1.0
+ ],
+ [
+ 70,
+ 1.0
+ ],
+ [
+ 80,
+ 1.0
+ ],
+ [
+ 90,
+ 1.0
+ ],
+ [
+ 100,
+ 1.0
+ ]
+ ],
+ "AP_green": [
+ [
+ 10,
+ 0.69421
+ ],
+ [
+ 20,
+ 0.81818
+ ],
+ [
+ 30,
+ 0.8986
+ ],
+ [
+ 40,
+ 0.8986
+ ],
+ [
+ 50,
+ 0.8986
+ ],
+ [
+ 60,
+ 0.90909
+ ],
+ [
+ 70,
+ 0.90909
+ ],
+ [
+ 80,
+ 0.90909
+ ],
+ [
+ 90,
+ 0.90909
+ ],
+ [
+ 100,
+ 0.90909
+ ]
+ ],
+ "AP_red": [
+ [
+ 10,
+ 0.90909
+ ],
+ [
+ 20,
+ 1.0
+ ],
+ [
+ 30,
+ 1.0
+ ],
+ [
+ 40,
+ 1.0
+ ],
+ [
+ 50,
+ 1.0
+ ],
+ [
+ 60,
+ 1.0
+ ],
+ [
+ 70,
+ 1.0
+ ],
+ [
+ 80,
+ 1.0
+ ],
+ [
+ 90,
+ 1.0
+ ],
+ [
+ 100,
+ 1.0
+ ]
+ ],
+ "mean": 123.5,
+ "std": 58.395,
+ "label_type": "detection",
+ "data_type": "image",
+ "labels": [
+ "light",
+ "yellow",
+ "green",
+ "red"
+ ],
+ "best_eval": {
+ "epoch": 100,
+ "val_acc": 0.975,
+ "val_info": [
+ {
+ "name": "0d2ec2d7722e45fa931ae176cddaeaba.jpg",
+ "nickname": "congche_hld_500.jpg",
+ "pred": [
+ [
+ 119,
+ 43,
+ 146,
+ 128,
+ 0.9829508066177368,
+ 0.0
+ ],
+ [
+ 120,
+ 43,
+ 146,
+ 70,
+ 0.9661509990692139,
+ 3.0
+ ]
+ ],
+ "gt": [
+ [
+ 119,
+ 44,
+ 146,
+ 127,
+ 1.0,
+ 0.0
+ ],
+ [
+ 120,
+ 44,
+ 145,
+ 70,
+ 1.0,
+ 3.0
+ ]
+ ],
+ "ok": true
+ },
+ {
+ "name": "24af822ea7e94aa39a0f4828b694bff9.jpg",
+ "nickname": "B_105.jpg",
+ "pred": [
+ [
+ 134,
+ 16,
+ 159,
+ 106,
+ 0.9393616318702698,
+ 0.0
+ ],
+ [
+ 138,
+ 49,
+ 161,
+ 74,
+ 0.9140685796737671,
+ 1.0
+ ]
+ ],
+ "gt": [
+ [
+ 135,
+ 17,
+ 158,
+ 107,
+ 1.0,
+ 0.0
+ ],
+ [
+ 138,
+ 50,
+ 157,
+ 73,
+ 1.0,
+ 1.0
+ ]
+ ],
+ "ok": true
+ },
+ {
+ "name": "bec7d109303e4769a544e88f30d216e9.jpg",
+ "nickname": "B_71.jpg",
+ "pred": [
+ [
+ 138,
+ 40,
+ 162,
+ 139,
+ 0.9695276618003845,
+ 0.0
+ ],
+ [
+ 138,
+ 76,
+ 162,
+ 103,
+ 0.8642435669898987,
+ 1.0
+ ]
+ ],
+ "gt": [
+ [
+ 137,
+ 40,
+ 162,
+ 138,
+ 1.0,
+ 0.0
+ ],
+ [
+ 140,
+ 76,
+ 160,
+ 102,
+ 1.0,
+ 1.0
+ ]
+ ],
+ "ok": true
+ },
+ {
+ "name": "eecb93c2daba4efe89220e98b99c59d1.jpg",
+ "nickname": "B_504.jpg",
+ "pred": [
+ [
+ 132,
+ 34,
+ 158,
+ 132,
+ 0.992346465587616,
+ 0.0
+ ],
+ [
+ 135,
+ 34,
+ 158,
+ 56,
+ 0.9793149828910828,
+ 3.0
+ ]
+ ],
+ "gt": [
+ [
+ 133,
+ 37,
+ 159,
+ 136,
+ 1.0,
+ 0.0
+ ],
+ [
+ 138,
+ 36,
+ 157,
+ 55,
+ 1.0,
+ 3.0
+ ]
+ ],
+ "ok": true
+ },
+ {
+ "name": "f4d87fce4b3e4787aa1d2d9ccf4b218e.jpg",
+ "nickname": "congche_hld_69.jpg",
+ "pred": [
+ [
+ 141,
+ 6,
+ 182,
+ 111,
+ 0.6584206819534302,
+ 0.0
+ ],
+ [
+ 142,
+ 66,
+ 182,
+ 108,
+ 0.9517905712127686,
+ 2.0
+ ]
+ ],
+ "gt": [
+ [
+ 138,
+ 6,
+ 182,
+ 106,
+ 1.0,
+ 0.0
+ ],
+ [
+ 139,
+ 61,
+ 181,
+ 104,
+ 1.0,
+ 2.0
+ ]
+ ],
+ "ok": true
+ },
+ {
+ "name": "4f48f768e6084ad1a0f7792a5303d32b.jpg",
+ "nickname": "congche_hld_485.jpg",
+ "pred": [
+ [
+ 119,
+ 41,
+ 147,
+ 130,
+ 0.9853512048721313,
+ 0.0
+ ],
+ [
+ 120,
+ 100,
+ 146,
+ 129,
+ 0.9472789764404297,
+ 2.0
+ ]
+ ],
+ "gt": [
+ [
+ 120,
+ 42,
+ 148,
+ 127,
+ 1.0,
+ 0.0
+ ],
+ [
+ 122,
+ 97,
+ 145,
+ 127,
+ 1.0,
+ 2.0
+ ]
+ ],
+ "ok": true
+ },
+ {
+ "name": "7ff479002f4d4b71a1ed1c02a0a0c00a.jpg",
+ "nickname": "B_619.jpg",
+ "pred": [
+ [
+ 147,
+ 42,
+ 176,
+ 163,
+ 0.9898746609687805,
+ 0.0
+ ],
+ [
+ 145,
+ 135,
+ 173,
+ 163,
+ 0.9800438284873962,
+ 2.0
+ ]
+ ],
+ "gt": [
+ [
+ 147,
+ 138,
+ 176,
+ 162,
+ 1.0,
+ 2.0
+ ],
+ [
+ 150,
+ 43,
+ 177,
+ 162,
+ 1.0,
+ 2.0
+ ]
+ ],
+ "ok": false
+ }
+ ]
+ }
+}
\ No newline at end of file
diff --git a/002_B_Car/model-178950.nncase/main.py b/002_B_Car/model-178950.nncase/main.py
new file mode 100644
index 0000000..96aa7e6
--- /dev/null
+++ b/002_B_Car/model-178950.nncase/main.py
@@ -0,0 +1,112 @@
+# generated by maixhub, tested on maixpy3 v0.4.8
+# copy files to TF card and plug into board and power on
+import sensor, image, lcd, time
+import KPU as kpu
+from machine import UART
+import gc, sys
+from fpioa_manager import fm
+
+input_size = (224, 224)
+labels = ['light']
+anchors = [1.06, 1.05, 0.91, 0.97, 0.72, 0.72, 0.84, 0.84, 1.06, 0.88]
+
+def lcd_show_except(e):
+ import uio
+ err_str = uio.StringIO()
+ sys.print_exception(e, err_str)
+ err_str = err_str.getvalue()
+ img = image.Image(size=input_size)
+ img.draw_string(0, 10, err_str, scale=1, color=(0xff,0x00,0x00))
+ lcd.display(img)
+
+class Comm:
+ def __init__(self, uart):
+ self.uart = uart
+
+ def send_detect_result(self, objects, labels):
+ msg = ""
+ for obj in objects:
+ pos = obj.rect()
+ p = obj.value()
+ idx = obj.classid()
+ label = labels[idx]
+ msg += "{}:{}:{}:{}:{}:{:.2f}:{}, ".format(pos[0], pos[1], pos[2], pos[3], idx, p, label)
+ if msg:
+ msg = msg[:-2] + "\n"
+ self.uart.write(msg.encode())
+
+def init_uart():
+ fm.register(10, fm.fpioa.UART1_TX, force=True)
+ fm.register(11, fm.fpioa.UART1_RX, force=True)
+
+ uart = UART(UART.UART1, 115200, 8, 0, 0, timeout=1000, read_buf_len=256)
+ return uart
+
+def main(anchors, labels = None, model_addr="/sd/m.kmodel", sensor_window=input_size, lcd_rotation=0, sensor_hmirror=False, sensor_vflip=False):
+ sensor.reset()
+ sensor.set_pixformat(sensor.RGB565)
+ sensor.set_framesize(sensor.QVGA)
+ sensor.set_windowing(sensor_window)
+ sensor.set_hmirror(sensor_hmirror)
+ sensor.set_vflip(sensor_vflip)
+ sensor.run(1)
+
+ lcd.init(type=1)
+ lcd.rotation(lcd_rotation)
+ lcd.clear(lcd.WHITE)
+
+ if not labels:
+ with open('labels.txt','r') as f:
+ exec(f.read())
+ if not labels:
+ print("no labels.txt")
+ img = image.Image(size=(320, 240))
+ img.draw_string(90, 110, "no labels.txt", color=(255, 0, 0), scale=2)
+ lcd.display(img)
+ return 1
+ try:
+ img = image.Image("startup.jpg")
+ lcd.display(img)
+ except Exception:
+ img = image.Image(size=(320, 240))
+ img.draw_string(90, 110, "loading model...", color=(255, 255, 255), scale=2)
+ lcd.display(img)
+
+ uart = init_uart()
+ comm = Comm(uart)
+
+ try:
+ task = None
+ task = kpu.load(model_addr)
+ kpu.init_yolo2(task, 0.4, 0.3, 5, anchors) # threshold:[0,1], nms_value: [0, 1]
+ while(True):
+ img = sensor.snapshot()
+ t = time.ticks_ms()
+ objects = kpu.run_yolo2(task, img)
+ t = time.ticks_ms() - t
+ if objects:
+ for obj in objects:
+ pos = obj.rect()
+ img.draw_rectangle(pos)
+ img.draw_string(pos[0], pos[1], "%s : %.2f" %(labels[obj.classid()], obj.value()), scale=2, color=(255, 0, 0))
+ comm.send_detect_result(objects, labels)
+ img.draw_string(0, 200, "t:%dms" %(t), scale=2, color=(255, 0, 0))
+ img.draw_string(0, 2, "Upgrade to MaixCAM to use YOLOv8", scale=1.2, color=(255, 0, 0))
+ img.draw_string(0, 30, "wiki.sipeed.com/maixcam", scale=1.2, color=(255, 0, 0))
+ lcd.display(img)
+ except Exception as e:
+ raise e
+ finally:
+ if not task is None:
+ kpu.deinit(task)
+
+
+if __name__ == "__main__":
+ try:
+ # main(anchors = anchors, labels=labels, model_addr=0x300000, lcd_rotation=0)
+ main(anchors = anchors, labels=labels, model_addr="/sd/model-178950.kmodel")
+ except Exception as e:
+ sys.print_exception(e)
+ lcd_show_except(e)
+ finally:
+ gc.collect()
diff --git a/002_B_Car/model-178950.nncase/model-178950.kmodel b/002_B_Car/model-178950.nncase/model-178950.kmodel
new file mode 100644
index 0000000..b268144
Binary files /dev/null and b/002_B_Car/model-178950.nncase/model-178950.kmodel differ
diff --git a/002_B_Car/model-178950.nncase/report.json b/002_B_Car/model-178950.nncase/report.json
new file mode 100644
index 0000000..53fc572
--- /dev/null
+++ b/002_B_Car/model-178950.nncase/report.json
@@ -0,0 +1,1503 @@
+{
+ "anchors": [
+ 1.06,
+ 1.05,
+ 0.91,
+ 0.97,
+ 0.72,
+ 0.72,
+ 0.84,
+ 0.84,
+ 1.06,
+ 0.88
+ ],
+ "inputs": {
+ "input0": [
+ 224,
+ 224,
+ 3
+ ]
+ },
+ "outputs": {
+ "output0": [
+ 7,
+ 7,
+ 30
+ ]
+ },
+ "params_quantity": 1856046,
+ "loss": [
+ [
+ 1,
+ 0.17391
+ ],
+ [
+ 2,
+ 0.08725
+ ],
+ [
+ 3,
+ 0.06588
+ ],
+ [
+ 4,
+ 0.05287
+ ],
+ [
+ 5,
+ 0.04661
+ ],
+ [
+ 6,
+ 0.08786
+ ],
+ [
+ 7,
+ 0.04733
+ ],
+ [
+ 8,
+ 0.0374
+ ],
+ [
+ 9,
+ 0.03213
+ ],
+ [
+ 10,
+ 0.02791
+ ],
+ [
+ 11,
+ 0.02486
+ ],
+ [
+ 12,
+ 0.02266
+ ],
+ [
+ 13,
+ 0.0212
+ ],
+ [
+ 14,
+ 0.02086
+ ],
+ [
+ 15,
+ 0.02113
+ ],
+ [
+ 16,
+ 0.01962
+ ],
+ [
+ 17,
+ 0.01877
+ ],
+ [
+ 18,
+ 0.01764
+ ],
+ [
+ 19,
+ 0.01698
+ ],
+ [
+ 20,
+ 0.01654
+ ],
+ [
+ 21,
+ 0.01602
+ ],
+ [
+ 22,
+ 0.01557
+ ],
+ [
+ 23,
+ 0.01537
+ ],
+ [
+ 24,
+ 0.01518
+ ],
+ [
+ 25,
+ 0.01501
+ ],
+ [
+ 26,
+ 0.01466
+ ],
+ [
+ 27,
+ 0.01446
+ ],
+ [
+ 28,
+ 0.01421
+ ],
+ [
+ 29,
+ 0.01406
+ ],
+ [
+ 30,
+ 0.01399
+ ],
+ [
+ 31,
+ 0.01376
+ ],
+ [
+ 32,
+ 0.01365
+ ],
+ [
+ 33,
+ 0.01359
+ ],
+ [
+ 34,
+ 0.01345
+ ],
+ [
+ 35,
+ 0.01341
+ ],
+ [
+ 36,
+ 0.01339
+ ],
+ [
+ 37,
+ 0.01348
+ ],
+ [
+ 38,
+ 0.01331
+ ],
+ [
+ 39,
+ 0.01316
+ ],
+ [
+ 40,
+ 0.01327
+ ],
+ [
+ 41,
+ 0.0131
+ ],
+ [
+ 42,
+ 0.01316
+ ],
+ [
+ 43,
+ 0.0131
+ ],
+ [
+ 44,
+ 0.01295
+ ],
+ [
+ 45,
+ 0.01296
+ ],
+ [
+ 46,
+ 0.01295
+ ],
+ [
+ 47,
+ 0.01294
+ ],
+ [
+ 48,
+ 0.01283
+ ],
+ [
+ 49,
+ 0.01282
+ ],
+ [
+ 50,
+ 0.01283
+ ],
+ [
+ 51,
+ 0.0128
+ ],
+ [
+ 52,
+ 0.01276
+ ],
+ [
+ 53,
+ 0.0127
+ ],
+ [
+ 54,
+ 0.01266
+ ],
+ [
+ 55,
+ 0.01263
+ ],
+ [
+ 56,
+ 0.01261
+ ],
+ [
+ 57,
+ 0.01262
+ ],
+ [
+ 58,
+ 0.01262
+ ],
+ [
+ 59,
+ 0.01259
+ ],
+ [
+ 60,
+ 0.01263
+ ],
+ [
+ 61,
+ 0.01258
+ ],
+ [
+ 62,
+ 0.01255
+ ],
+ [
+ 63,
+ 0.01253
+ ],
+ [
+ 64,
+ 0.01251
+ ],
+ [
+ 65,
+ 0.01254
+ ],
+ [
+ 66,
+ 0.0125
+ ],
+ [
+ 67,
+ 0.01248
+ ],
+ [
+ 68,
+ 0.01244
+ ],
+ [
+ 69,
+ 0.01246
+ ],
+ [
+ 70,
+ 0.01243
+ ],
+ [
+ 71,
+ 0.01246
+ ],
+ [
+ 72,
+ 0.01256
+ ],
+ [
+ 73,
+ 0.01246
+ ],
+ [
+ 74,
+ 0.01243
+ ],
+ [
+ 75,
+ 0.01242
+ ],
+ [
+ 76,
+ 0.01244
+ ],
+ [
+ 77,
+ 0.01243
+ ],
+ [
+ 78,
+ 0.0124
+ ],
+ [
+ 79,
+ 0.01242
+ ],
+ [
+ 80,
+ 0.0124
+ ],
+ [
+ 81,
+ 0.01255
+ ],
+ [
+ 82,
+ 0.01239
+ ],
+ [
+ 83,
+ 0.01239
+ ],
+ [
+ 84,
+ 0.01238
+ ],
+ [
+ 85,
+ 0.01239
+ ],
+ [
+ 86,
+ 0.01238
+ ],
+ [
+ 87,
+ 0.01235
+ ],
+ [
+ 88,
+ 0.01235
+ ],
+ [
+ 89,
+ 0.01233
+ ],
+ [
+ 90,
+ 0.01234
+ ],
+ [
+ 91,
+ 0.01235
+ ],
+ [
+ 92,
+ 0.01247
+ ],
+ [
+ 93,
+ 0.01233
+ ],
+ [
+ 94,
+ 0.01233
+ ],
+ [
+ 95,
+ 0.01234
+ ],
+ [
+ 96,
+ 0.01245
+ ],
+ [
+ 97,
+ 0.01233
+ ],
+ [
+ 98,
+ 0.01233
+ ],
+ [
+ 99,
+ 0.01231
+ ],
+ [
+ 100,
+ 0.01231
+ ]
+ ],
+ "acc": [
+ [
+ 1,
+ 0.33034
+ ],
+ [
+ 2,
+ 0.34642
+ ],
+ [
+ 3,
+ 0.35676
+ ],
+ [
+ 4,
+ 0.34471
+ ],
+ [
+ 5,
+ 0.36597
+ ],
+ [
+ 6,
+ 0.47446
+ ],
+ [
+ 7,
+ 0.49297
+ ],
+ [
+ 8,
+ 0.50653
+ ],
+ [
+ 9,
+ 0.51814
+ ],
+ [
+ 10,
+ 0.51187
+ ],
+ [
+ 11,
+ 0.51509
+ ],
+ [
+ 12,
+ 0.51547
+ ],
+ [
+ 13,
+ 0.51679
+ ],
+ [
+ 14,
+ 0.51727
+ ],
+ [
+ 15,
+ 0.52781
+ ],
+ [
+ 16,
+ 0.53138
+ ],
+ [
+ 17,
+ 0.53227
+ ],
+ [
+ 18,
+ 0.53446
+ ],
+ [
+ 19,
+ 0.53107
+ ],
+ [
+ 20,
+ 0.52732
+ ],
+ [
+ 21,
+ 0.52929
+ ],
+ [
+ 22,
+ 0.52986
+ ],
+ [
+ 23,
+ 0.53367
+ ],
+ [
+ 24,
+ 0.52991
+ ],
+ [
+ 25,
+ 0.53143
+ ],
+ [
+ 26,
+ 0.53381
+ ],
+ [
+ 27,
+ 0.53165
+ ],
+ [
+ 28,
+ 0.53306
+ ],
+ [
+ 29,
+ 0.52996
+ ],
+ [
+ 30,
+ 0.53134
+ ],
+ [
+ 31,
+ 0.53052
+ ],
+ [
+ 32,
+ 0.53153
+ ],
+ [
+ 33,
+ 0.53264
+ ],
+ [
+ 34,
+ 0.53177
+ ],
+ [
+ 35,
+ 0.53482
+ ],
+ [
+ 36,
+ 0.53255
+ ],
+ [
+ 37,
+ 0.5323
+ ],
+ [
+ 38,
+ 0.52955
+ ],
+ [
+ 39,
+ 0.53212
+ ],
+ [
+ 40,
+ 0.5299
+ ],
+ [
+ 41,
+ 0.5292
+ ],
+ [
+ 42,
+ 0.52717
+ ],
+ [
+ 43,
+ 0.5268
+ ],
+ [
+ 44,
+ 0.52881
+ ],
+ [
+ 45,
+ 0.52693
+ ],
+ [
+ 46,
+ 0.5279
+ ],
+ [
+ 47,
+ 0.52839
+ ],
+ [
+ 48,
+ 0.52916
+ ],
+ [
+ 49,
+ 0.52705
+ ],
+ [
+ 50,
+ 0.52753
+ ],
+ [
+ 51,
+ 0.52835
+ ],
+ [
+ 52,
+ 0.52801
+ ],
+ [
+ 53,
+ 0.52889
+ ],
+ [
+ 54,
+ 0.52795
+ ],
+ [
+ 55,
+ 0.52537
+ ],
+ [
+ 56,
+ 0.52522
+ ],
+ [
+ 57,
+ 0.52518
+ ],
+ [
+ 58,
+ 0.52631
+ ],
+ [
+ 59,
+ 0.52653
+ ],
+ [
+ 60,
+ 0.52506
+ ],
+ [
+ 61,
+ 0.52379
+ ],
+ [
+ 62,
+ 0.52605
+ ],
+ [
+ 63,
+ 0.52556
+ ],
+ [
+ 64,
+ 0.52462
+ ],
+ [
+ 65,
+ 0.52514
+ ],
+ [
+ 66,
+ 0.52388
+ ],
+ [
+ 67,
+ 0.52369
+ ],
+ [
+ 68,
+ 0.52347
+ ],
+ [
+ 69,
+ 0.52444
+ ],
+ [
+ 70,
+ 0.52473
+ ],
+ [
+ 71,
+ 0.52438
+ ],
+ [
+ 72,
+ 0.52495
+ ],
+ [
+ 73,
+ 0.52485
+ ],
+ [
+ 74,
+ 0.52537
+ ],
+ [
+ 75,
+ 0.52501
+ ],
+ [
+ 76,
+ 0.5243
+ ],
+ [
+ 77,
+ 0.5246
+ ],
+ [
+ 78,
+ 0.52312
+ ],
+ [
+ 79,
+ 0.52417
+ ],
+ [
+ 80,
+ 0.52381
+ ],
+ [
+ 81,
+ 0.5244
+ ],
+ [
+ 82,
+ 0.52341
+ ],
+ [
+ 83,
+ 0.52379
+ ],
+ [
+ 84,
+ 0.52398
+ ],
+ [
+ 85,
+ 0.52362
+ ],
+ [
+ 86,
+ 0.52393
+ ],
+ [
+ 87,
+ 0.52316
+ ],
+ [
+ 88,
+ 0.52332
+ ],
+ [
+ 89,
+ 0.52309
+ ],
+ [
+ 90,
+ 0.52281
+ ],
+ [
+ 91,
+ 0.52337
+ ],
+ [
+ 92,
+ 0.52342
+ ],
+ [
+ 93,
+ 0.5232
+ ],
+ [
+ 94,
+ 0.52212
+ ],
+ [
+ 95,
+ 0.52194
+ ],
+ [
+ 96,
+ 0.52259
+ ],
+ [
+ 97,
+ 0.5224
+ ],
+ [
+ 98,
+ 0.52269
+ ],
+ [
+ 99,
+ 0.52198
+ ],
+ [
+ 100,
+ 0.52091
+ ]
+ ],
+ "lr": [
+ [
+ 1,
+ 0.001
+ ],
+ [
+ 2,
+ 0.00098
+ ],
+ [
+ 3,
+ 0.00095
+ ],
+ [
+ 4,
+ 0.00093
+ ],
+ [
+ 5,
+ 0.00091
+ ],
+ [
+ 6,
+ 0.00089
+ ],
+ [
+ 7,
+ 0.00087
+ ],
+ [
+ 8,
+ 0.00085
+ ],
+ [
+ 9,
+ 0.00083
+ ],
+ [
+ 10,
+ 0.00081
+ ],
+ [
+ 11,
+ 0.00079
+ ],
+ [
+ 12,
+ 0.00078
+ ],
+ [
+ 13,
+ 0.00076
+ ],
+ [
+ 14,
+ 0.00074
+ ],
+ [
+ 15,
+ 0.00072
+ ],
+ [
+ 16,
+ 0.00071
+ ],
+ [
+ 17,
+ 0.00069
+ ],
+ [
+ 18,
+ 0.00068
+ ],
+ [
+ 19,
+ 0.00066
+ ],
+ [
+ 20,
+ 0.00065
+ ],
+ [
+ 21,
+ 0.00063
+ ],
+ [
+ 22,
+ 0.00062
+ ],
+ [
+ 23,
+ 0.0006
+ ],
+ [
+ 24,
+ 0.00059
+ ],
+ [
+ 25,
+ 0.00058
+ ],
+ [
+ 26,
+ 0.00056
+ ],
+ [
+ 27,
+ 0.00055
+ ],
+ [
+ 28,
+ 0.00054
+ ],
+ [
+ 29,
+ 0.00052
+ ],
+ [
+ 30,
+ 0.00051
+ ],
+ [
+ 31,
+ 0.0005
+ ],
+ [
+ 32,
+ 0.00049
+ ],
+ [
+ 33,
+ 0.00048
+ ],
+ [
+ 34,
+ 0.00047
+ ],
+ [
+ 35,
+ 0.00046
+ ],
+ [
+ 36,
+ 0.00045
+ ],
+ [
+ 37,
+ 0.00044
+ ],
+ [
+ 38,
+ 0.00043
+ ],
+ [
+ 39,
+ 0.00042
+ ],
+ [
+ 40,
+ 0.00041
+ ],
+ [
+ 41,
+ 0.0004
+ ],
+ [
+ 42,
+ 0.00039
+ ],
+ [
+ 43,
+ 0.00038
+ ],
+ [
+ 44,
+ 0.00037
+ ],
+ [
+ 45,
+ 0.00036
+ ],
+ [
+ 46,
+ 0.00035
+ ],
+ [
+ 47,
+ 0.00035
+ ],
+ [
+ 48,
+ 0.00034
+ ],
+ [
+ 49,
+ 0.00033
+ ],
+ [
+ 50,
+ 0.00032
+ ],
+ [
+ 51,
+ 0.00032
+ ],
+ [
+ 52,
+ 0.00031
+ ],
+ [
+ 53,
+ 0.0003
+ ],
+ [
+ 54,
+ 0.0003
+ ],
+ [
+ 55,
+ 0.00029
+ ],
+ [
+ 56,
+ 0.00028
+ ],
+ [
+ 57,
+ 0.00028
+ ],
+ [
+ 58,
+ 0.00027
+ ],
+ [
+ 59,
+ 0.00026
+ ],
+ [
+ 60,
+ 0.00026
+ ],
+ [
+ 61,
+ 0.00025
+ ],
+ [
+ 62,
+ 0.00025
+ ],
+ [
+ 63,
+ 0.00024
+ ],
+ [
+ 64,
+ 0.00023
+ ],
+ [
+ 65,
+ 0.00023
+ ],
+ [
+ 66,
+ 0.00022
+ ],
+ [
+ 67,
+ 0.00022
+ ],
+ [
+ 68,
+ 0.00021
+ ],
+ [
+ 69,
+ 0.00021
+ ],
+ [
+ 70,
+ 0.0002
+ ],
+ [
+ 71,
+ 0.0002
+ ],
+ [
+ 72,
+ 0.00019
+ ],
+ [
+ 73,
+ 0.00019
+ ],
+ [
+ 74,
+ 0.00019
+ ],
+ [
+ 75,
+ 0.00018
+ ],
+ [
+ 76,
+ 0.00018
+ ],
+ [
+ 77,
+ 0.00017
+ ],
+ [
+ 78,
+ 0.00017
+ ],
+ [
+ 79,
+ 0.00017
+ ],
+ [
+ 80,
+ 0.00016
+ ],
+ [
+ 81,
+ 0.00016
+ ],
+ [
+ 82,
+ 0.00015
+ ],
+ [
+ 83,
+ 0.00015
+ ],
+ [
+ 84,
+ 0.00015
+ ],
+ [
+ 85,
+ 0.00014
+ ],
+ [
+ 86,
+ 0.00014
+ ],
+ [
+ 87,
+ 0.00014
+ ],
+ [
+ 88,
+ 0.00013
+ ],
+ [
+ 89,
+ 0.00013
+ ],
+ [
+ 90,
+ 0.00013
+ ],
+ [
+ 91,
+ 0.00013
+ ],
+ [
+ 92,
+ 0.00012
+ ],
+ [
+ 93,
+ 0.00012
+ ],
+ [
+ 94,
+ 0.00012
+ ],
+ [
+ 95,
+ 0.00011
+ ],
+ [
+ 96,
+ 0.00011
+ ],
+ [
+ 97,
+ 0.00011
+ ],
+ [
+ 98,
+ 0.00011
+ ],
+ [
+ 99,
+ 0.0001
+ ],
+ [
+ 100,
+ 0.0001
+ ]
+ ],
+ "val_acc": [
+ [
+ 10,
+ 0.0
+ ],
+ [
+ 20,
+ 0.42678
+ ],
+ [
+ 30,
+ 0.89731
+ ],
+ [
+ 40,
+ 0.8987
+ ],
+ [
+ 50,
+ 0.90647
+ ],
+ [
+ 60,
+ 0.89653
+ ],
+ [
+ 70,
+ 0.89863
+ ],
+ [
+ 80,
+ 0.89863
+ ],
+ [
+ 90,
+ 0.89875
+ ],
+ [
+ 100,
+ 0.89863
+ ]
+ ],
+ "AP_light": [
+ [
+ 10,
+ 0.0
+ ],
+ [
+ 20,
+ 0.42678
+ ],
+ [
+ 30,
+ 0.89731
+ ],
+ [
+ 40,
+ 0.8987
+ ],
+ [
+ 50,
+ 0.90647
+ ],
+ [
+ 60,
+ 0.89653
+ ],
+ [
+ 70,
+ 0.89863
+ ],
+ [
+ 80,
+ 0.89863
+ ],
+ [
+ 90,
+ 0.89875
+ ],
+ [
+ 100,
+ 0.89863
+ ]
+ ],
+ "mean": 123.5,
+ "std": 58.395,
+ "label_type": "detection",
+ "data_type": "image",
+ "labels": [
+ "light"
+ ],
+ "best_eval": {
+ "epoch": 50,
+ "val_acc": 0.906,
+ "val_info": [
+ {
+ "name": "eadc065910444670b27ef6369071c0f0.jpg",
+ "nickname": "hld2_135_16.jpg",
+ "pred": [
+ [
+ 130,
+ 55,
+ 166,
+ 95,
+ 0.9348750114440918,
+ 0.0
+ ],
+ [
+ 118,
+ 125,
+ 169,
+ 166,
+ 0.8402513265609741,
+ 0.0
+ ],
+ [
+ 121,
+ 201,
+ 164,
+ 237,
+ 0.8965619802474976,
+ 0.0
+ ]
+ ],
+ "gt": [
+ [
+ 126,
+ 56,
+ 171,
+ 99,
+ 1.0,
+ 0.0
+ ],
+ [
+ 120,
+ 129,
+ 170,
+ 170,
+ 1.0,
+ 0.0
+ ],
+ [
+ 120,
+ 194,
+ 170,
+ 231,
+ 1.0,
+ 0.0
+ ]
+ ],
+ "ok": true
+ },
+ {
+ "name": "daa9ba61ae4440df830f6f2d18b00e0b.jpg",
+ "nickname": "hld2_135_8.jpg",
+ "pred": [
+ [
+ 130,
+ 55,
+ 166,
+ 96,
+ 0.9444698095321655,
+ 0.0
+ ],
+ [
+ 129,
+ 130,
+ 164,
+ 165,
+ 0.8673686981201172,
+ 0.0
+ ],
+ [
+ 124,
+ 194,
+ 173,
+ 234,
+ 0.9432332515716553,
+ 0.0
+ ]
+ ],
+ "gt": [
+ [
+ 130,
+ 60,
+ 173,
+ 99,
+ 1.0,
+ 0.0
+ ],
+ [
+ 121,
+ 126,
+ 173,
+ 164,
+ 1.0,
+ 0.0
+ ],
+ [
+ 113,
+ 191,
+ 169,
+ 239,
+ 1.0,
+ 0.0
+ ]
+ ],
+ "ok": true
+ },
+ {
+ "name": "7cd7ef14f87846ffa966113be9694adc.jpg",
+ "nickname": "hld2_136_9.jpg",
+ "pred": [
+ [
+ 132,
+ 55,
+ 173,
+ 101,
+ 0.9230822920799255,
+ 0.0
+ ],
+ [
+ 131,
+ 127,
+ 174,
+ 171,
+ 0.9599894881248474,
+ 0.0
+ ],
+ [
+ 127,
+ 202,
+ 171,
+ 236,
+ 0.9505459666252136,
+ 0.0
+ ]
+ ],
+ "gt": [
+ [
+ 139,
+ 57,
+ 173,
+ 97,
+ 1.0,
+ 0.0
+ ],
+ [
+ 133,
+ 133,
+ 169,
+ 170,
+ 1.0,
+ 0.0
+ ],
+ [
+ 130,
+ 206,
+ 170,
+ 239,
+ 1.0,
+ 0.0
+ ]
+ ],
+ "ok": true
+ }
+ ]
+ }
+}
\ No newline at end of file
diff --git a/002_B_Car/model-179117.nncase.zip b/002_B_Car/model-179117.nncase.zip
new file mode 100644
index 0000000..d5992bc
Binary files /dev/null and b/002_B_Car/model-179117.nncase.zip differ
diff --git a/002_B_Car/model-179147.nncase.zip b/002_B_Car/model-179147.nncase.zip
new file mode 100644
index 0000000..d867a4d
Binary files /dev/null and b/002_B_Car/model-179147.nncase.zip differ
diff --git a/002_B_Car/model-179299.nncase/main.py b/002_B_Car/model-179299.nncase/main.py
new file mode 100644
index 0000000..11127e8
--- /dev/null
+++ b/002_B_Car/model-179299.nncase/main.py
@@ -0,0 +1,113 @@
+# generated by maixhub, tested on maixpy3 v0.4.8
+# copy files to TF card and plug into board and power on
+import sensor, image, lcd, time
+import KPU as kpu
+from machine import UART
+import gc, sys
+from fpioa_manager import fm
+
+input_size = (224, 224)
+labels = ['light']
+anchors = [0.84, 0.5, 0.78, 0.81, 0.84, 0.91, 0.78, 0.34, 0.72, 0.72]
+
+def lcd_show_except(e):
+ import uio
+ err_str = uio.StringIO()
+ sys.print_exception(e, err_str)
+ err_str = err_str.getvalue()
+ img = image.Image(size=input_size)
+ img.draw_string(0, 10, err_str, scale=1, color=(0xff,0x00,0x00))
+ lcd.display(img)
+
+class Comm:
+ def __init__(self, uart):
+ self.uart = uart
+
+ def send_detect_result(self, objects, labels):
+ msg = ""
+ for obj in objects:
+ pos = obj.rect()
+ p = obj.value()
+ idx = obj.classid()
+ label = labels[idx]
+ msg += "{}:{}:{}:{}:{}:{:.2f}:{}, ".format(pos[0], pos[1], pos[2], pos[3], idx, p, label)
+ if msg:
+ msg = msg[:-2] + "\n"
+ self.uart.write(msg.encode())
+
+def init_uart():
+ fm.register(10, fm.fpioa.UART1_TX, force=True)
+ fm.register(11, fm.fpioa.UART1_RX, force=True)
+
+ uart = UART(UART.UART1, 115200, 8, 0, 0, timeout=1000, read_buf_len=256)
+ return uart
+
+def main(anchors, labels = None, model_addr="/sd/m.kmodel", sensor_window=input_size, lcd_rotation=0, sensor_hmirror=False, sensor_vflip=False):
+ sensor.reset()
+ sensor.set_pixformat(sensor.RGB565)
+ sensor.set_framesize(sensor.QVGA)
+ sensor.set_windowing(sensor_window)
+ sensor.set_hmirror(sensor_hmirror)
+ sensor.set_vflip(sensor_vflip)
+ sensor.run(1)
+
+ lcd.init(type=1)
+ lcd.rotation(lcd_rotation)
+ lcd.clear(lcd.WHITE)
+
+ if not labels:
+ with open('labels.txt','r') as f:
+ exec(f.read())
+ if not labels:
+ print("no labels.txt")
+ img = image.Image(size=(320, 240))
+ img.draw_string(90, 110, "no labels.txt", color=(255, 0, 0), scale=2)
+ lcd.display(img)
+ return 1
+ try:
+ img = image.Image("startup.jpg")
+ lcd.display(img)
+ except Exception:
+ img = image.Image(size=(320, 240))
+ img.draw_string(90, 110, "loading model...", color=(255, 255, 255), scale=2)
+ lcd.display(img)
+
+ uart = init_uart()
+ comm = Comm(uart)
+
+ try:
+ task = None
+ task = kpu.load(model_addr)
+ kpu.init_yolo2(task, 0.3, 0.3, 5, anchors) # threshold:[0,1], nms_value: [0, 1]
+ while(True):
+ t = time.ticks_ms()
+ img = sensor.snapshot()
+ img = img.rotation_corr(z_rotation=180) # 旋转90度
+ objects = kpu.run_yolo2(task, img)
+ t = time.ticks_ms() - t
+ if objects:
+ for obj in objects:
+ pos = obj.rect()
+ img.draw_rectangle(pos)
+ img.draw_string(pos[0], pos[1], "%s : %.2f" %(labels[obj.classid()], obj.value()), scale=2, color=(255, 0, 0))
+ comm.send_detect_result(objects, labels)
+ img.draw_string(0, 200, "t:%dms" %(t), scale=2, color=(255, 0, 0))
+ # img.draw_string(0, 2, "Upgrade to MaixCAM to use YOLOv8", scale=1.2, color=(255, 0, 0))
+ # img.draw_string(0, 30, "wiki.sipeed.com/maixcam", scale=1.2, color=(255, 0, 0))
+ lcd.display(img)
+ except Exception as e:
+ raise e
+ finally:
+ if not task is None:
+ kpu.deinit(task)
+
+
+if __name__ == "__main__":
+ try:
+ # main(anchors = anchors, labels=labels, model_addr=0x300000, lcd_rotation=0)
+ main(anchors = anchors, labels=labels, model_addr="/sd/model-179299.kmodel")
+ except Exception as e:
+ sys.print_exception(e)
+ lcd_show_except(e)
+ finally:
+ gc.collect()
diff --git a/002_B_Car/model-179299.nncase/model-179299.kmodel b/002_B_Car/model-179299.nncase/model-179299.kmodel
new file mode 100644
index 0000000..413bafd
Binary files /dev/null and b/002_B_Car/model-179299.nncase/model-179299.kmodel differ
diff --git a/002_B_Car/model-179299.nncase/report.json b/002_B_Car/model-179299.nncase/report.json
new file mode 100644
index 0000000..2ff8f4d
--- /dev/null
+++ b/002_B_Car/model-179299.nncase/report.json
@@ -0,0 +1,1503 @@
+{
+ "anchors": [
+ 0.84,
+ 0.5,
+ 0.78,
+ 0.81,
+ 0.84,
+ 0.91,
+ 0.78,
+ 0.34,
+ 0.72,
+ 0.72
+ ],
+ "inputs": {
+ "input0": [
+ 224,
+ 224,
+ 3
+ ]
+ },
+ "outputs": {
+ "output0": [
+ 7,
+ 7,
+ 30
+ ]
+ },
+ "params_quantity": 1856046,
+ "loss": [
+ [
+ 1,
+ 0.1779
+ ],
+ [
+ 2,
+ 0.11251
+ ],
+ [
+ 3,
+ 0.05848
+ ],
+ [
+ 4,
+ 0.04483
+ ],
+ [
+ 5,
+ 0.03644
+ ],
+ [
+ 6,
+ 0.07361
+ ],
+ [
+ 7,
+ 0.02901
+ ],
+ [
+ 8,
+ 0.02153
+ ],
+ [
+ 9,
+ 0.01644
+ ],
+ [
+ 10,
+ 0.0141
+ ],
+ [
+ 11,
+ 0.01192
+ ],
+ [
+ 12,
+ 0.00935
+ ],
+ [
+ 13,
+ 0.00771
+ ],
+ [
+ 14,
+ 0.00637
+ ],
+ [
+ 15,
+ 0.00539
+ ],
+ [
+ 16,
+ 0.00463
+ ],
+ [
+ 17,
+ 0.00403
+ ],
+ [
+ 18,
+ 0.00349
+ ],
+ [
+ 19,
+ 0.00318
+ ],
+ [
+ 20,
+ 0.00291
+ ],
+ [
+ 21,
+ 0.0028
+ ],
+ [
+ 22,
+ 0.00251
+ ],
+ [
+ 23,
+ 0.0022
+ ],
+ [
+ 24,
+ 0.00212
+ ],
+ [
+ 25,
+ 0.00206
+ ],
+ [
+ 26,
+ 0.00194
+ ],
+ [
+ 27,
+ 0.00183
+ ],
+ [
+ 28,
+ 0.00163
+ ],
+ [
+ 29,
+ 0.00153
+ ],
+ [
+ 30,
+ 0.00134
+ ],
+ [
+ 31,
+ 0.00132
+ ],
+ [
+ 32,
+ 0.0013
+ ],
+ [
+ 33,
+ 0.00141
+ ],
+ [
+ 34,
+ 0.00127
+ ],
+ [
+ 35,
+ 0.00132
+ ],
+ [
+ 36,
+ 0.00124
+ ],
+ [
+ 37,
+ 0.00113
+ ],
+ [
+ 38,
+ 0.00105
+ ],
+ [
+ 39,
+ 0.00097
+ ],
+ [
+ 40,
+ 0.00086
+ ],
+ [
+ 41,
+ 0.00086
+ ],
+ [
+ 42,
+ 0.00083
+ ],
+ [
+ 43,
+ 0.00083
+ ],
+ [
+ 44,
+ 0.00082
+ ],
+ [
+ 45,
+ 0.00076
+ ],
+ [
+ 46,
+ 0.00076
+ ],
+ [
+ 47,
+ 0.00074
+ ],
+ [
+ 48,
+ 0.00071
+ ],
+ [
+ 49,
+ 0.00068
+ ],
+ [
+ 50,
+ 0.00063
+ ],
+ [
+ 51,
+ 0.00066
+ ],
+ [
+ 52,
+ 0.00061
+ ],
+ [
+ 53,
+ 0.0006
+ ],
+ [
+ 54,
+ 0.0006
+ ],
+ [
+ 55,
+ 0.00061
+ ],
+ [
+ 56,
+ 0.0006
+ ],
+ [
+ 57,
+ 0.00055
+ ],
+ [
+ 58,
+ 0.00052
+ ],
+ [
+ 59,
+ 0.00052
+ ],
+ [
+ 60,
+ 0.00046
+ ],
+ [
+ 61,
+ 0.00046
+ ],
+ [
+ 62,
+ 0.00044
+ ],
+ [
+ 63,
+ 0.00045
+ ],
+ [
+ 64,
+ 0.00045
+ ],
+ [
+ 65,
+ 0.00045
+ ],
+ [
+ 66,
+ 0.0004
+ ],
+ [
+ 67,
+ 0.00041
+ ],
+ [
+ 68,
+ 0.00039
+ ],
+ [
+ 69,
+ 0.0004
+ ],
+ [
+ 70,
+ 0.00042
+ ],
+ [
+ 71,
+ 0.00039
+ ],
+ [
+ 72,
+ 0.00037
+ ],
+ [
+ 73,
+ 0.00039
+ ],
+ [
+ 74,
+ 0.00036
+ ],
+ [
+ 75,
+ 0.00036
+ ],
+ [
+ 76,
+ 0.00033
+ ],
+ [
+ 77,
+ 0.00038
+ ],
+ [
+ 78,
+ 0.00034
+ ],
+ [
+ 79,
+ 0.00034
+ ],
+ [
+ 80,
+ 0.00034
+ ],
+ [
+ 81,
+ 0.00033
+ ],
+ [
+ 82,
+ 0.00032
+ ],
+ [
+ 83,
+ 0.0003
+ ],
+ [
+ 84,
+ 0.00029
+ ],
+ [
+ 85,
+ 0.00028
+ ],
+ [
+ 86,
+ 0.00027
+ ],
+ [
+ 87,
+ 0.00028
+ ],
+ [
+ 88,
+ 0.00027
+ ],
+ [
+ 89,
+ 0.00031
+ ],
+ [
+ 90,
+ 0.00027
+ ],
+ [
+ 91,
+ 0.00028
+ ],
+ [
+ 92,
+ 0.00025
+ ],
+ [
+ 93,
+ 0.00028
+ ],
+ [
+ 94,
+ 0.00029
+ ],
+ [
+ 95,
+ 0.00026
+ ],
+ [
+ 96,
+ 0.00027
+ ],
+ [
+ 97,
+ 0.00027
+ ],
+ [
+ 98,
+ 0.00027
+ ],
+ [
+ 99,
+ 0.00026
+ ],
+ [
+ 100,
+ 0.00027
+ ]
+ ],
+ "acc": [
+ [
+ 1,
+ 0.48564
+ ],
+ [
+ 2,
+ 0.43267
+ ],
+ [
+ 3,
+ 0.40026
+ ],
+ [
+ 4,
+ 0.39878
+ ],
+ [
+ 5,
+ 0.40469
+ ],
+ [
+ 6,
+ 0.43561
+ ],
+ [
+ 7,
+ 0.502
+ ],
+ [
+ 8,
+ 0.51537
+ ],
+ [
+ 9,
+ 0.51657
+ ],
+ [
+ 10,
+ 0.52604
+ ],
+ [
+ 11,
+ 0.52667
+ ],
+ [
+ 12,
+ 0.52168
+ ],
+ [
+ 13,
+ 0.51994
+ ],
+ [
+ 14,
+ 0.52391
+ ],
+ [
+ 15,
+ 0.52405
+ ],
+ [
+ 16,
+ 0.52383
+ ],
+ [
+ 17,
+ 0.52209
+ ],
+ [
+ 18,
+ 0.52022
+ ],
+ [
+ 19,
+ 0.51866
+ ],
+ [
+ 20,
+ 0.51673
+ ],
+ [
+ 21,
+ 0.51539
+ ],
+ [
+ 22,
+ 0.51326
+ ],
+ [
+ 23,
+ 0.51559
+ ],
+ [
+ 24,
+ 0.51302
+ ],
+ [
+ 25,
+ 0.50924
+ ],
+ [
+ 26,
+ 0.50878
+ ],
+ [
+ 27,
+ 0.50847
+ ],
+ [
+ 28,
+ 0.50827
+ ],
+ [
+ 29,
+ 0.50598
+ ],
+ [
+ 30,
+ 0.50338
+ ],
+ [
+ 31,
+ 0.50324
+ ],
+ [
+ 32,
+ 0.50135
+ ],
+ [
+ 33,
+ 0.5008
+ ],
+ [
+ 34,
+ 0.50374
+ ],
+ [
+ 35,
+ 0.50677
+ ],
+ [
+ 36,
+ 0.50451
+ ],
+ [
+ 37,
+ 0.50399
+ ],
+ [
+ 38,
+ 0.50409
+ ],
+ [
+ 39,
+ 0.50311
+ ],
+ [
+ 40,
+ 0.5038
+ ],
+ [
+ 41,
+ 0.50279
+ ],
+ [
+ 42,
+ 0.50332
+ ],
+ [
+ 43,
+ 0.50228
+ ],
+ [
+ 44,
+ 0.50175
+ ],
+ [
+ 45,
+ 0.50289
+ ],
+ [
+ 46,
+ 0.50226
+ ],
+ [
+ 47,
+ 0.50297
+ ],
+ [
+ 48,
+ 0.5034
+ ],
+ [
+ 49,
+ 0.50378
+ ],
+ [
+ 50,
+ 0.5036
+ ],
+ [
+ 51,
+ 0.50289
+ ],
+ [
+ 52,
+ 0.50385
+ ],
+ [
+ 53,
+ 0.50431
+ ],
+ [
+ 54,
+ 0.50496
+ ],
+ [
+ 55,
+ 0.50464
+ ],
+ [
+ 56,
+ 0.50443
+ ],
+ [
+ 57,
+ 0.50466
+ ],
+ [
+ 58,
+ 0.50522
+ ],
+ [
+ 59,
+ 0.50374
+ ],
+ [
+ 60,
+ 0.50293
+ ],
+ [
+ 61,
+ 0.5023
+ ],
+ [
+ 62,
+ 0.502
+ ],
+ [
+ 63,
+ 0.50216
+ ],
+ [
+ 64,
+ 0.50143
+ ],
+ [
+ 65,
+ 0.50151
+ ],
+ [
+ 66,
+ 0.50161
+ ],
+ [
+ 67,
+ 0.50147
+ ],
+ [
+ 68,
+ 0.50253
+ ],
+ [
+ 69,
+ 0.50305
+ ],
+ [
+ 70,
+ 0.50212
+ ],
+ [
+ 71,
+ 0.50176
+ ],
+ [
+ 72,
+ 0.50169
+ ],
+ [
+ 73,
+ 0.50202
+ ],
+ [
+ 74,
+ 0.50115
+ ],
+ [
+ 75,
+ 0.50159
+ ],
+ [
+ 76,
+ 0.50117
+ ],
+ [
+ 77,
+ 0.5022
+ ],
+ [
+ 78,
+ 0.50159
+ ],
+ [
+ 79,
+ 0.50145
+ ],
+ [
+ 80,
+ 0.501
+ ],
+ [
+ 81,
+ 0.50104
+ ],
+ [
+ 82,
+ 0.50175
+ ],
+ [
+ 83,
+ 0.50167
+ ],
+ [
+ 84,
+ 0.50249
+ ],
+ [
+ 85,
+ 0.50105
+ ],
+ [
+ 86,
+ 0.50113
+ ],
+ [
+ 87,
+ 0.50178
+ ],
+ [
+ 88,
+ 0.50038
+ ],
+ [
+ 89,
+ 0.50157
+ ],
+ [
+ 90,
+ 0.50169
+ ],
+ [
+ 91,
+ 0.50163
+ ],
+ [
+ 92,
+ 0.50111
+ ],
+ [
+ 93,
+ 0.50131
+ ],
+ [
+ 94,
+ 0.50072
+ ],
+ [
+ 95,
+ 0.50066
+ ],
+ [
+ 96,
+ 0.50102
+ ],
+ [
+ 97,
+ 0.50123
+ ],
+ [
+ 98,
+ 0.50192
+ ],
+ [
+ 99,
+ 0.50137
+ ],
+ [
+ 100,
+ 0.5018
+ ]
+ ],
+ "lr": [
+ [
+ 1,
+ 0.001
+ ],
+ [
+ 2,
+ 0.00098
+ ],
+ [
+ 3,
+ 0.00095
+ ],
+ [
+ 4,
+ 0.00093
+ ],
+ [
+ 5,
+ 0.00091
+ ],
+ [
+ 6,
+ 0.00089
+ ],
+ [
+ 7,
+ 0.00087
+ ],
+ [
+ 8,
+ 0.00085
+ ],
+ [
+ 9,
+ 0.00083
+ ],
+ [
+ 10,
+ 0.00081
+ ],
+ [
+ 11,
+ 0.00079
+ ],
+ [
+ 12,
+ 0.00078
+ ],
+ [
+ 13,
+ 0.00076
+ ],
+ [
+ 14,
+ 0.00074
+ ],
+ [
+ 15,
+ 0.00072
+ ],
+ [
+ 16,
+ 0.00071
+ ],
+ [
+ 17,
+ 0.00069
+ ],
+ [
+ 18,
+ 0.00068
+ ],
+ [
+ 19,
+ 0.00066
+ ],
+ [
+ 20,
+ 0.00065
+ ],
+ [
+ 21,
+ 0.00063
+ ],
+ [
+ 22,
+ 0.00062
+ ],
+ [
+ 23,
+ 0.0006
+ ],
+ [
+ 24,
+ 0.00059
+ ],
+ [
+ 25,
+ 0.00058
+ ],
+ [
+ 26,
+ 0.00056
+ ],
+ [
+ 27,
+ 0.00055
+ ],
+ [
+ 28,
+ 0.00054
+ ],
+ [
+ 29,
+ 0.00052
+ ],
+ [
+ 30,
+ 0.00051
+ ],
+ [
+ 31,
+ 0.0005
+ ],
+ [
+ 32,
+ 0.00049
+ ],
+ [
+ 33,
+ 0.00048
+ ],
+ [
+ 34,
+ 0.00047
+ ],
+ [
+ 35,
+ 0.00046
+ ],
+ [
+ 36,
+ 0.00045
+ ],
+ [
+ 37,
+ 0.00044
+ ],
+ [
+ 38,
+ 0.00043
+ ],
+ [
+ 39,
+ 0.00042
+ ],
+ [
+ 40,
+ 0.00041
+ ],
+ [
+ 41,
+ 0.0004
+ ],
+ [
+ 42,
+ 0.00039
+ ],
+ [
+ 43,
+ 0.00038
+ ],
+ [
+ 44,
+ 0.00037
+ ],
+ [
+ 45,
+ 0.00036
+ ],
+ [
+ 46,
+ 0.00035
+ ],
+ [
+ 47,
+ 0.00035
+ ],
+ [
+ 48,
+ 0.00034
+ ],
+ [
+ 49,
+ 0.00033
+ ],
+ [
+ 50,
+ 0.00032
+ ],
+ [
+ 51,
+ 0.00032
+ ],
+ [
+ 52,
+ 0.00031
+ ],
+ [
+ 53,
+ 0.0003
+ ],
+ [
+ 54,
+ 0.0003
+ ],
+ [
+ 55,
+ 0.00029
+ ],
+ [
+ 56,
+ 0.00028
+ ],
+ [
+ 57,
+ 0.00028
+ ],
+ [
+ 58,
+ 0.00027
+ ],
+ [
+ 59,
+ 0.00026
+ ],
+ [
+ 60,
+ 0.00026
+ ],
+ [
+ 61,
+ 0.00025
+ ],
+ [
+ 62,
+ 0.00025
+ ],
+ [
+ 63,
+ 0.00024
+ ],
+ [
+ 64,
+ 0.00023
+ ],
+ [
+ 65,
+ 0.00023
+ ],
+ [
+ 66,
+ 0.00022
+ ],
+ [
+ 67,
+ 0.00022
+ ],
+ [
+ 68,
+ 0.00021
+ ],
+ [
+ 69,
+ 0.00021
+ ],
+ [
+ 70,
+ 0.0002
+ ],
+ [
+ 71,
+ 0.0002
+ ],
+ [
+ 72,
+ 0.00019
+ ],
+ [
+ 73,
+ 0.00019
+ ],
+ [
+ 74,
+ 0.00019
+ ],
+ [
+ 75,
+ 0.00018
+ ],
+ [
+ 76,
+ 0.00018
+ ],
+ [
+ 77,
+ 0.00017
+ ],
+ [
+ 78,
+ 0.00017
+ ],
+ [
+ 79,
+ 0.00017
+ ],
+ [
+ 80,
+ 0.00016
+ ],
+ [
+ 81,
+ 0.00016
+ ],
+ [
+ 82,
+ 0.00015
+ ],
+ [
+ 83,
+ 0.00015
+ ],
+ [
+ 84,
+ 0.00015
+ ],
+ [
+ 85,
+ 0.00014
+ ],
+ [
+ 86,
+ 0.00014
+ ],
+ [
+ 87,
+ 0.00014
+ ],
+ [
+ 88,
+ 0.00013
+ ],
+ [
+ 89,
+ 0.00013
+ ],
+ [
+ 90,
+ 0.00013
+ ],
+ [
+ 91,
+ 0.00013
+ ],
+ [
+ 92,
+ 0.00012
+ ],
+ [
+ 93,
+ 0.00012
+ ],
+ [
+ 94,
+ 0.00012
+ ],
+ [
+ 95,
+ 0.00011
+ ],
+ [
+ 96,
+ 0.00011
+ ],
+ [
+ 97,
+ 0.00011
+ ],
+ [
+ 98,
+ 0.00011
+ ],
+ [
+ 99,
+ 0.0001
+ ],
+ [
+ 100,
+ 0.0001
+ ]
+ ],
+ "val_acc": [
+ [
+ 10,
+ 0.0
+ ],
+ [
+ 20,
+ 0.47849
+ ],
+ [
+ 30,
+ 0.85644
+ ],
+ [
+ 40,
+ 0.8894
+ ],
+ [
+ 50,
+ 0.90638
+ ],
+ [
+ 60,
+ 0.89675
+ ],
+ [
+ 70,
+ 0.89675
+ ],
+ [
+ 80,
+ 0.89675
+ ],
+ [
+ 90,
+ 0.87362
+ ],
+ [
+ 100,
+ 0.87362
+ ]
+ ],
+ "AP_light": [
+ [
+ 10,
+ 0.0
+ ],
+ [
+ 20,
+ 0.47849
+ ],
+ [
+ 30,
+ 0.85644
+ ],
+ [
+ 40,
+ 0.8894
+ ],
+ [
+ 50,
+ 0.90638
+ ],
+ [
+ 60,
+ 0.89675
+ ],
+ [
+ 70,
+ 0.89675
+ ],
+ [
+ 80,
+ 0.89675
+ ],
+ [
+ 90,
+ 0.87362
+ ],
+ [
+ 100,
+ 0.87362
+ ]
+ ],
+ "mean": 123.5,
+ "std": 58.395,
+ "label_type": "detection",
+ "data_type": "image",
+ "labels": [
+ "light"
+ ],
+ "best_eval": {
+ "epoch": 50,
+ "val_acc": 0.906,
+ "val_info": [
+ {
+ "name": "b07416149b15452abc57fcdcb85786e9.jpg",
+ "nickname": "198.jpg",
+ "pred": [
+ [
+ 147,
+ 23,
+ 178,
+ 56,
+ 0.781109094619751,
+ 0.0
+ ],
+ [
+ 146,
+ 97,
+ 176,
+ 129,
+ 0.8338834047317505,
+ 0.0
+ ],
+ [
+ 146,
+ 155,
+ 178,
+ 188,
+ 0.8692611455917358,
+ 0.0
+ ]
+ ],
+ "gt": [
+ [
+ 144,
+ 30,
+ 173,
+ 63,
+ 1.0,
+ 0.0
+ ],
+ [
+ 146,
+ 96,
+ 176,
+ 124,
+ 1.0,
+ 0.0
+ ],
+ [
+ 150,
+ 161,
+ 177,
+ 191,
+ 1.0,
+ 0.0
+ ]
+ ],
+ "ok": true
+ },
+ {
+ "name": "09795c3370664840844645f7ccacd1ac.jpg",
+ "nickname": "44.jpg",
+ "pred": [
+ [
+ 149,
+ 0,
+ 185,
+ 33,
+ 0.9628742933273315,
+ 0.0
+ ],
+ [
+ 150,
+ 65,
+ 184,
+ 98,
+ 0.9749233722686768,
+ 0.0
+ ],
+ [
+ 152,
+ 136,
+ 188,
+ 169,
+ 0.9425005912780762,
+ 0.0
+ ]
+ ],
+ "gt": [
+ [
+ 150,
+ 1,
+ 183,
+ 30,
+ 1.0,
+ 0.0
+ ],
+ [
+ 153,
+ 70,
+ 186,
+ 100,
+ 1.0,
+ 0.0
+ ],
+ [
+ 153,
+ 139,
+ 190,
+ 174,
+ 1.0,
+ 0.0
+ ]
+ ],
+ "ok": true
+ },
+ {
+ "name": "9399b25fb43f47f08e78fb3f62963d9f.jpg",
+ "nickname": "62.jpg",
+ "pred": [
+ [
+ 175,
+ 0,
+ 216,
+ 22,
+ 0.9558107256889343,
+ 0.0
+ ],
+ [
+ 175,
+ 65,
+ 213,
+ 107,
+ 0.9367208480834961,
+ 0.0
+ ],
+ [
+ 175,
+ 146,
+ 213,
+ 189,
+ 0.9285430908203125,
+ 0.0
+ ]
+ ],
+ "gt": [
+ [
+ 171,
+ 0,
+ 219,
+ 23,
+ 1.0,
+ 0.0
+ ],
+ [
+ 173,
+ 64,
+ 213,
+ 106,
+ 1.0,
+ 0.0
+ ],
+ [
+ 174,
+ 146,
+ 214,
+ 187,
+ 1.0,
+ 0.0
+ ]
+ ],
+ "ok": true
+ }
+ ]
+ }
+}
\ No newline at end of file
diff --git a/002_B_Car/xdl/len_B_sai_2.py b/002_B_Car/xdl/len_B_sai_2.py
index e1cec0f..706c6c4 100644
--- a/002_B_Car/xdl/len_B_sai_2.py
+++ b/002_B_Car/xdl/len_B_sai_2.py
@@ -128,7 +128,7 @@ class Mainlen():
self.tracking(img)
elif Flag_qr: # 二维码#
print("识别二维码")
- self.discem_QR(img)#r任务9
+ self.discem_QR(img)
# Flag_qr = False
elif Flag_light: # 红绿灯
print("识别红绿灯")
@@ -252,16 +252,14 @@ class Mainlen():
# result = self.remove_chinese_chars(result) # 剔除中文
print(result)
- if len(result) > 10: # 车牌
- result = self.jiajin(result)
- result = self.binary_to_hexadecimal(result)
- count = 2
-
- else: # 公式
+ if len(result) < 10: # 车牌
+ # result = self.jiajin(result)
+ # result = self.binary_to_hexadecimal(result)
count = 1
- if res == len(res_QR) - 1:
- if_end = 0x00
+ else: # 公式
+ continue
+
print("count: ", count)
print("if_end: ", if_end)
@@ -275,12 +273,8 @@ class Mainlen():
self.uart.write(bytes([0x06]))
print(len(result))
self.uart.write(bytes([count]))
- self.uart.write(bytes([if_end]))
- if count==2:
- self.uart.write(bytes([2]))
- self.uart.write(bytes([int(result[1]+result[0], 16)]))
- self.uart.write(bytes([int(result[3]+result[2], 16)]))
- else:
+ # self.uart.write(bytes([if_end]))
+ if count==1:
self.uart.write(bytes([len(result)]))
for qr_data in result:
self.uart.write(bytes([ord(qr_data)]))
diff --git a/002_B_Car/xdl/保存图片.py b/002_B_Car/xdl/保存图片.py
deleted file mode 100644
index 5ac7e72..0000000
--- a/002_B_Car/xdl/保存图片.py
+++ /dev/null
@@ -1,28 +0,0 @@
-import sensor, image, lcd, time
-from Maix import UART
-from fpioa_manager import fm
-
-# 初始化摄像头
-sensor.reset()
-sensor.set_pixformat(sensor.RGB565)
-sensor.set_framesize(sensor.QVGA)
-sensor.skip_frames(time=2000)
-
-# 初始化 LCD
-lcd.init()
-lcd.rotation(2)
-
-# 初始化 UART
-fm.register(7, fm.fpioa.UART1_TX) # TX 引脚映射
-fm.register(6, fm.fpioa.UART1_RX) # RX 引脚映射
-uart = UART(UART.UART1, 115200, read_buf_len=4096)
-
-print("按下按钮捕获图片并发送")
-
-while True:
- img = sensor.snapshot()
- lcd.display(img)
- if uart.read(1): # 等待串口信号（例如从电脑发送字符）
- print("开始发送图片...")
- uart.write(img.compress(quality=90).to_bytes()) # 压缩并发送图片
- print("图片发送完成")
diff --git a/002_B_Car/xdl/找圆.py b/002_B_Car/xdl/找圆.py
index 747436e..c268095 100644
--- a/002_B_Car/xdl/找圆.py
+++ b/002_B_Car/xdl/找圆.py
@@ -5,25 +5,21 @@
import sensor, lcd, time
from machine import Timer, PWM
-lcd.init(freq=15000000) # 初始化LCD
-sensor.reset() # 复位和初始化摄像头
-sensor.set_vflip(1) # 将摄像头设置成后置方式（所见即所得）
-sensor.set_pixformat(sensor.RGB565) # 设置像素格式为彩色 RGB565
-sensor.set_framesize(sensor.QVGA) # 设置帧大小为 QVGA (320x240)
+lcd.init(freq=15000000) # 初始化LCD
+sensor.reset() # 复位和初始化摄像头
+sensor.set_vflip(1) # 将摄像头设置成后置方式（所见即所得）
+sensor.set_pixformat(sensor.RGB565)# 设置像素格式为彩色 RGB565
+sensor.set_framesize(sensor.QVGA) # 设置帧大小为 QVGA (320x240)
sensor.set_auto_gain(False)
sensor.set_auto_whitebal(False)
sensor.set_auto_gain(0, 0)
-sensor.skip_frames(time=200) # 等待设置生效
+sensor.skip_frames(time=200) # 等待设置生效
def draw_circles():
"""
检测并绘制图像中的圆形，仅在图像的右半部分进行检测，
- 累计识别条件为超过4秒或识别到3个圆
"""
- result_cache = [] # 缓存每次有结果的识别
- max_results = 3 # 累计识别到的圆形数量
start_time = time.ticks_ms() # 记录开始时间（毫秒）
-
def circles_overlap(c1, c2):
"""判断两个圆是否重叠"""
dx = c1.x() - c2.x()
@@ -35,109 +31,111 @@ def draw_circles():
"""尝试合并新检测到的圆与缓存中的圆"""
for existing in cache:
if circles_overlap(new_circle, existing):
- return # 如果重叠，不加入新的圆
+ return False # 如果重叠，不加入新的圆
cache.append(new_circle) # 如果没有重叠，添加到缓存中
+ return True
def calculate_lab(area, img):
- """计算给定区域的中间 1/3 的平均 LAB 值"""
- mean_l, mean_a, mean_b = 0, 0, 0
- pixel_count = 0
- start_y = area[1] + area[3] // 3
- end_y = start_y + area[3] // 3
- for y in range(start_y, end_y):
- for x in range(area[0], area[0] + area[2]):
- l, a, b = img.get_statistics(roi=(x, y, 1, 1)).l_mean(), \
- img.get_statistics(roi=(x, y, 1, 1)).a_mean(), \
- img.get_statistics(roi=(x, y, 1, 1)).b_mean()
- mean_l += l
- mean_a += a
- mean_b += b
- pixel_count += 1
- mean_l //= pixel_count
- mean_a //= pixel_count
- mean_b //= pixel_count
- return mean_l, mean_a, mean_b
-
- def determine_farthest_circle(circles, img):
- """找到与其他圆 LAB 差异最大的圆"""
- start_time = time.ticks_ms() # 开始时间
-
- max_distance = -1
- farthest_circle = None
- farthest_lab = None
-
- for i, c1 in enumerate(circles):
- area1 = (c1.x() - c1.r(), c1.y() - c1.r(), 2 * c1.r(), 2 * c1.r())
- lab1 = calculate_lab(area1, img)
- total_distance = 0
-
- for j, c2 in enumerate(circles):
- if i != j:
- area2 = (c2.x() - c2.r(), c2.y() - c2.r(), 2 * c2.r(), 2 * c2.r())
- lab2 = calculate_lab(area2, img)
- total_distance += ((lab1[0] - lab2[0]) ** 2 + (lab1[1] - lab2[1]) ** 2 + (lab1[2] - lab2[2]) ** 2) ** 0.5
-
- if total_distance > max_distance:
- max_distance = total_distance
- farthest_circle = c1
- farthest_lab = lab1
-
- elapsed_time = time.ticks_diff(time.ticks_ms(), start_time) # 计算耗时
- print("determine_farthest_circle 耗时:" ,elapsed_time," ms")
- return farthest_circle, farthest_lab
+ """计算给定区域的平均 LAB 值"""
+ stats = img.get_statistics(roi=area)
+ return stats.l_mean(), stats.a_mean(), stats.b_mean()
def determine_color_from_lab(lab):
- """根据 LAB 值判断颜色倾向"""
+ """
+ 返回最接近的颜色以及对应的距离
+
+ :param lab: (l, a, b) 三元组
+ :return: (closest_color, min_distance)
+ """
+ # 你可以根据实际情况添加或修改更多预设颜色及其 LAB
predefined_colors = {
- "红色": (53, 80, 67),
- "黄色": (97, -21, 94),
- "绿色": (87, -86, 83)
+ "红色": (50, 38, 1),
+ "黄色": (86, 2, 27),
+ "绿色": (75, -29, 15)
}
min_distance = float('inf')
- closest_color = "灰色"
- for color_name, (l, a, b) in predefined_colors.items():
- distance = ((lab[0] - l) ** 2 + (lab[1] - a) ** 2 + (lab[2] - b) ** 2) ** 0.5
+ closest_color = "灰色" # 默认或者未匹配时
+
+ # 打印一下实际检测到的LAB，便于调试
+ # print("检测到的LAB:", lab)
+
+ for color_name, (l_ref, a_ref, b_ref) in predefined_colors.items():
+ distance = ((lab[0] - l_ref) ** 2 +
+ (lab[1] - a_ref) ** 2 +
+ (lab[2] - b_ref) ** 2) ** 0.5
if distance < min_distance:
min_distance = distance
closest_color = color_name
- return closest_color
+ if min_distance < 10:
+ print("最接近的颜色={}, 距离={:.2f}".format(closest_color, min_distance))
+ # print("最接近的颜色={}, 距离={:.2f}".format(closest_color, min_distance))
+ return closest_color, min_distance
+ count=0
+
+ min_color='red'
+ min_distance=50
while True:
img = sensor.snapshot()
- roi = (img.width() // 2, 0, img.width() // 2, img.height()) # 只处理右半部分
- # 查找圆形
- circles = img.find_circles(roi=roi, threshold=2500, x_margin=10, y_margin=10, r_margin=10,
- r_min=2, r_max=100, r_step=2)
+ # 只处理右半部分
+ roi = (img.width() // 2, 0, img.width() // 2, img.height())
+
+ # 在此根据整体亮度简单调参，阈值可自行调整
+ threshold = max(1000, min(5000, int(img.get_statistics().l_mean() * 50)))
+
+ # find_circles 参数可根据实际效果调整
+ circles = img.find_circles(roi=roi,
+ threshold=2500,
+ x_margin=20,
+ y_margin=20,
+ r_margin=20,
+ r_min=15,
+ r_max=30,
+ r_step=2)
+
if circles:
- # 如果有结果，尝试合并到结果缓存
+ closest_circle = None
+ closest_distance = float('inf')
+ closest_color = "未知"
+
for c in circles:
- merge_circles(c, result_cache)
+ area = (c.x() - c.r(), c.y() - c.r(), 2*c.r(), 2*c.r())
+ lab_val = calculate_lab(area, img)
+ color_name, distance = determine_color_from_lab(lab_val)
+ # 找出最小距离（与预设颜色最相似）的那个圆
+ if distance < closest_distance:
+ closest_distance = distance
+ closest_color = color_name
+ closest_circle = c
+ if closest_circle is None:
+ print('没有识别到')
+ continue
+ elif closest_distance<20:
+ area = (closest_circle.x() - closest_circle.r(),
+ closest_circle.y() - closest_circle.r(),
+ 2*closest_circle.r(), 2*closest_circle.r())
+ # 画框和画圆
+ img.draw_rectangle(area, color=(255, 0, 0))
+ img.draw_circle(closest_circle.x(), closest_circle.y(), closest_circle.r(), color=(255, 0, 0))
+ print("最终识别到的颜色={}, 距离={:.2f}".format(closest_color, closest_distance))
+ min_distance=50
+ print("{}轮检测耗时: {} ms".format(count,time.ticks_diff(time.ticks_ms(), start_time)))
+ start_time = time.ticks_ms()
+ count+=1
- # 检查是否达到累计条件或超时
- current_time = time.ticks_ms()
- elapsed_time = time.ticks_diff(current_time, start_time) # 计算耗时
- if len(result_cache) >= max_results or elapsed_time > 4000:
- # 找到 LAB 最不一样的圆
- farthest_circle, farthest_lab = determine_farthest_circle(result_cache, img)
+ else:
+ if closest_distance 5000:
+ print("最终识别到的颜色={}, 距离={:.2f}".format(min_color, min_distance))
+ print("{}轮检测耗时: {} ms".format(count,time.ticks_diff(time.ticks_ms(), start_time)))
+ start_time = time.ticks_ms()
- # 绘制并输出结果
- if farthest_circle:
- area = (farthest_circle.x() - farthest_circle.r(), farthest_circle.y() - farthest_circle.r(),
- 2 * farthest_circle.r(), 2 * farthest_circle.r())
- color = determine_color_from_lab(farthest_lab)
- if color in ["红色", "黄色", "绿色"]:
- img.draw_rectangle(area, color=(255, 255, 255)) # 画圆外接矩形
- img.draw_circle(farthest_circle.x(), farthest_circle.y(), farthest_circle.r(), color=(255, 0, 0))
- print("颜色倾向: ", color)
+ min_distance=50
+ count+=1
- current_time = time.ticks_ms()
- elapsed_time = time.ticks_diff(current_time, start_time) # 计算耗时
- print("保留的圆形数量：", len(result_cache))
- print("耗时:", elapsed_time, "ms")
- result_cache = [] # 清空缓存
- start_time = time.ticks_ms() # 重置开始时间
- time.sleep(0.1)
# 主函数调用
if __name__ == "__main__":
diff --git a/002_B_Car/xdl/找圆2.py b/002_B_Car/xdl/找圆2.py
new file mode 100644
index 0000000..c268095
--- /dev/null
+++ b/002_B_Car/xdl/找圆2.py
@@ -0,0 +1,142 @@
+# -*- coding:utf-8 -*-
+# @Author len
+# @Create 2023/11/30 11:08
+
+import sensor, lcd, time
+from machine import Timer, PWM
+
+lcd.init(freq=15000000) # 初始化LCD
+sensor.reset() # 复位和初始化摄像头
+sensor.set_vflip(1) # 将摄像头设置成后置方式（所见即所得）
+sensor.set_pixformat(sensor.RGB565)# 设置像素格式为彩色 RGB565
+sensor.set_framesize(sensor.QVGA) # 设置帧大小为 QVGA (320x240)
+sensor.set_auto_gain(False)
+sensor.set_auto_whitebal(False)
+sensor.set_auto_gain(0, 0)
+sensor.skip_frames(time=200) # 等待设置生效
+
+def draw_circles():
+ """
+ 检测并绘制图像中的圆形，仅在图像的右半部分进行检测，
+ """
+ start_time = time.ticks_ms() # 记录开始时间（毫秒）
+ def circles_overlap(c1, c2):
+ """判断两个圆是否重叠"""
+ dx = c1.x() - c2.x()
+ dy = c1.y() - c2.y()
+ distance = (dx**2 + dy**2)**0.5
+ return distance < (c1.r() + c2.r())
+
+ def merge_circles(new_circle, cache):
+ """尝试合并新检测到的圆与缓存中的圆"""
+ for existing in cache:
+ if circles_overlap(new_circle, existing):
+ return False # 如果重叠，不加入新的圆
+ cache.append(new_circle) # 如果没有重叠，添加到缓存中
+ return True
+
+ def calculate_lab(area, img):
+ """计算给定区域的平均 LAB 值"""
+ stats = img.get_statistics(roi=area)
+ return stats.l_mean(), stats.a_mean(), stats.b_mean()
+
+ def determine_color_from_lab(lab):
+ """
+ 返回最接近的颜色以及对应的距离
+
+ :param lab: (l, a, b) 三元组
+ :return: (closest_color, min_distance)
+ """
+ # 你可以根据实际情况添加或修改更多预设颜色及其 LAB
+ predefined_colors = {
+ "红色": (50, 38, 1),
+ "黄色": (86, 2, 27),
+ "绿色": (75, -29, 15)
+ }
+ min_distance = float('inf')
+ closest_color = "灰色" # 默认或者未匹配时
+
+ # 打印一下实际检测到的LAB，便于调试
+ # print("检测到的LAB:", lab)
+
+ for color_name, (l_ref, a_ref, b_ref) in predefined_colors.items():
+ distance = ((lab[0] - l_ref) ** 2 +
+ (lab[1] - a_ref) ** 2 +
+ (lab[2] - b_ref) ** 2) ** 0.5
+ if distance < min_distance:
+ min_distance = distance
+ closest_color = color_name
+ if min_distance < 10:
+ print("最接近的颜色={}, 距离={:.2f}".format(closest_color, min_distance))
+ # print("最接近的颜色={}, 距离={:.2f}".format(closest_color, min_distance))
+ return closest_color, min_distance
+
+ count=0
+
+ min_color='red'
+ min_distance=50
+ while True:
+ img = sensor.snapshot()
+ # 只处理右半部分
+ roi = (img.width() // 2, 0, img.width() // 2, img.height())
+
+ # 在此根据整体亮度简单调参，阈值可自行调整
+ threshold = max(1000, min(5000, int(img.get_statistics().l_mean() * 50)))
+
+ # find_circles 参数可根据实际效果调整
+ circles = img.find_circles(roi=roi,
+ threshold=2500,
+ x_margin=20,
+ y_margin=20,
+ r_margin=20,
+ r_min=15,
+ r_max=30,
+ r_step=2)
+
+ if circles:
+ closest_circle = None
+ closest_distance = float('inf')
+ closest_color = "未知"
+
+ for c in circles:
+ area = (c.x() - c.r(), c.y() - c.r(), 2*c.r(), 2*c.r())
+ lab_val = calculate_lab(area, img)
+ color_name, distance = determine_color_from_lab(lab_val)
+ # 找出最小距离（与预设颜色最相似）的那个圆
+ if distance < closest_distance:
+ closest_distance = distance
+ closest_color = color_name
+ closest_circle = c
+ if closest_circle is None:
+ print('没有识别到')
+ continue
+ elif closest_distance<20:
+ area = (closest_circle.x() - closest_circle.r(),
+ closest_circle.y() - closest_circle.r(),
+ 2*closest_circle.r(), 2*closest_circle.r())
+ # 画框和画圆
+ img.draw_rectangle(area, color=(255, 0, 0))
+ img.draw_circle(closest_circle.x(), closest_circle.y(), closest_circle.r(), color=(255, 0, 0))
+ print("最终识别到的颜色={}, 距离={:.2f}".format(closest_color, closest_distance))
+ min_distance=50
+ print("{}轮检测耗时: {} ms".format(count,time.ticks_diff(time.ticks_ms(), start_time)))
+ start_time = time.ticks_ms()
+ count+=1
+
+ else:
+ if closest_distance 5000:
+ print("最终识别到的颜色={}, 距离={:.2f}".format(min_color, min_distance))
+ print("{}轮检测耗时: {} ms".format(count,time.ticks_diff(time.ticks_ms(), start_time)))
+ start_time = time.ticks_ms()
+
+ min_distance=50
+ count+=1
+
+
+
+# 主函数调用
+if __name__ == "__main__":
+ draw_circles()
diff --git a/002_B_Car/主程序/len_B_240110_waste.py b/002_B_Car/主程序/len_B_240110_waste.py
index 64c51e4..f46ea59 100644
--- a/002_B_Car/主程序/len_B_240110_waste.py
+++ b/002_B_Car/主程序/len_B_240110_waste.py
@@ -141,6 +141,8 @@ class Mainlen():
sensor.set_pixformat(sensor.RGB565) # 设置像素格式为彩色 RGB565
# sensor.set_pixformat(sensor.GRAYSCALE) # 设置像素格式为灰色
sensor.set_framesize(sensor.QVGA) # 设置帧大小为QVGA（320×240）
+ # sensor.set_framesize(sensor.FRAME_240X240 ) # 设置帧大小为QVGA（320×240）
+
sensor.set_vflip(1) # 后置模式
sensor.skip_frames(30) # # 跳过前30帧
@@ -151,6 +153,8 @@ class Mainlen():
sensor.set_pixformat(sensor.RGB565) # 设置像素格式为彩色 RGB565
# sensor.set_pixformat(sensor.GRAYSCALE) # 设置像素格式为灰色
sensor.set_framesize(sensor.QVGA) # 设置帧大小为 QVGA (320x240)
+ # sensor.set_framesize(sensor.FRAME_240X240) # 设置帧大小为 QVGA (320x240)
+
# sensor.set_windowing((224, 224)) # 设置摄像头的窗口大小
sensor.set_auto_gain(False)
sensor.set_auto_whitebal(False)
diff --git a/002_B_Car/红绿灯len.py b/002_B_Car/红绿灯len.py
index 8a9091a..6905330 100644
--- a/002_B_Car/红绿灯len.py
+++ b/002_B_Car/红绿灯len.py
@@ -11,9 +11,10 @@ sensor.set_auto_gain(0,0)
sensor.skip_frames(time = 200) # 等待设置生效
# 阈值设置
-thresholds = [(35, 100, 6, 127, 0, 127), # 红色阈值
- (25, 100, -6, 127, 5, 43), # 黄色阈值
- (87, 100, -59, 127, -10, 127)] # 绿色阈值
+thresholds = [(123, 143, 161, 181, 136, 156), # 红色阈值
+ (207, 227, 126, 146, 155, 175), # 黄色阈值
+ (196, 216, 103, 123, 135, 155)] # 绿色阈值
+
# while True:
img = sensor.snapshot()
diff --git a/002_B_Car/红绿灯len2.py b/002_B_Car/红绿灯len2.py
index 25483e6..d875b50 100644
--- a/002_B_Car/红绿灯len2.py
+++ b/002_B_Car/红绿灯len2.py
@@ -34,9 +34,9 @@ Servo(15)
# (25, 100, -6, 127, 5, 43, "黄色"), # 黄色阈值
# (52, 100, -128, -5, 5, 127, "绿色")] # 绿色阈值
-thresholds = [(51, 62, 33, 101, 14, 127, "红色"), # 红色阈值
- (39, 98, -17, 21, 17, 57, "黄色"), # 黄色阈值
- (44, 90, -61, -13, -3, 30, "绿色")] # 绿色阈值
+thresholds = [(43, 68, 7, 74, 1, 46, "红色"), # 红色阈值
+ (48, 100, -23, 19, 4, 55, "黄色"), # 黄色阈值
+ (40, 93, -57, -7, -8, 29, "绿色")] # 绿色阈值
while True:
img = sensor.snapshot()
@@ -45,9 +45,9 @@ while True:
max_blob_color = ""
max_blob_color_index = 0
- # 找圆
- circles = img.find_circles(threshold=3500, x_margin=10, y_margin=10, r_margin=10,
- r_min=2, r_max=100, r_step=2)
+ # # 找圆
+ # circles = img.find_circles(threshold=3500, x_margin=10, y_margin=10, r_margin=10,
+ # r_min=2, r_max=100, r_step=2)
for index, threshold in enumerate(thresholds):
# 查找每种颜色的色块
@@ -60,15 +60,17 @@ while True:
max_blob_size = blob.pixels()
max_blob_color = threshold[-1] # 颜色标签
max_blob_color_index = index
- print(circles)
+ # print(circles)
# 画圆
- for c in circles:
- area = (c.x() - c.r(), c.y() - c.r(), 2 * c.r(), 2 * c.r())
- img.draw_rectangle(area, color=(255, 255, 255))
+ # for c in circles:
+ # area = (c.x() - c.r(), c.y() - c.r(), 2 * c.r(), 2 * c.r())
+ # img.draw_rectangle(area, color=(255, 255, 255))
# 绘制最大色块的矩形和中心十字，并输出颜色
if max_blob:
img.draw_rectangle(max_blob[0:4])
img.draw_cross(max_blob[5], max_blob[6])
print("最大色块的颜色是:", max_blob_color, max_blob_color_index)
+ else:
+ print("没有找到色块")
lcd.display(img)
diff --git a/002_B_Car/获取图片.py b/002_B_Car/获取图片.py
index 4d95026..192aab1 100644
--- a/002_B_Car/获取图片.py
+++ b/002_B_Car/获取图片.py
@@ -22,22 +22,22 @@ def Servo(angle):
S1 = PWM(tim_pwm, freq=50, duty=0, pin=17)
S1.duty((angle+90)/180*10+2.5)
-Servo(15)
+# Servo(15)
# 摄像头初始化
sensor.reset()
sensor.set_pixformat(sensor.RGB565)
sensor.set_framesize(sensor.QVGA)
+sensor.skip_frames(40)
sensor.set_vflip(1) # 摄像头后置方式
sensor.set_pixformat(sensor.RGB565) # 设置像素格式为彩色 RGB565
#sensor.set_pixformat(sensor.GRAYSCALE) # 设置像素格式为灰色
-sensor.set_framesize(sensor.QVGA) # 设置帧大小为 QVGA (320x240)
sensor.set_auto_gain(False)
sensor.set_auto_whitebal(False)
sensor.set_auto_gain(0,0)
-lcd.init() # LCD初始化
+# lcd.init() # LCD初始化
# 初始化 BOOT 键
boot_pin = 16 # 根据您的板子修改对应的引脚号
@@ -51,35 +51,7 @@ taking_pictures = False
count = 0
while(True):
- img = sensor.snapshot()
-
- # 检查 BOOT 键是否被按下
- if key.value() == 0:
- # 等待按键释放
- while key.value() == 0:
- time.sleep_ms(20)
- # 改变拍照状态
- taking_pictures = not taking_pictures
- if taking_pictures:
- print("Start taking pictures")
- else:
- print("Stop taking pictures")
-
- # 根据拍照状态决定是否保存图片
- if taking_pictures:
- img = img.resize(224, 224)
- image_save = "/sd/img/example{}.jpg".format(count)
- img.save(image_save)
- print("Image saved:", image_save)
- count += 1
- a = lcd.display(img)
-
- else:
- a = lcd.display(img)
-
- # 打印FPS
- print(clock.fps())
-
- # 稍作延迟，避免过快的循环
- time.sleep_ms(20)
-
+ sensor.snapshot().save("/sd/imgs/{}.jpg".format(count))
+ count+=1
+ print(count)
+ time.sleep_ms(100)
diff --git a/002_B_Car/获取颜色阈值LAB.py b/002_B_Car/获取颜色阈值LAB.py
index 220d160..38e24b5 100644
--- a/002_B_Car/获取颜色阈值LAB.py
+++ b/002_B_Car/获取颜色阈值LAB.py
@@ -6,45 +6,68 @@ import cv2
import numpy as np
# 读取图像
-image_path = 'xdl/1.png' # 请替换为您图像的实际路径
+image_path =r"E:\imgs\4.jpg"
image_bgr = cv2.imread(image_path)
+# 全局变量
+roi = []
+drawing = False
+
+def draw_rectangle(event, x, y, flags, param):
+ global roi, drawing, image_bgr_copy
+
+ if event == cv2.EVENT_LBUTTONDOWN:
+ roi = [(x, y)]
+ drawing = True
+
+ elif event == cv2.EVENT_MOUSEMOVE:
+ if drawing:
+ image_bgr_copy = image_bgr.copy()
+ cv2.rectangle(image_bgr_copy, roi[0], (x, y), (0, 0, 255), 2)
+ cv2.imshow('Select ROI', image_bgr_copy)
+
+ elif event == cv2.EVENT_LBUTTONUP:
+ roi.append((x, y))
+ drawing = False
+ cv2.rectangle(image_bgr, roi[0], roi[1], (0, 0, 255), 2)
+ cv2.imshow('Select ROI', image_bgr)
+
# 检查图像是否成功加载
if image_bgr is not None:
- # 将图像从 BGR 颜色空间转换到 LAB 颜色空间
- image_lab = cv2.cvtColor(image_bgr, cv2.COLOR_BGR2LAB)
+ image_bgr_copy = image_bgr.copy()
+ cv2.imshow('Select ROI', image_bgr_copy)
+ cv2.setMouseCallback('Select ROI', draw_rectangle)
- # 定义红框区域的坐标 (x, y, width, height)
- # 请用实际的红框坐标替换
- x, y, w, h = 280, 340, 60, 60 # 替换为红框的实际坐标
+ while True:
+ key = cv2.waitKey(1) & 0xFF
+ if key == ord('q'):
+ break
- # 裁剪 LAB 图像以获取红框内的区域
- cropped_lab = image_lab[y:y+h, x:x+w]
-
- # 计算裁剪区域的平均 LAB 值
- avg_lab = np.mean(cropped_lab.reshape(-1, 3), axis=0)
-
- # 定义阈值范围
- threshold_range = 10
- lower_red_threshold = np.clip(avg_lab - threshold_range, 0, 255)
- upper_red_threshold = np.clip(avg_lab + threshold_range, 0, 255)
-
- # 输出红色 LAB 颜色阈值
- lower = lower_red_threshold.astype(int)
- upper = upper_red_threshold.astype(int)
- print('红色的 LAB 颜色阈值下界:', lower)
- print('红色的 LAB 颜色阈值上界:', upper)
- print("{}, {}, {}, {}, {}, {}".format(lower[0], upper[0], lower[1], upper[1], lower[2], upper[2]))
-
- # 在原图像上绘制红框
- cv2.rectangle(image_bgr, (x, y), (x + w, y + h), (0, 0, 255), 2)
-
- # 显示图像
- cv2.imshow('Image with Red Rectangle', image_bgr)
- cv2.waitKey(0)
cv2.destroyAllWindows()
- # 保存处理后的图像
- cv2.imwrite("image_with_red_rectangle.jpg", image_bgr)
+ if len(roi) == 2:
+ x1, y1 = roi[0]
+ x2, y2 = roi[1]
+ x, y, w, h = min(x1, x2), min(y1, y2), abs(x2 - x1), abs(y2 - y1)
+
+ # 转换到 LAB 颜色空间
+ image_lab = cv2.cvtColor(image_bgr, cv2.COLOR_BGR2LAB)
+ cropped_lab = image_lab[y:y+h, x:x+w]
+
+ # 计算裁剪区域的 LAB 颜色阈值下界和上界
+ lower_red_threshold = np.min(cropped_lab.reshape(-1, 3), axis=0)
+ upper_red_threshold = np.max(cropped_lab.reshape(-1, 3), axis=0)
+
+ # 输出红色 LAB 颜色阈值
+ lower = lower_red_threshold.astype(int)
+ upper = upper_red_threshold.astype(int)
+ print('红色的 LAB 颜色阈值下界:', lower)
+ print('红色的 LAB 颜色阈值上界:', upper)
+ print("{}, {}, {}, {}, {}, {}".format(lower[0], upper[0], lower[1], upper[1], lower[2], upper[2]))
+
+ # 保存处理后的图像
+ cv2.imwrite("image_with_red_rectangle.jpg", image_bgr)
+ else:
+ print("未正确标注区域。")
else:
print("图像无法加载。")
diff --git a/002_B_Car/颜色识别.py b/002_B_Car/颜色识别.py
index d8ea7c1..52a776e 100644
--- a/002_B_Car/颜色识别.py
+++ b/002_B_Car/颜色识别.py
@@ -28,9 +28,9 @@ clock=time.clock()
# thresholds = [(59, 100, 40, 127, 5, 127), # 红色阈值
# (87, 100, -59, 127, -10, 127), # 绿色阈值
# (0, 30, 0, 64, -128, -20)] # 蓝色阈值
-thresholds = [(35, 100, 6, 127, 0, 127), # 红色阈值
- (25, 100, -6, 127, 5, 43), # 黄色阈值
- (87, 100, -59, 127, -10, 127)] # 绿色阈值
+thresholds = [(123, 143, 161, 181, 136, 156), # 红色阈值
+ (207, 227, 126, 146, 155, 175), # 黄色阈值
+ (196, 216, 103, 123, 135, 155)] # 绿色阈值
while True:
diff --git a/img2bin.py b/img2bin.py
index 014f775..bff8853 100644
--- a/img2bin.py
+++ b/img2bin.py
@@ -1,102 +1,91 @@
-# -*- coding:utf-8 -*-
-# @Author len
-# @Create 2023/10/26 9:18
-
from PIL import Image
import numpy as np
import os
-def convert_jpg_to_bin(input_file, output_file, max_width, max_height):
- # 打开JPEG图像
- image = Image.open(input_file)
+def convert_jpg_to_bin(input_file, output_file):
print(f"Processing: {input_file}")
+ # 打开已经调整尺寸并转换为 JPEG 的图像
+ image = Image.open(input_file).convert("RGB")
- # 限制图像尺寸
- image.thumbnail((max_width, max_height))
+ # 转换为 NumPy 数组
+ img_array = np.array(image, dtype=np.uint16)
+ r = img_array[:, :, 0]
+ g = img_array[:, :, 1]
+ b = img_array[:, :, 2]
- # 将图像转换为RGB模式
- image = image.convert("RGB")
+ # 计算 RGB565 格式的值
+ rgb565 = ((r & 0xF8) << 8) | ((g & 0xFC) << 3) | (b >> 3)
- # 创建一个空白的16位彩色图像
- result_image = np.zeros((image.size[1], image.size[0]), dtype=np.uint16)
+ # 展平数组并保存为 BIN 文件
+ rgb565.flatten().tofile(output_file)
- # 遍历图像的每个像素
- for y in range(image.size[1]):
- for x in range(image.size[0]):
- # 获取RGB值
- r, g, b = image.getpixel((x, y))
-
- # 将RGB值转换为RGB565格式
- pixel_value = ((r & 0b11111000) << 8) | ((g & 0b11111100) << 3) | (b >> 3)
-
- # 将RGB565值存储到16位彩色图像中
- result_image[y, x] = pixel_value
-
- # 将三维数组展平为一维数组
- flattened_array = result_image.flatten()
-
- # 保存为BIN文件
- with open(output_file, "wb") as file:
- file.write(flattened_array.tobytes())
-
-def ensure_jpg_format(input_path, temp_dir):
- # 确保文件是JPEG格式，如果不是则转换为JPEG
+def prepare_image(input_path, temp_dir, target_width, target_height):
+ """
+ 先调整图片尺寸，保持原始比例，然后将缩放后的图片粘贴到固定尺寸的画布上，
+ 最终保存为 JPEG 格式，确保输出尺寸为 target_width x target_height。
+ """
with Image.open(input_path) as img:
- if img.format != "JPEG":
- base_name = os.path.splitext(os.path.basename(input_path))[0]
- jpg_path = os.path.join(temp_dir, f"{base_name}.jpg")
- img = img.convert("RGB") # 转换为RGB模式
- img.save(jpg_path, "JPEG")
- print(f"Converted to JPEG: {jpg_path}")
- return jpg_path
- return input_path
+ img = img.convert("RGB")
+ # 按比例缩放，使图片适应目标尺寸，但不会超出目标尺寸
+ img.thumbnail((target_width, target_height))
+
+ # 创建一个新的目标尺寸图像，背景填充黑色（可以根据需要修改背景色）
+ new_img = Image.new("RGB", (target_width, target_height), (0, 0, 0))
+ # 计算居中粘贴的位置
+ left = (target_width - img.width) // 2
+ top = (target_height - img.height) // 2
+ new_img.paste(img, (left, top))
+
+ base_name = os.path.splitext(os.path.basename(input_path))[0]
+ jpg_path = os.path.join(temp_dir, f"{base_name}.jpg")
+ new_img.save(jpg_path, "JPEG")
+ print(f"Converted, resized and padded to JPEG: {jpg_path}")
+ return jpg_path
def get_sort_key(filename):
- # 提取文件名前两位数字用于排序
base_name = os.path.splitext(filename)[0]
try:
- # 尝试将前两位数字转换为整数
return int(base_name[:2])
except ValueError:
- # 如果无法转换为数字，则返回一个较大的默认值
return float('inf')
-# 使用示例
-input_file = r"C:\Users\10561\Desktop\002\任务2" # 图片地址
-temp_dir = os.path.join(input_file, "temp") # 临时目录
-output_dir = os.path.join(input_file, "bin") # 输出目录
+def main():
+ input_dir = r"C:\Users\10561\Desktop\20250221\tft图片" # 图片目录
+ temp_dir = os.path.join(input_dir, "temp") # 临时目录
+ output_dir = os.path.join(input_dir, "bin") # 输出目录
-os.makedirs(temp_dir, exist_ok=True) # 创建临时目录
-os.makedirs(output_dir, exist_ok=True) # 创建输出目录
+ os.makedirs(temp_dir, exist_ok=True)
+ os.makedirs(output_dir, exist_ok=True)
-max_width = 800
-max_height = 480
+ target_width, target_height = 800, 480
-# 获取文件列表并忽略目录，同时按照前两位数字排序
-files = sorted(
- [file for file in os.listdir(input_file) if os.path.isfile(os.path.join(input_file, file))],
- key=get_sort_key
-)
+ # 获取所有文件，并按照文件名前两位数字排序
+ files = sorted(
+ [f for f in os.listdir(input_dir) if os.path.isfile(os.path.join(input_dir, f))],
+ key=get_sort_key
+ )
-for file in files:
- print(file)
- if file in ["System Volume Information"] or ".bin" in file:
- continue
+ for file in files:
+ if file in ["System Volume Information"] or ".bin" in file:
+ continue
- portion = os.path.splitext(file)
- input001 = os.path.join(input_file, file)
+ file_base = os.path.splitext(file)[0]
+ input_path = os.path.join(input_dir, file)
- # 确保输入文件是JPEG格式
- jpg_file = ensure_jpg_format(input001, temp_dir)
+ # 调整尺寸、填充背景并转换为 JPEG
+ jpg_file = prepare_image(input_path, temp_dir, target_width, target_height)
- # 生成输出文件路径
- out001 = os.path.join(output_dir, f"{portion[0]}.bin")
- print(f"Output BIN: {out001}")
+ # 生成输出 BIN 文件路径
+ output_file = os.path.join(output_dir, f"{file_base}.bin")
+ print(f"Output BIN: {output_file}")
- # 转换为BIN文件
- convert_jpg_to_bin(jpg_file, out001, max_width, max_height)
+ # 转换为 BIN 文件
+ convert_jpg_to_bin(jpg_file, output_file)
-# 清理临时目录
-for temp_file in os.listdir(temp_dir):
- os.remove(os.path.join(temp_dir, temp_file))
-os.rmdir(temp_dir)
+ # 如有需要，可清理临时目录
+ # for temp_file in os.listdir(temp_dir):
+ # os.remove(os.path.join(temp_dir, temp_file))
+ # os.rmdir(temp_dir)
+
+if __name__ == '__main__':
+ main()
diff --git a/txt2xml.py b/txt2xml.py
new file mode 100644
index 0000000..a735ea2
--- /dev/null
+++ b/txt2xml.py
@@ -0,0 +1,154 @@
+import os
+import xml.etree.ElementTree as ET
+from xml.dom import minidom
+from PIL import Image
+
+def convert_txt_to_xml(txt_path, image_path, xml_path):
+ # 打开图像，获取尺寸及通道数
+ with Image.open(image_path) as img:
+ img_width, img_height = img.size
+ mode = img.mode
+ if mode == "RGB":
+ depth = 3
+ elif mode == "L":
+ depth = 1
+ else:
+ depth = 3 # 默认3
+
+ # 读取TXT文件中的标注信息
+ with open(txt_path, "r") as f:
+ lines = f.readlines()
+
+ # 类别编号到标签名称的映射
+ id_to_label = {
+ 0: "light",
+ 1: "red",
+ 2: "yellow",
+ 3: "green"
+ }
+
+ # 构建XML结构
+ annotation = ET.Element("annotation", verified="yes")
+
+ folder_elem = ET.SubElement(annotation, "folder")
+ folder_elem.text = os.path.basename(os.path.dirname(image_path))
+
+ filename_elem = ET.SubElement(annotation, "filename")
+ image_filename = os.path.basename(image_path)
+ filename_elem.text = image_filename
+
+ path_elem = ET.SubElement(annotation, "path")
+ path_elem.text = os.path.abspath(image_path)
+
+ source_elem = ET.SubElement(annotation, "source")
+ database_elem = ET.SubElement(source_elem, "database")
+ database_elem.text = "Unknown"
+
+ size_elem = ET.SubElement(annotation, "size")
+ width_elem = ET.SubElement(size_elem, "width")
+ width_elem.text = str(img_width)
+ height_elem = ET.SubElement(size_elem, "height")
+ height_elem.text = str(img_height)
+ depth_elem = ET.SubElement(size_elem, "depth")
+ depth_elem.text = str(depth)
+
+ segmented_elem = ET.SubElement(annotation, "segmented")
+ segmented_elem.text = "0"
+
+ # 逐行处理TXT标注
+ for line in lines:
+ parts = line.strip().split()
+ if len(parts) != 5:
+ print(f"跳过格式不正确的行: {line.strip()}")
+ continue
+ class_id_str, x_center_norm_str, y_center_norm_str, width_norm_str, height_norm_str = parts
+ try:
+ class_id = int(class_id_str)
+ x_center_norm = float(x_center_norm_str)
+ y_center_norm = float(y_center_norm_str)
+ width_norm = float(width_norm_str)
+ height_norm = float(height_norm_str)
+ except Exception as e:
+ print(f"转换数据出错 {line.strip()}: {e}")
+ continue
+
+ if class_id not in id_to_label:
+ print(f"未知标签编号 {class_id}，跳过该行")
+ continue
+
+ label = id_to_label[class_id]
+
+ # 计算绝对坐标
+ x_center = x_center_norm * img_width
+ y_center = y_center_norm * img_height
+ bbox_width = width_norm * img_width
+ bbox_height = height_norm * img_height
+
+ xmin = int(x_center - bbox_width / 2)
+ ymin = int(y_center - bbox_height / 2)
+ xmax = int(x_center + bbox_width / 2)
+ ymax = int(y_center + bbox_height / 2)
+
+ # 构建object节点
+ object_elem = ET.SubElement(annotation, "object")
+ name_elem = ET.SubElement(object_elem, "name")
+ name_elem.text = label
+ pose_elem = ET.SubElement(object_elem, "pose")
+ pose_elem.text = "Unspecified"
+ truncated_elem = ET.SubElement(object_elem, "truncated")
+ truncated_elem.text = "0"
+ difficult_elem = ET.SubElement(object_elem, "difficult")
+ difficult_elem.text = "0"
+
+ bndbox_elem = ET.SubElement(object_elem, "bndbox")
+ xmin_elem = ET.SubElement(bndbox_elem, "xmin")
+ xmin_elem.text = str(xmin)
+ ymin_elem = ET.SubElement(bndbox_elem, "ymin")
+ ymin_elem.text = str(ymin)
+ xmax_elem = ET.SubElement(bndbox_elem, "xmax")
+ xmax_elem.text = str(xmax)
+ ymax_elem = ET.SubElement(bndbox_elem, "ymax")
+ ymax_elem.text = str(ymax)
+
+ # 美化输出的XML字符串
+ xml_str = ET.tostring(annotation, encoding="utf-8")
+ parsed_str = minidom.parseString(xml_str)
+ pretty_xml_str = parsed_str.toprettyxml(indent=" ")
+
+ with open(xml_path, "w", encoding="utf-8") as f:
+ f.write(pretty_xml_str)
+
+def batch_convert_txt_to_xml(txt_dir, image_dir, output_dir, image_ext=".jpg"):
+ """
+ 批量将TXT转换为XML：
+ txt_dir: TXT文件所在目录
+ image_dir: 对应图像所在目录（图片文件名与TXT同名，后缀可指定）
+ output_dir: XML文件保存目录
+ image_ext: 图像文件扩展名（默认 .jpg）
+ """
+ if not os.path.exists(output_dir):
+ os.makedirs(output_dir)
+
+ for file in os.listdir(txt_dir):
+ if file.lower().endswith(".txt"):
+ txt_path = os.path.join(txt_dir, file)
+ base_name = os.path.splitext(file)[0]
+ image_filename = base_name + image_ext
+ image_path = os.path.join(image_dir, image_filename)
+ if not os.path.exists(image_path):
+ print(f"图像文件不存在: {image_path}，跳过 {txt_path}")
+ continue
+ xml_filename = base_name + ".xml"
+ xml_path = os.path.join(output_dir, xml_filename)
+ try:
+ convert_txt_to_xml(txt_path, image_path, xml_path)
+ print(f"成功转换: {txt_path} --> {xml_path}")
+ except Exception as e:
+ print(f"处理 {txt_path} 时出错: {e}")
+
+if __name__ == "__main__":
+ # 请修改以下目录为实际路径
+ txt_directory = "path_to_txt_directory" # TXT文件所在目录
+ image_directory = "path_to_image_directory" # 图像文件所在目录
+ output_directory = "path_to_output_xml_directory" # 输出XML文件保存目录
+ batch_convert_txt_to_xml(txt_directory, image_directory, output_directory, image_ext=".jpg")
diff --git a/xml2txt.py b/xml2txt.py
new file mode 100644
index 0000000..58e8f68
--- /dev/null
+++ b/xml2txt.py
@@ -0,0 +1,82 @@
+import os
+import xml.etree.ElementTree as ET
+
+def convert_xml_to_txt(xml_path, txt_path):
+ # 解析XML文件
+ tree = ET.parse(xml_path)
+ root = tree.getroot()
+
+ # 获取图像尺寸信息
+ size = root.find("size")
+ img_width = int(size.find("width").text)
+ img_height = int(size.find("height").text)
+
+ # 标签映射字典
+ label_mapping = {
+ "light": 0,
+ "red": 1,
+ "yellow": 2,
+ "green": 3
+ }
+
+ lines = []
+ # 遍历每个object节点
+ for obj in root.findall("object"):
+ # 获取并转换标签名称，统一转为小写
+ class_name = obj.find("name").text.strip().lower()
+ if class_name in label_mapping:
+ class_id = label_mapping[class_name]
+ else:
+ print(f"未知标签 {class_name}，文件 {xml_path} 中跳过该对象")
+ continue
+
+ # 获取边界框坐标
+ bndbox = obj.find("bndbox")
+ xmin = float(bndbox.find("xmin").text)
+ ymin = float(bndbox.find("ymin").text)
+ xmax = float(bndbox.find("xmax").text)
+ ymax = float(bndbox.find("ymax").text)
+
+ # 计算边界框中心点和宽高
+ x_center = (xmin + xmax) / 2.0
+ y_center = (ymin + ymax) / 2.0
+ bbox_width = xmax - xmin
+ bbox_height = ymax - ymin
+
+ # 归一化
+ x_center_norm = x_center / img_width
+ y_center_norm = y_center / img_height
+ width_norm = bbox_width / img_width
+ height_norm = bbox_height / img_height
+
+ # 每行格式：类别编号 x_center y_center width height
+ line = f"{class_id} {x_center_norm:.6f} {y_center_norm:.6f} {width_norm:.6f} {height_norm:.6f}"
+ lines.append(line)
+
+ # 将转换结果写入TXT文件
+ with open(txt_path, "w") as f:
+ for line in lines:
+ f.write(line + "\n")
+
+def batch_convert_xml_to_txt(input_dir, output_dir):
+ # 如果输出目录不存在则创建
+ if not os.path.exists(output_dir):
+ os.makedirs(output_dir)
+
+ # 遍历输入目录下所有XML文件
+ for file in os.listdir(input_dir):
+ if file.lower().endswith(".xml"):
+ xml_path = os.path.join(input_dir, file)
+ txt_file = os.path.splitext(file)[0] + ".txt"
+ txt_path = os.path.join(output_dir, txt_file)
+ try:
+ convert_xml_to_txt(xml_path, txt_path)
+ print(f"成功转换: {xml_path} --> {txt_path}")
+ except Exception as e:
+ print(f"处理 {xml_path} 时出错: {e}")
+
+if __name__ == "__main__":
+ # 请替换下面的目录为实际路径
+ input_directory = r"C:\Users\10561\Downloads\Compressed\xml" # XML文件所在的目录
+ output_directory = r"C:\Users\10561\Downloads\Compressed\txt" # 转换后txt文件保存的目录
+ batch_convert_xml_to_txt(input_directory, output_directory)
diff --git a/xml删除标签.py b/xml删除标签.py
new file mode 100644
index 0000000..eb0c289
--- /dev/null
+++ b/xml删除标签.py
@@ -0,0 +1,51 @@
+import os
+import xml.etree.ElementTree as ET
+
+def remove_tag_from_element(element, tag):
+ """
+ 递归遍历element的所有子节点，删除所有匹配指定标签的子节点。
+ """
+ for child in list(element): # 用 list() 避免在遍历时修改子节点列表导致问题
+ if child.tag == tag:
+ element.remove(child)
+ else:
+ remove_tag_from_element(child, tag)
+
+def remove_tag_from_xml(xml_file, tag_to_remove, output_file):
+ """
+ 解析单个 XML 文件，删除其中所有指定的标签，然后保存为新的 XML 文件。
+ """
+ try:
+ tree = ET.parse(xml_file)
+ root = tree.getroot()
+ # 如果根节点本身就是要删除的标签，则提示并跳过处理
+ if root.tag == tag_to_remove:
+ print(f"警告：{xml_file} 的根标签就是要删除的标签({tag_to_remove})，跳过该文件。")
+ return
+
+ remove_tag_from_element(root, tag_to_remove)
+ tree.write(output_file, encoding="utf-8", xml_declaration=True)
+ print(f"处理完成: {xml_file} -> {output_file}")
+ except Exception as e:
+ print(f"处理 {xml_file} 时出错: {e}")
+
+def batch_remove_tag(input_dir, output_dir, tag_to_remove):
+ """
+ 批量处理：遍历 input_dir 下所有 XML 文件，删除指定标签后，将新文件保存到 output_dir。
+ """
+ if not os.path.exists(output_dir):
+ os.makedirs(output_dir)
+
+ for file in os.listdir(input_dir):
+ if file.lower().endswith(".xml"):
+ input_path = os.path.join(input_dir, file)
+ output_path = os.path.join(output_dir, file)
+ remove_tag_from_xml(input_path, tag_to_remove, output_path)
+
+if __name__ == "__main__":
+ # 请修改以下路径和要删除的标签名称
+ input_directory = r"C:\Users\10561\Downloads\Compressed\xml" # 输入 XML 文件所在目录
+ output_directory = "path_to_output_xml_directory" # 处理后 XML 文件保存目录
+ tag_to_delete = "segmented" # 例如：删除 标签，可改为你需要删除的标签名
+
+ batch_remove_tag(input_directory, output_directory, tag_to_delete)
diff --git a/图片分批.py b/图片分批.py
new file mode 100644
index 0000000..c0c1b5c
--- /dev/null
+++ b/图片分批.py
@@ -0,0 +1,20 @@
+import os
+
+
+if __name__ == '__main__':
+ dir_path=r"C:\Users\10561\Desktop\frames-3000"
+ output_dir=r"C:\Users\10561\Desktop\frames-3000-批次"
+ if not os.path.exists(output_dir):
+ os.mkdir(output_dir)
+ images=os.listdir(dir_path)
+ images=[os.path.join(dir_path,i) for i in images]
+ for i in range(len(images)//100):
+ out=os.path.join(output_dir,f"{i}")
+ os.mkdir(out)
+ if i == len(images)//100:
+ #复制
+ for j in range(i*100,len(images)):
+ os.system(f"copy {images[j]} {out}")
+ else:
+ for j in range(i*100,(i+1)*100):
+ os.system(f"copy {images[j]} {out}")
diff --git a/图片尺寸检查.py b/图片尺寸检查.py
new file mode 100644
index 0000000..bc86171
--- /dev/null
+++ b/图片尺寸检查.py
@@ -0,0 +1,21 @@
+import os
+from PIL import Image
+
+# 指定要检查的目录路径
+directory = r"C:\Users\10561\Downloads\Compressed\images" # 替换为你的目录路径
+
+# 目标尺寸
+target_size = (224, 224)
+
+# 遍历目录中的所有文件
+for filename in os.listdir(directory):
+ # 只处理图片文件（根据文件后缀判断）
+ if filename.lower().endswith(('.png', '.jpg', '.jpeg', '.bmp', '.gif')):
+ file_path = os.path.join(directory, filename)
+ try:
+ with Image.open(file_path) as img:
+ # 检查图片尺寸是否与目标尺寸不一致
+ if img.size != target_size:
+ print(f"{filename}: {img.size}")
+ except Exception as e:
+ print(f"处理 {filename} 时出错: {e}")
diff --git a/抽帧.py b/抽帧.py
index 5aa01b9..32473a6 100644
--- a/抽帧.py
+++ b/抽帧.py
@@ -38,8 +38,8 @@ def extract_frames(video_path, output_dir, interval=0.5,output_name='frame'):
cap.release()
print(f"共保存了 {saved_count} 帧.")
-video_dir=r'C:\Users\10561\Desktop\bcar红绿灯\1'
-output_directory = r"C:\Users\10561\Desktop\frames" # 替换为你希望保存的目录路径
+video_dir=r'C:\Users\10561\Desktop\bcar红绿灯\3'
+output_directory = r"C:\Users\10561\Desktop\frames123" # 替换为你希望保存的目录路径
for file in os.listdir(video_dir):
if file.endswith('.mp4'):