Yolov8 照片偵測後如何抓物體座標位置的方法

woff

1. 法一

1. import os
   
2. import numpy as np
   
3. from ultralytics import YOLO
   
4. os.environ["KMP_DUPLICATE_LIB_OK"]  =  "TRUE"
   
5. 
   
6. if __name__ == '__main__':
   
7.     model = YOLO("best.pt")
   
8. 
   
9.     results = model("bus.jpg")
   
10.     boxes = results[0].boxes
    
11.     arrxy = results[0].boxes.xyxy.cpu() #numpy只能用CPU運算
    
12.     coordinates = np.array(arrxy)
    
13.     x_coords = (coordinates[:, 0] + coordinates[:, 2]) / 2
    
14.     y_coords = (coordinates[:, 1] + coordinates[:, 3]) / 2
    
15.     midpoints = np.column_stack((x_coords, y_coords))
    
16.     rounded_n_sorted_arr = np.round(midpoints[midpoints[:, 1].argsort()]).astype(int)
    
17.     print(rounded_n_sorted_arr[:]) #列出圖片預測物體所有座標
    
18. 
    
19.     for box in boxes: #一項一項列出圖片預測物體座標
    
20.         print(box.cls.item()) # class id
    
21.         print(box.xyxy)  # box coordinates (tensor)
    
22.         print("準確率",box.conf.item()) # confidence value)
    
23.         coordinates = np.array(box.xyxy.cpu()) #numpy只能用CPU運算
    
24.         x_coords = (coordinates[:, 0] + coordinates[:, 2]) / 2
    
25.         y_coords = (coordinates[:, 1] + coordinates[:, 3]) / 2
    
26.         midpoints = np.column_stack((x_coords, y_coords))
    
27.         rounded_n_sorted_arr = np.round(midpoints[midpoints[:, 1].argsort()]).astype(int)
    
28.         print("x,y:",rounded_n_sorted_arr[:])

複製代碼

結果圖：


座標：

1. [[ 16 290]
   
2. [412 491]
   
3. [740 626]
   
4. [283 631]
   
5. [146 651]
   
6. [ 32 710]]
   
7. 5.0
   
8. image 1/1 D:\yolo\bus.jpg: 640x480 4 persons, 1 bus, 1 stop sign, 83.9ms
   
9. Speed: 3.0ms preprocess, 83.9ms inference, 4.0ms postprocess per image at shape (1, 3, 640, 480)
   
10. tensor([[ 22.3412, 228.0822, 802.0841, 754.3939]], device='cuda:0')
    
11. 準確率 0.879738450050354
    
12. x,y: [[412 491]]
    
13. 0.0
    
14. tensor([[ 47.5999, 398.8344, 244.2552, 903.1386]], device='cuda:0')
    
15. 準確率 0.873720109462738
    
16. x,y: [[146 651]]
    
17. 0.0
    
18. tensor([[670.3670, 376.9174, 810.0000, 875.0829]], device='cuda:0')
    
19. 準確率 0.8693700432777405
    
20. x,y: [[740 626]]
    
21. 0.0
    
22. tensor([[220.5713, 405.7240, 344.5589, 857.2506]], device='cuda:0')
    
23. 準確率 0.819391667842865
    
24. x,y: [[283 631]]
    
25. 11.0
    
26. tensor([[7.7698e-02, 2.5441e+02, 3.2119e+01, 3.2465e+02]], device='cuda:0')
    
27. 準確率 0.44594067335128784
    
28. x,y: [[ 16 290]]
    
29. 0.0
    
30. tensor([[3.2650e-02, 5.4988e+02, 6.4001e+01, 8.6930e+02]], device='cuda:0')
    
31. 準確率 0.29976797103881836
    
32. x,y: [[ 32 710]]
    

複製代碼

法二、程式比法一簡單(速度慢數幾ms)程式碼

1. from ultralytics import YOLO
   
2. model = YOLO("best.pt")
   
3. result = model.predict(
   
4.     source="bus.jpg",
   
5.     mode="predict",
   
6.     save=True
   
7. )
   
8. 
   
9. print('模型可偵測項目的名稱與編號')
   
10. print(model.names)
    
11. print('模型偵測結果')
    
12. boxes = result[0].boxes.cpu().numpy()
    
13. for box in boxes:
    
14.         print("分類編號：", box.cls)
    
15.         print("座標位置：", box.xyxy)
    
16.         print("機率：", box.conf)

複製代碼

1. image 1/1 D:\yolo\bus.jpg: 640x480 4 persons, 1 bus, 1 stop sign, 86.8ms
   
2. Speed: 3.0ms preprocess, 86.8ms inference, 5.0ms postprocess per image at shape (1, 3, 640, 480)
   
3. Results saved to runs\detect\predict21
   
4. 模型可偵測項目的名稱與編號
   
5. {0: 'person', 1: 'bicycle', 2: 'car', 3: 'motorcycle', 4: 'airplane', 5: 'bus', 6: 'train', 7: 'truck', 8: 'boat', 9: 'traffic light', 10: 'fire hydrant', 11: 'stop sign', 12: 'parking meter', 13: 'bench', 14: 'bird', 15: 'cat', 16: 'dog', 17: 'horse', 18: 'sheep', 19: 'cow', 20: 'elephant', 21: 'bear', 22: 'zebra', 23: 'giraffe', 24: 'backpack', 25: 'umbrella', 26: 'handbag', 27: 'tie', 28: 'suitcase', 29: 'frisbee', 30: 'skis', 31: 'snowboard', 32: 'sports ball', 33: 'kite', 34: 'baseball bat', 35: 'baseball glove', 36: 'skateboard', 37: 'surfboard', 38: 'tennis racket', 39: 'bottle', 40: 'wine glass', 41: 'cup', 42: 'fork', 43: 'knife', 44: 'spoon', 45: 'bowl', 46: 'banana', 47: 'apple', 48: 'sandwich', 49: 'orange', 50: 'broccoli', 51: 'carrot', 52: 'hot dog', 53: 'pizza', 54: 'donut', 55: 'cake', 56: 'chair', 57: 'couch', 58: 'potted plant', 59: 'bed', 60: 'dining table', 61: 'toilet', 62: 'tv', 63: 'laptop', 64: 'mouse', 65: 'remote', 66: 'keyboard', 67: 'cell phone', 68: 'microwave', 69: 'oven', 70: 'toaster', 71: 'sink', 72: 'refrigerator', 73: 'book', 74: 'clock', 75: 'vase', 76: 'scissors', 77: 'teddy bear', 78: 'hair drier', 79: 'toothbrush'}
   
6. 模型偵測結果
   
7. 分類編號： [          5]
   
8. 座標位置： [[     22.341      228.08      802.08      754.39]]
   
9. 機率： [    0.87974]
   
10. 分類編號： [          0]
    
11. 座標位置： [[       47.6      398.83      244.26      903.14]]
    
12. 機率： [    0.87372]

複製代碼

文章出處： NetYea 網頁設計

參考文章
https://medium.com/@batuhansener ... yolov8-682dcdc0f451
https://stackoverflow.com/questi ... ect-detection-model