YOLO之boxes小记

python 复制代码

import cv2
from ultralytics import YOLO
# 加载模型
model = YOLO(model='yolov8n.pt')
results = model(source='animal.jpg')

result = results[0]
img = result.plot()
from matplotlib import pyplot as plt
# matplotlib :rgb模式
# cv:bgr模式
plt.imshow(X=img[:,:,::-1])

python 复制代码

result.boxes



```python
"""
# 类别
cls: tensor([22., 20., 23., 20., 20.])
# 置信度 也就是概率
conf: tensor([0.8806, 0.8598, 0.5866, 0.5551, 0.3515])
data: tensor([[2.4926e+01, 5.1402e+02, 3.7411e+02, 7.9772e+02, 8.8055e-01, 2.2000e+01],
        [3.2387e+02, 3.8592e+02, 7.2567e+02, 8.0409e+02, 8.5977e-01, 2.0000e+01],
        [5.9173e+02, 2.6204e+02, 8.0608e+02, 7.7721e+02, 5.8656e-01, 2.3000e+01],
        [7.1596e+02, 6.3707e+02, 9.0578e+02, 8.5622e+02, 5.5513e-01, 2.0000e+01],
        [7.2943e+02, 5.2940e+02, 1.1668e+03, 7.8796e+02, 3.5151e-01, 2.0000e+01]])
id: None
is_track: False
orig_shape: (897, 1200)
shape: torch.Size([5, 6])
# 中心点坐标 （x,y） ，w宽 h高  n是规范化以后的数据，所有跟x相关的除以宽度，所有与y相关的除以高度,相当于归一化了 （相对坐标是用的比较多的）
xywh: tensor([[199.5182, 655.8711, 349.1838, 283.6943],
        [524.7690, 595.0039, 401.7979, 418.1656],
        [698.9015, 519.6245, 214.3517, 515.1679],
        [810.8722, 746.6467, 189.8149, 219.1466],
        [948.1376, 658.6791, 437.4136, 258.5655]])
xywhn: tensor([[0.1663, 0.7312, 0.2910, 0.3163],
        [0.4373, 0.6633, 0.3348, 0.4662],
        [0.5824, 0.5793, 0.1786, 0.5743],
        [0.6757, 0.8324, 0.1582, 0.2443],
        [0.7901, 0.7343, 0.3645, 0.2883]])
# 这个是左上角坐标（x,y）和右下角坐标（x,y）        
xyxy: tensor([[  24.9263,  514.0239,  374.1101,  797.7183],
        [ 323.8700,  385.9211,  725.6679,  804.0867],
        [ 591.7256,  262.0406,  806.0774,  777.2085],
        [ 715.9648,  637.0734,  905.7797,  856.2200],
        [ 729.4308,  529.3964, 1166.8444,  787.9619]])
xyxyn: tensor([[0.0208, 0.5730, 0.3118, 0.8893],
        [0.2699, 0.4302, 0.6047, 0.8964],
        [0.4931, 0.2921, 0.6717, 0.8665],
        [0.5966, 0.7102, 0.7548, 0.9545],
        [0.6079, 0.5902, 0.9724, 0.8784]])
"""

复制代码