YOLOv8+DeepSort实现目标检测与追踪

YOLOv8+DeepSort实现目标检测与追踪 - 迷途小书童的Note迷途小书童的Note

本篇介绍一个使用 YOLOv8 目标检测算法和 DeepSort 目标跟踪算法的开源项目。

1. git clone https://github.com/MuhammadMoinFaisal/YOLOv8-DeepSORT-Object-Tracking.git
2. cd YOLOv8-DeepSORT-Object-Tracking
3. # 创建一个全新的虚拟环境
4. conda create -n py38 python=3.8
5. # 激活虚拟环境
6. conda activate py38
7. # 安装依赖
8. pip install -e .

https://drive.google.com/drive/folders/1kna8eWGrSfzaR6DtNJ8_GchGgPMv3VC8，将整个文件夹下载，下来的是一个 zip ，将 zip 解压后， 文件夹放在目录 ultralytics/yolo/v8/detect 下

准备一段视频检测素材，放到目录 ultralytics/yolo/v8/detect 下，然后执行命令

1. # 使用yolov8l.pt模型，首次执行，会自动取下载，测试视频是10s.mp4，show参数表示处理过程中显示处理结果图像
2. python predict.py model=yolov8l.pt source="10s.mp4" show=True

1. File "E:\Playground\YOLOv8-DeepSORT-Object-Tracking\ultralytics\yolo\v8\detect\deep_sort_pytorch\deep_sort\deep_sort.py", line 31, in update
2. detections = [Detection(bbox_tlwh[i], conf, features[i],oid) for i, (conf,oid) in enumerate(zip(confidences,oids)) if conf > self.min_confidence]
3. File "E:\Playground\YOLOv8-DeepSORT-Object-Tracking\ultralytics\yolo\v8\detect\deep_sort_pytorch\deep_sort\deep_sort.py", line 31, in <listcomp>
4. detections = [Detection(bbox_tlwh[i], conf, features[i],oid) for i, (conf,oid) in enumerate(zip(confidences,oids)) if conf > self.min_confidence]
5. File "E:\Playground\YOLOv8-DeepSORT-Object-Tracking\ultralytics\yolo\v8\detect\deep_sort_pytorch\deep_sort\sort\detection.py", line 30, in __init__
6. self.tlwh = np.asarray(tlwh, dtype=np.float)
7. File "D:\Tools\anaconda3\envs\py38\lib\site-packages\numpy\__init__.py", line 305, in __getattr__
8. raise AttributeError(__former_attrs__[attr])
9. AttributeError: module 'numpy' has no attribute 'float'.
10. `np.float` was a deprecated alias for the builtin `float`. To avoid this error in existing code, use `float` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.float64` here.
11. The aliases was originally deprecated in NumPy 1.20; for more details and guidance see the original release note at:
12. https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
13. Set the environment variable HYDRA_FULL_ERROR=1 for a complete stack trace.

这是由于 numpy 新版已经没有 float 属性了，降版本解决 pip install numpy==1.23.5

1. Traceback (most recent call last):
2. File "predict.py", line 253, in predict
3. predictor()
4. File "D:\Tools\anaconda3\envs\py38\lib\site-packages\torch\utils\_contextlib.py", line 115, in decorate_context
5. return func(*args, **kwargs)
6. File "e:\playground\yolov8-deepsort-object-tracking\ultralytics\yolo\engine\predictor.py", line 183, in __call__
7. preds = self.postprocess(preds, im, im0s)
8. File "predict.py", line 181, in postprocess
9. preds = ops.non_max_suppression(preds,
10. File "e:\playground\yolov8-deepsort-object-tracking\ultralytics\yolo\utils\ops.py", line 212, in non_max_suppression
11. i = torchvision.ops.nms(boxes, scores, iou_thres) # NMS
12. File "D:\Tools\anaconda3\envs\py38\lib\site-packages\torchvision\ops\boxes.py", line 41, in nms
13. return torch.ops.torchvision.nms(boxes, scores, iou_threshold)
14. File "D:\Tools\anaconda3\envs\py38\lib\site-packages\torch\_ops.py", line 755, in __call__
15. return self._op(*args, **(kwargs or {}))
16. NotImplementedError: Could not run 'torchvision::nms' with arguments from the 'CUDA' backend. This could be because the operator doesn't exist for this backend, or was omitted during the selective/custom build process (if using custom build). If you are a Facebook employee using PyTorch on mobile, please visit https://fburl.com/ptmfixes for possible resolutions. 'torchvision::nms' is only available for these backends: [CPU, Meta, QuantizedCPU, BackendSelect, Python, FuncTorchDynamicLayerBackMode, Functionalize, Named, Conjugate, Negative, ZeroTensor, ADInplaceOrView, AutogradOther, AutogradCPU, AutogradCUDA, AutogradXLA, AutogradMPS, AutogradXPU, AutogradHPU, AutogradLazy, AutogradMeta, Tracer, AutocastCPU, AutocastCUDA, FuncTorchBatched, BatchedNestedTensor, FuncTorchVmapMode, Batched, VmapMode, FuncTorchGradWrapper, PythonTLSSnapshot, FuncTorchDynamicLayerFrontMode, PreDispatch, PythonDispatcher].

这个错误一般是由于 torch 和 torchvision 版本不匹配引起的，可以先卸载再重新安装。

1. pip uninstall torch torchvison
2. # 以安装gpu版本为例
3. pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118

链接：https://pan.baidu.com/s/10cTn3_idG9ouUKyuisb5Kg
提取码：18tw