Cropping of Detected object 裁剪照片功能

[jocelyntehsr]

Env

• GPU Nano
• OS, Ubuntu16.0
• Cuda version 10.2
• TensorRT version 8.0

About this repo

• which model? yolov5

Your problem

Hi, i am able to run the yolov5 successfully on my nano. However, I have a question, does TensorRTx has the function to save the detected object in a cropped image? Something like ultralytics/yolov5#5412 (comment)

This is because I'd like to process the image in OpenCV after it is detected by yolov5. Thank you

您好，我成功执行到了yolov5 engine model。但是，我想请问TensorRTx 有没有功能去把检测到的物体裁剪起来然后另存为图像？如链接所示：ultralytics/yolov5#5412 (comment)

我想要把检测到的物体串流进 opencv去处理。谢谢！

[wang-xinyu]

This part you need to implement it by yourself. Not difficult.