[go: up one dir, main page]

Skip to content

Multi-thread tracking of YOLOv5 and ByteTrack implemented by C++, accelerated by TensorRT. YOLOv5 和 ByteTrack 的多线程追踪 C++ 实现, 使用 TensorRT 进行推理加速

License

Notifications You must be signed in to change notification settings

xieincz/YOLOv5_ByteTrack_Multithreading_TensorRT

Repository files navigation

YOLOv5_ByteTrack_Multithreading_TensorRT

Multi-thread tracking of YOLOv5 and ByteTrack implemented by C++, accelerated by TensorRT.

检测器采用的是 YOLOv5 (可以轻松替换成其他同类的检测器),追踪器采用的是 ByteTrack 。和其他同类项目不同的是加入了多线程处理以及用 C++和 TensorRT 加快推理速度。一个线程负责读取来自视频文件(可以轻松更改为摄像头)的帧,一个线程负责用 YOLOv5 得到检测框,一个线程用 ByteTrack 给各个检测框 reid ,还有一个线程负责将结果绘制到视频文件中。在多 CPU 核心(>=3 核)的设备上的效果比目前其他同类的项目要更快。而且还可以根据需要给各个线程设置 CPU 亲和性(将某线程绑定到某个 CPU 核心,该功能仅限于 Linux 平台)。

此外还利用 SWIG 包装了接口,方便在 python 中像调用一个库一样使用本项目。

TODO

  • YOLOv8

Usage and demo

Click the button shown below to try this project in colab.

Open In Colab

note: You can also upload the notebook in the colab folder of this project to the colab for running.

Acknowledgement

A large part of the code is borrowed from yolov5_deepsort_tensorrt, TensorRTx, YOLOv5 and ByteTrack. Thanks for their wonderful works.

About

Multi-thread tracking of YOLOv5 and ByteTrack implemented by C++, accelerated by TensorRT. YOLOv5 和 ByteTrack 的多线程追踪 C++ 实现, 使用 TensorRT 进行推理加速

Topics

Resources

License

Stars

Watchers

Forks