Skip to content

A C++ implementation of Deepsort in Jetson Xavier nx and Jetson nano

License

Notifications You must be signed in to change notification settings

RichardoMrMu/deepsort-tensorrt

Repository files navigation

DeepSort_TensorRT

Use tensorrt to accelerate deepsort tracking. 使用tensorrt来加速deepsort的特征提取。

Install

[REQUIRE] TensorRT 7

[REQUIRE] OpenCV

下载本库和deepsort的pytorch库

git clone [email protected]:ZQPei/deep_sort_pytorch.git
git clone https://github.com/RichardoMrMu/deepsort-tensorrt.git

配置好pytorch工程后,将exportOnnx.py复制到pytorch工程的项目目录下运行,将会生成deepsort.onnx。

把deepsort.onnx移到本工程的resources目录下。然后回到工程的根目录

mkdir build
cd build
cmake ..
make
./onnx2engine ../resources/deepsort.onnx ../resources/deepsort.engine

然后就可以运行demo测试了

./demo ../resources/deepsort.engine ../resources/track.txt

然后就可以把自己TensorRT的目标检测工程进行对接。

git clone https://github.com/wang-xinyu/tensorrtx.git

注:视频文件太大,所以只把检测结果保存成txt进行测试。

About

A C++ implementation of Deepsort in Jetson Xavier nx and Jetson nano

Topics

Resources

License

Stars

Watchers

Forks

Packages

No packages published