Coder-TieHH/Yolov5s_ros

借鉴了官方的实例以及git上的一位大佬，记不得仓库了

ROS_Tengine_NPU

This is a ROS package for Tengine, a high-performance neural network inference framework by OPEN AI LAB optimized for embedded devices:

• ARM NEON assembly level optimization
• Sophisticated memory management and data structure design, very low memory footprint
• Completely separated front-end/back-end design, Supports CPU, GPU, NPU and other heterogeneous computing units
• Extensible model design, supports uint8/int8 quantization storage
• Can import caffe/pytorch/mxnet/onnx/ncnn models

Setting up

• 晶晨A311D
• USB Camera
• Tengine Compile with TIM-VX

Library

ROS package

git clone git@github.com:Coder-TieHH/Yolov5n_ros.git
cd ros_tengine
git submodule update --init

├── Catkin_make.sh
├── Clean.sh
├── Run.sh
└── src
    ├── ros_tengine
    │   ├── assets
    │   │   └── models
    │   │       └── yolov3_uint8.tmfile
    │   ├── CMakeLists.txt
    │   ├── include
    │   │   └── ros_tengine
    │   │       ├── c_api.h
    │   │       ├── defines.h
    │   │       ├── timer.hpp
    │   │       ├── timvx_device.h
    │   │       ├── types.hpp
    │   │       ├── yolo.hpp
    │   │       └── yolo_layer.hpp
    │   ├── launch
    │   │   └── yolov3.launch
    │   ├── lib
    │   │   └── libtengine-lite.so
    │   ├── msg
    │   │   ├── Euler.msg
    │   │   ├── FaceObject.msg
    │   │   ├── Object.msg
    │   │   ├── Rectangle.msg
    │   │   └── Vector2D.msg
    │   ├── package.xml
    │   └── src
    │       ├── timer.cc
    │       ├── yolo.cpp
    │       ├── yolo_layer.cpp
    │       └── yolov3_node.cpp
    └── usb_cam

How to build

bash Catkin_make.sh

How to run

bash Run.sh

✌️ Acknowledgements

Tengine

Insightface-NPU