https://github.com/bhoja1234/yolo-onnxruntime-cpp

Cpp-based onnx deployment, supporting yolo and unsupervised models
https://github.com/bhoja1234/yolo-onnxruntime-cpp

anomalib cpp onnx onnxruntime opencv yolo

Last synced: 2 months ago
JSON representation

Cpp-based onnx deployment, supporting yolo and unsupervised models

• Host: GitHub
• URL: https://github.com/bhoja1234/yolo-onnxruntime-cpp
• Owner: Bhoja1234
• Created: 2025-05-06T02:51:55.000Z (about 1 year ago)
• Default Branch: main
• Last Pushed: 2025-05-06T03:37:35.000Z (about 1 year ago)
• Last Synced: 2025-05-06T04:19:00.277Z (about 1 year ago)
• Topics: anomalib, cpp, onnx, onnxruntime, opencv, yolo
• Language: C++
• Homepage:
• Size: 22.3 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# yolo-onnxruntime-cpp

基于Onnxruntime的Yolo模型部署，无监督模型部署

支持检测、分割、分类、关键点检测、无监督模型

## 环境

- Opencv 4.x

- Onnxruntime 1.9

- CUDA 12.5

- OS: windows11 x64（x86不支持CUDA）

## 配置

需要配置好opencv和onnxruntime环境

## 使用说明

以检测模型为例：

```cpp

DetectTest(){

    DCSP_CORE* yoloDetector = new DCSP_CORE;

    yoloDetector->classes = { "person" }; 

    DL_INIT_PARAM params;

    params.ModelPath = "./models/yolov8n.onnx";

    params.rectConfidenceThreshold = 0.5;

    params.iouThreshold = 0.5;

    params.imgSize = { 640, 640 };

    params.modelType = YOLO_DETECT;

    ...

    yoloDetector->CreateSession(params);

    Detector(yoloDetector);

}

```

1. 指定模型检测类别：yoloDetector->classes = { "person" };写几个就检测几个。例如检测模型类别0表示"person"，类别1表示"bicycle", 类别2表示"car"...大括号里面只写"person"，其他检测类别就会被过滤；大括号里面写两个字符串，就会只检测前两个。

2. 指定模型路径：params.ModelPath

3. 具体图片路径需要在Detector()中设置

## 运行

在`main(){}`中，选择自己需要的模式，运行 `源.cpp`即可，自动检测GPU，没有GPU默认使用CPU

```cpp

main(){

...

    //DetectTest();  // 检测

    //ClsTest();  // 分类

    SegmentTest();  //分割

    //KeypointTest();  // 关键点检测

    //UnsuperviedTest();  //无监督

...

}

```

## 效果