OpenCV4.8 Java SDK实现YOLOv5模型部署

OpenCV学堂

发布于 2024-05-17 18:29:58

7191

发布于 2024-05-17 18:29:58

文章被收录于专栏：贾志刚-OpenCV学堂

OpenCV Java开发环境配置

Java 开发的IDE，推荐选择IDEA这个，真的比较好用，我下载的是社区版，免费可用。

安装好以后直接点击打开，选择 New Project

添加本地的OpenCV Java SDK的 Jar文件依赖包

从Project Structure到 Modules 到 Module SDK 添加Jars or Dependence

OpenCV Java 代码测试

OpenCV Java配置好以后，测试代码首先需要加载DLL，然后依赖Java SDK，可以实现图像加载与显示。测试代码如下：

package org.example;

import org.opencv.core.*;
import org.opencv.highgui.HighGui;
import org.opencv.imgcodecs.Imgcodecs;


//TIP To <b>Run</b> code, press <shortcut actionId="Run"/> or
// click the <icon src="AllIcons.Actions.Execute"/> icon in the gutter.
public class YOLOv5DNNDetector {
    public static void main(String[] args) {
        System.load("D:/opencv-4.8.0/opencv/build/java/x64/opencv_java480.dll");
        System.out.println("start to read image...");
        Mat inputImage = Imgcodecs.imread("D:/images/lena.jpg");
        HighGui.imshow("Java OpenCV开发测试", inputImage);
        HighGui.waitKey(0);
        HighGui.destroyAllWindows();
        System.exit(0);
    }
}

运行结果如下：

YOLOv5模型部署

基于YOLOv5 7.0版本，完成了对象检测

public static void main(String[] args) {

    System.load("D:/opencv-4.8.0/opencv/build/java/x64/opencv_java480.dll");
    System.out.println("start to read image...");
    String file_path = "D:/python/yolov5-7.0/classes.txt";
    try {
        Mat inputImage = Imgcodecs.imread("D:/bird_test/city-walk.jpg");
        java.util.ArrayList<Rect> boxes = new ArrayList<>();
        FileInputStream fis = new FileInputStream(file_path);
        BufferedReader br = new BufferedReader(new InputStreamReader(fis));
        YOLOv5DNNDetector detector = new YOLOv5DNNDetector("D:/python/yolov5-7.0/yolov5s.onnx", 0.5f, 0.5f, 640, 640);
        String line;
        ArrayList<String> map_labels = new ArrayList<>();
        while ((line = br.readLine()) != null) {
            map_labels.add(line);
        }
        detector.addAllLabels(map_labels);
        detector.infer_image(inputImage, boxes);
        HighGui.imshow("OpenCV4.8 for Java YOLOv5对象检测演示", inputImage);
        HighGui.waitKey(0);
        HighGui.destroyAllWindows();
    } catch (Exception e) {
        e.printStackTrace();
    }
    System.exit(0);
}