结果

基本实现了识别的功能。基本的界面如下

界面长得比较丑，主要是JavaSwing写界面比较麻烦，写个菜单栏都要那么多代码。目前不打算改了。

实现的思路是：使用opencv中自带的OpenCVFrameGrabber获取摄像头的数据，CanvasFrame来显示摄像头捕获的画面。

点击注册和识别都会在捕获的视频流自动抓取一帧图片，点下面的按钮可以查看拍摄效果

当输入想要比对的身份，点击识别之后，会与注册保存的数据进行比对

准确率还行。

附上部分代码

public static void OpenCamera()throwsException, InterruptedException{

OpenCVFrameGrabber grabber= new OpenCVFrameGrabber(0);//0表示本机摄像头 当然这里也可以换成网络摄像头地址

grabber.start(); //开始获取摄像头数据

CanvasFrame canvas = new CanvasFrame("倒计时5秒自动拍照注册");//新建一个窗口

canvas.setDefaultCloseOperation(JFrame.DISPOSE_ON_CLOSE);//窗口关闭时程序运行结束

canvas.setAlwaysOnTop(true);int i=0;while(true){if(i==30){//窗口是否关闭

System.out.println("已关闭");

grabber.stop();//停止抓取

canvas.dispose();//System.exit(2);//退出

}

canvas.showImage(grabber.grab());//获取摄像头图像并放到窗口上显示， 这里的Frame frame=grabber.grab(); frame表示一帧视频图像//调用doExecuteFrame()方法，将截取的图片保存在本地

if(i==1)CatchPhoto(grabber.grabFrame(),"./register"+"/"+ MainWindow.ID +".jpg");

Thread.sleep(50);//50毫秒刷新一次图像

i++;

}

}

//人脸检测

ImageInfo imageInfo = getRGBData(newFile(register));

List faceInfoList = new ArrayList();

errorCode=faceEngine.detectFaces(imageInfo.getImageData(), imageInfo.getWidth(), imageInfo.getHeight(), imageInfo.getImageFormat(), faceInfoList);

System.out.println(faceInfoList);//特征提取

FaceFeature faceFeature = newFaceFeature();

errorCode= faceEngine.extractFaceFeature(imageInfo.getImageData(), imageInfo.getWidth(), imageInfo.getHeight(), imageInfo.getImageFormat(), faceInfoList.get(0), faceFeature);

System.out.println("特征值大小：" +faceFeature.getFeatureData().length);//人脸检测2

ImageInfo imageInfo2 = getRGBData(newFile(discern));

List faceInfoList2 = new ArrayList();

errorCode=faceEngine.detectFaces(imageInfo2.getImageData(), imageInfo2.getWidth(), imageInfo2.getHeight(),imageInfo.getImageFormat(), faceInfoList2);

System.out.println(faceInfoList);//特征提取2

FaceFeature faceFeature2 = newFaceFeature();

errorCode= faceEngine.extractFaceFeature(imageInfo2.getImageData(), imageInfo2.getWidth(), imageInfo2.getHeight(), imageInfo.getImageFormat(), faceInfoList2.get(0), faceFeature2);

System.out.println("特征值大小：" +faceFeature.getFeatureData().length);//特征比对

FaceFeature targetFaceFeature = newFaceFeature();

targetFaceFeature.setFeatureData(faceFeature.getFeatureData());

FaceFeature sourceFaceFeature= newFaceFeature();

sourceFaceFeature.setFeatureData(faceFeature2.getFeatureData());

FaceSimilar faceSimilar= newFaceSimilar();

errorCode=faceEngine.compareFaceFeature(targetFaceFeature, sourceFaceFeature, faceSimilar);//System.out.println("相似度：" + faceSimilar.getScore());

MainWindow.Similarity.setText("相似度：" +faceSimilar.getScore());//设置活体测试

errorCode = faceEngine.setLivenessParam(0.5f, 0.7f);//人脸属性检测

FunctionConfiguration configuration = newFunctionConfiguration();

configuration.setSupportAge(true);

configuration.setSupportFace3dAngle(true);

configuration.setSupportGender(true);

configuration.setSupportLiveness(true);

errorCode=faceEngine.process(imageInfo.getImageData(), imageInfo.getWidth(), imageInfo.getHeight(), imageInfo.getImageFormat(), faceInfoList, configuration);//性别检测

List genderInfoList = new ArrayList();

errorCode=faceEngine.getGender(genderInfoList);//System.out.println("性别：" + genderInfoList.get(0).getGender());

if(genderInfoList.get(0).getGender()==0){

MainWindow.Sex.setText("性别：男");

}else if(genderInfoList.get(0).getGender()==1){

MainWindow.Sex.setText("性别：女");

}else{

MainWindow.Sex.setText("性别：未知");

}//MainWindow.Sex.setText("性别：" + genderInfoList.get(0).getGender());//年龄检测

List ageInfoList = new ArrayList();

errorCode=faceEngine.getAge(ageInfoList);//System.out.println("年龄：" + ageInfoList.get(0).getAge());

MainWindow.Age.setText("年龄：" + ageInfoList.get(0).getAge());//3D信息检测

List face3DAngleList = new ArrayList();

errorCode=faceEngine.getFace3DAngle(face3DAngleList);

System.out.println("3D角度：" + face3DAngleList.get(0).getPitch() + "," + face3DAngleList.get(0).getRoll() + "," + face3DAngleList.get(0).getYaw());

MainWindow.Angle.setText("3D角度：" + face3DAngleList.get(0).getPitch() + "," + face3DAngleList.get(0).getRoll() + "," + face3DAngleList.get(0).getYaw());//活体检测

List livenessInfoList = new ArrayList();

errorCode=faceEngine.getLiveness(livenessInfoList);//System.out.println("活体：" + livenessInfoList.get(0).getLiveness());

MainWindow.LiVing.setText("活体：" + livenessInfoList.get(0).getLiveness());//IR属性处理

ImageInfo imageInfoGray = getGrayData(new File("C:\\Users\\user\\Desktop\\test\\photo\\1.jpg"));

List faceInfoListGray = new ArrayList();

errorCode=faceEngine.detectFaces(imageInfoGray.getImageData(), imageInfoGray.getWidth(), imageInfoGray.getHeight(), imageInfoGray.getImageFormat(), faceInfoListGray);

FunctionConfiguration configuration2= newFunctionConfiguration();

configuration2.setSupportIRLiveness(true);

errorCode=faceEngine.processIr(imageInfoGray.getImageData(), imageInfoGray.getWidth(), imageInfoGray.getHeight(), imageInfoGray.getImageFormat(), faceInfoListGray, configuration2);//IR活体检测

List irLivenessInfo = new ArrayList<>();

errorCode=faceEngine.getLivenessIr(irLivenessInfo);

System.out.println("IR活体：" + irLivenessInfo.get(0).getLiveness());

ImageInfoEx imageInfoEx= newImageInfoEx();

imageInfoEx.setHeight(imageInfo.getHeight());

imageInfoEx.setWidth(imageInfo.getWidth());

imageInfoEx.setImageFormat(imageInfo.getImageFormat());

imageInfoEx.setImageDataPlanes(new byte[][]{imageInfo.getImageData()});

imageInfoEx.setImageStrides(new int[]{imageInfo.getWidth() * 3});

List faceInfoList1 = new ArrayList<>();

errorCode=faceEngine.detectFaces(imageInfoEx, DetectModel.ASF_DETECT_MODEL_RGB, faceInfoList1);

FunctionConfiguration fun= newFunctionConfiguration();

fun.setSupportAge(true);

errorCode=faceEngine.process(imageInfoEx, faceInfoList1, functionConfiguration);

List ageInfoList1 = new ArrayList<>();int age =faceEngine.getAge(ageInfoList1);

System.out.println("年龄：" + ageInfoList1.get(0).getAge());

FaceFeature feature= newFaceFeature();

errorCode= faceEngine.extractFaceFeature(imageInfoEx, faceInfoList1.get(0), feature);

问题记录

1.捕获视频流

网上找到了很多方法，比如JMF、ffmpeg等。JMF以前用过，过于老旧了，而且只支持32位系统，想用的话还得用32位的IDE，所以忽略。然后发现用opencv集成了ffmpeg，通过grabber.grab()方法就可以获取。很简单。

2.视频显示

一开始的想法，grabber.grab()获取的是一帧一帧的图片，那么可以再frame里加一个显示图片的label，每过十毫秒刷新一次图片，这样就起到了播放视频的效果。但是遇到了问题

使用java.awt.Image 中的方法无法获取grabber.grab()中的数据，强制类型转换无效。后来查看OpenCVFrameGrabber的源码。最后在Frame.Class里发现image的类型

NIO与IO之间是有一些区别的，NIO要更优秀一点，可惜Java界面没有

所以转换思路，没办法在主窗口播放，那就在新建一个窗口，看起来有点别扭，但受限于技术水平，还是先把功能实现了。

javacv里有一个简单的新建窗口方式

CanvasFrame canvas = new CanvasFrame("Camera");//新建一个窗口

通过源码可以看到CanvasFrame是继承了javax.swing.JFrame类的。使用这个类主要是它里面的showimage方法

里面有适合的数据类型。

3.窗口问题

使用新建窗口显示视频时，关闭视频播放窗口，主窗口也会关闭。

原来关闭窗口使用的是System.exit(0);这个方法直接终止了虚拟机。后来改成了dispose()方法(需要.setDefaultCloseOperation设置为JFrame.DISPOSE_ON_CLOSE)