人工智能Java SDK：读取本地MP4文件，实时检测口罩

最新推荐文章于 2022-02-24 10:53:50 发布

Calvin.AIAS

最新推荐文章于 2022-02-24 10:53:50 发布

阅读量201

点赞数

分类专栏： AIAS 文章标签：人工智能 java 开发语言

本文链接：https://blog.csdn.net/weixin_39355136/article/details/121158199

版权

AIAS 专栏收录该内容

77 篇文章 20 订阅

订阅专栏

口罩检测

口罩检测助力抗击肺炎，人工智能技术正被应用到疫情防控中来。抗疫切断传播途径中，
佩戴口罩已经几乎成为了最重要的举措之一。但是在实际场景中，仍然有不重视、不注意、
侥幸心理的人员不戴口罩，尤其在公众场合，给个人和公众造成极大的风险隐患。
而基于人工智能的口罩检测功能可以基于摄像头视频流进行实时检测。

SDK功能

读取本地MP4文件，实时检测口罩。

运行例子

首先下载例子代码

git clone https://github.com/mymagicpower/AIAS.git

运行例子代码：MP4FaceMaskDetectionExample

/**
 * 本地视频人脸检测
 *
 * @author Calvin
 */
public class MP4FaceMaskDetectionExample {
  public static void main(String[] args) throws IOException, ModelException, TranslateException {
    faceMaskDetection("src/test/resources/mask.mp4");
  }

  /**
   * 人脸检测
   *
   * @param input 视频源
   */
  public static void faceMaskDetection(String input)
      throws IOException, ModelException, TranslateException {
    float shrink = 0.5f;
    float threshold = 0.85f;
    Criteria<Image, DetectedObjects> criteria = new FaceDetection().criteria(shrink, threshold);
    Criteria<Image, Classifications> maskCriteria = new FaceMaskDetect().criteria();

    // 读取视频文件或者视频流获取图像（得到的图像为frame类型，需要转换为mat类型进行检测和识别）
    FFmpegFrameGrabber grabber = new FFmpegFrameGrabber(input);
    grabber.start();

    // Frame与Mat转换
    OpenCVFrameConverter.ToMat converter = new OpenCVFrameConverter.ToMat();

    CanvasFrame canvas = new CanvasFrame("人脸检测"); // 新建一个预览窗口
    canvas.setDefaultCloseOperation(JFrame.EXIT_ON_CLOSE);
    canvas.setVisible(true);
    canvas.setFocusable(true);
    // 窗口置顶
    if (canvas.isAlwaysOnTopSupported()) {
      canvas.setAlwaysOnTop(true);
    }
    Frame frame = null;

    try (ZooModel model = ModelZoo.loadModel(criteria);
        Predictor<Image, DetectedObjects> predictor = model.newPredictor();
        ZooModel classifyModel = ModelZoo.loadModel(maskCriteria);
        Predictor<Image, Classifications> classifier = classifyModel.newPredictor()) {
      // 获取图像帧
      for (; canvas.isVisible() && (frame = grabber.grabImage()) != null; ) {

        // 将获取的frame转化成mat数据类型
        Mat img = converter.convert(frame);
        BufferedImage buffImg = OpenCVImageUtil.mat2BufferedImage(img);

        Image image = ImageFactory.getInstance().fromImage(buffImg);
        int imageWidth = image.getWidth();
        int imageHeight = image.getHeight();

        DetectedObjects detections = predictor.predict(image);
        List<DetectedObjects.DetectedObject> items = detections.items();

        // 遍历人脸
        for (DetectedObjects.DetectedObject item : items) {
          Image subImg = getSubImage(image, item.getBoundingBox());
          Classifications classifications = classifier.predict(subImg);
          String className = classifications.best().getClassName();

          BoundingBox box = item.getBoundingBox();
          Rectangle rectangle = box.getBounds();
          int x = (int) (rectangle.getX() * imageWidth);
          int y = (int) (rectangle.getY() * imageHeight);
          Rect face =
              new Rect(
                  x,
                  y,
                  (int) (rectangle.getWidth() * imageWidth),
                  (int) (rectangle.getHeight() * imageHeight));

          // 绘制人脸矩形区域，scalar色彩顺序：BGR(蓝绿红)
          rectangle(img, face, new Scalar(0, 0, 255, 1));

          int pos_x = Math.max(face.tl().x() - 10, 0);
          int pos_y = Math.max(face.tl().y() - 10, 0);
          // 在人脸矩形上面绘制文字
          putText(
              img,
              className,
              new Point(pos_x, pos_y),
              FONT_HERSHEY_COMPLEX,
              1.0,
              new Scalar(0, 0, 255, 2.0));
        }

        // 显示视频图像
        canvas.showImage(frame);
      }
    }

    canvas.dispose();
    grabber.close();
  }

  private static int[] extendSquare(
      double xmin, double ymin, double width, double height, double percentage) {
    double centerx = xmin + width / 2;
    double centery = ymin + height / 2;
    double maxDist = Math.max(width / 2, height / 2) * (1 + percentage);
    return new int[] {(int) (centerx - maxDist), (int) (centery - maxDist), (int) (2 * maxDist)};
    //    return new int[] {(int) xmin, (int) ymin, (int) width, (int) height};
  }

  private static Image getSubImage(Image img, BoundingBox box) {
    Rectangle rect = box.getBounds();
    int width = img.getWidth();
    int height = img.getHeight();
    int[] squareBox =
        extendSquare(
            rect.getX() * width,
            rect.getY() * height,
            rect.getWidth() * width,
            rect.getHeight() * height,
            0); // 0.18

    if (squareBox[0] < 0) squareBox[0] = 0;
    if (squareBox[1] < 0) squareBox[1] = 0;
    if (squareBox[0] > width) squareBox[0] = width;
    if (squareBox[1] > height) squareBox[1] = height;
    if ((squareBox[0] + squareBox[2]) > width) squareBox[2] = width - squareBox[0];
    if ((squareBox[1] + squareBox[2]) > height) squareBox[2] = height - squareBox[1];
    return img.getSubimage(squareBox[0], squareBox[1], squareBox[2], squareBox[2]);
    //    return img.getSubimage(squareBox[0], squareBox[1], squareBox[2], squareBox[3]);
  }
}