java截取视频的帧

淩亂°似流年 2022-10-27 05:22 245阅读 0赞

### 使用javacv获取视频(本地视频文件、视频链接)中的帧，保存成图片或者base64串 ###

#### POM文件引入依赖 ####

<dependency>
        <groupId>org.bytedeco</groupId>
        <artifactId>javacv</artifactId>
        <version>1.4.1</version>
    </dependency>
    <dependency>
        <groupId>org.bytedeco</groupId>
        <artifactId>javacpp</artifactId>
        <version>1.4.1</version>
    </dependency>
    <dependency>
        <groupId>org.bytedeco.javacpp-presets</groupId>
        <artifactId>opencv-platform</artifactId>
        <version>3.4.1-1.4.1</version>
    </dependency>
    <dependency>
        <groupId>org.bytedeco.javacpp-presets</groupId>
        <artifactId>ffmpeg-platform</artifactId>
        <version>3.4.2-1.4.1</version>
    </dependency>

#### 提取视频帧 ####

package com.mf.video;
    
    import org.bytedeco.javacv.FFmpegFrameGrabber;
    import org.bytedeco.javacv.Frame;
    import org.bytedeco.javacv.Java2DFrameConverter;
    import sun.misc.BASE64Decoder;
    import sun.misc.BASE64Encoder;
    
    import javax.imageio.ImageIO;
    import java.awt.image.BufferedImage;
    import java.io.ByteArrayOutputStream;
    import java.io.File;
    import java.io.FileOutputStream;
    import java.util.*;
    import java.util.List;
    
    /**
     * @Author KeXin
     * @Date 2021/2/1 下午3:07
     **/
    public class VideoUtil {
    
        /**
         * base64字符串转换成图片
         * @param imgStr      base64字符串
         * @param imgFilePath 图片存放路径
         */
        private static void base64ToImage(String imgStr, String imgFilePath) {
            try {
                byte[] buffer = new BASE64Decoder().decodeBuffer(imgStr);
                FileOutputStream out = new FileOutputStream(imgFilePath);
                out.write(buffer);
                out.close();
            }catch (Exception e){
                e.printStackTrace();
            }
        }
    
        /**
         * 将截取的帧生成文件或者生成base64
         * @param frame
         * @return
         */
        private static String getBase64FromFrame(Frame frame) {
            String imgFormat = "jpg";
            Java2DFrameConverter converter = new Java2DFrameConverter();
            BufferedImage srcBi = converter.getBufferedImage(frame);
            // 可以选择对截取的帧进行等比例缩放
    //        int owidth = srcBi.getWidth();
    //        int oheight = srcBi.getHeight();
    //        int width = 800;
    //        int height = (int) (((double) width / owidth) * oheight);
    //        BufferedImage bi = new BufferedImage(width, height, BufferedImage.TYPE_3BYTE_BGR);
    //        bi.getGraphics().drawImage(srcBi.getScaledInstance(width, height, Image.SCALE_SMOOTH), 0, 0, null);
            try {
                ByteArrayOutputStream output = new ByteArrayOutputStream();
                // ImageIO工具类提供了将图片写成文件或者outputStream中
    //            ImageIO.write(bi, imgFormat, targetFile);
                ImageIO.write(srcBi, imgFormat, output);
                // 这里需要获取图片的base64数据串，所以将图片写到流里面
                return new BASE64Encoder().encode(output.toByteArray());
            } catch (Exception e) {
                e.printStackTrace();
            }
            return null;
        }
    
        /**
         * 获取指定视频的帧并保存为图片至指定目录
         *
         * @param fileUrl  源视频文件Url路径
         * @param stepSecond 每隔几秒取一帧，默认1200s
         * @param count 最多获取几帧
         * @return 返回截取的帧的Base64串列表
         * @throws Exception
         */
        public static List<String> fetchPicFromVideo(String fileUrl, Integer stepSecond, Integer count) throws Exception {
            List<String> picBase64List = new ArrayList<>();
            stepSecond = stepSecond == null ? 1200 : stepSecond;
    
            FFmpegFrameGrabber ff = new FFmpegFrameGrabber(fileUrl);
            ff.start();
            long timeLength = ff.getLengthInTime();
            Frame frame = ff.grabImage();
            long startTime = frame.timestamp;
            long timestamp = 0;
            int second = 0;
            int picNum = 0;
            while (timestamp <= timeLength) {
                timestamp = startTime + second * 1000000L;
                ff.setTimestamp(timestamp);
                frame = ff.grabImage();
                if (frame != null) {
                    if (frame.image != null) {
                        picBase64List.add(getBase64FromFrame(frame));
                        picNum++;
                        if (count != null && picNum == count) {
                            break;
                        }
                    }
                }
                second += stepSecond;
            }
            ff.stop();
            return picBase64List;
        }
    
    
        public static void main(String[] args) {
            try {
                String outputPath = "/Users/kexin/work/projects/learn/MavenUtil/output";
                String videoPath = "https:*********165dd278762865.mp4";
                List<String> files = VideoUtil.fetchPicFromVideo(videoPath, 600, 8);
                // 为了验证base64串的正确性，解码生成图片文件
                for (String imgBase64 : files) {
                    String fileName = String.valueOf(System.currentTimeMillis());
                    String targetFile = outputPath + File.separator + fileName + ".jpg";
                    base64ToImage(imgBase64, targetFile);
                }
            } catch (Exception e) {
                e.printStackTrace();
            }
        }
    
    }