前言
只是最近碰到有这方面的项目需求，所以简单 Mark 下本文。下面的示例是参考过他人分享的文章，之后本人再自行实践、调整和测试过的，希望对有这方面需求的人有所帮助。
示例
- 添加依赖
<dependency>
<groupId>org.bytedeco</groupId>
<artifactId>javacv-platform</artifactId>
<version>1.4.4</version>
</dependency>
上述这段 maven 依赖包含了完整的 javacv 功能 (非常多，依赖
Jar就占大概有500MB)，由于这里只使用到了其中ffmpeg这块的特性，因此也可以像下面这样排除掉无关的部分
<dependency>
<groupId>org.bytedeco</groupId>
<artifactId>javacv</artifactId>
<version>1.4.4</version>
<exclusions>
<exclusion>
<groupId>org.bytedeco</groupId>
<artifactId>javacpp</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>flycapture</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>libdc1394</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>libfreenect</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>libfreenect2</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>librealsense</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>videoinput</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>opencv</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>tesseract</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>leptonica</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>flandmark</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>artoolkitplus</artifactId>
</exclusion>
</exclusions>
</dependency>
<dependency>
<groupId>org.bytedeco</groupId>
<artifactId>javacv-platform</artifactId>
<version>1.4.4</version>
<exclusions>
<exclusion>
<groupId>org.bytedeco</groupId>
<artifactId>javacv</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>flycapture-platform</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>libdc1394-platform</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>libfreenect-platform</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>libfreenect2-platform</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>librealsense-platform</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>videoinput-platform</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>opencv-platform</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>tesseract-platform</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>leptonica-platform</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>flandmark-platform</artifactId>
</exclusion>
<exclusion>
<groupId>org.bytedeco.javacpp-presets</groupId>
<artifactId>artoolkitplus-platform</artifactId>
</exclusion>
</exclusions>
</dependency>
- 核心代码
- 获取视频时长
/**
* 获取视频时长，单位为秒
*
* @param video 源视频文件
* @return 时长（s）
*/
public static long getVideoDuration(File video) {
long duration = 0L;
FFmpegFrameGrabber ff = new FFmpegFrameGrabber(video);
try {
ff.start();
duration = ff.getLengthInTime() / (1000 * 1000);
ff.stop();
} catch (FrameGrabber.Exception e) {
e.printStackTrace();
}
return duration;
}
- 截取视频指定帧为图片
/**
* 截取视频获得指定帧的图片
*
* @param video 源视频文件
* @param picPath 截图存放路径
*/
public static void getVideoPic(File video, String picPath) {
FFmpegFrameGrabber ff = new FFmpegFrameGrabber(video);
try {
ff.start();
// 截取中间帧图片(具体依实际情况而定)
int i = 0;
int length = ff.getLengthInFrames();
int middleFrame = length / 2;
Frame frame = null;
while (i < length) {
frame = ff.grabFrame();
if ((i > middleFrame) && (frame.image != null)) {
break;
}
i++;
}
// 截取的帧图片
Java2DFrameConverter converter = new Java2DFrameConverter();
BufferedImage srcImage = converter.getBufferedImage(frame);
int srcImageWidth = srcImage.getWidth();
int srcImageHeight = srcImage.getHeight();
// 对截图进行等比例缩放(缩略图)
int width = 480;
int height = (int) (((double) width / srcImageWidth) * srcImageHeight);
BufferedImage thumbnailImage = new BufferedImage(width, height, BufferedImage.TYPE_3BYTE_BGR);
thumbnailImage.getGraphics().drawImage(srcImage.getScaledInstance(width, height, Image.SCALE_SMOOTH), 0, 0, null);
File picFile = new File(picPath);
ImageIO.write(thumbnailImage, "jpg", picFile);
ff.stop();
} catch (IOException e) {
e.printStackTrace();
}
}
- 测试用例
public static void main(String[] args) {
String videoPath = ResourceUtils.CLASSPATH_URL_PREFIX + "video.mp4";
File video = null;
try {
video = ResourceUtils.getFile(videoPath);
} catch (FileNotFoundException e) {
e.printStackTrace();
}
String picPath = "video.jpg";
getVideoPic(video, picPath);
long duration = getVideoDuration(video);
System.out.println("videoDuration = " + duration);
}
文章已授权转载，原文链接：Java 获取视频时长及截取帧截图
