Question

我正在尝试从UIImage捕获的视频帧中生成GPUImage。我已经完成了很多AVFoundation视频工作，但我不习惯使用GPUImage。我subclassed GPUImageVideoCamera并添加了此方法，但UIImage始终为零。如果有人能告诉我我哪里走得太可怕了，我会非常感激！

- (void)processVideoSampleBuffer:( CMSampleBufferRef )sampleBuffer

{ [super processVideoSampleBuffer：sampleBuffer]; //让GPUImage先进行处理

if (!self.thumbnailGenerated)
{
    CMTime timestamp = CMSampleBufferGetPresentationTimeStamp(sampleBuffer);
    NSLog(@"%f", (float)timestamp.value / timestamp.timescale);

    self.thumbnailGenerated = YES;

    dispatch_sync(dispatch_get_main_queue(), ^
                  {
                      // generate a preview frame from the last filter in the camera filter chain

                      UIImage *thumbnailImage = [UIImage imageWithCGImage:[[self.targets lastObject] newCGImageFromCurrentlyProcessedOutput]];
                      NSString *pathToMovie = [NSHomeDirectory() stringByAppendingPathComponent:@"Documents/Thumbnail.png"];
                      [UIImagePNGRepresentation(thumbnailImage) writeToFile:pathToMovie atomically:YES];
                  });
}
}

Answer 1

我已使用此代码生成第一帧的CGImageRef，用于缩略图

AVURLAsset *asset = [[AVURLAsset alloc] initWithURL:videoURl options:nil];
AVAssetImageGenerator *imageGenerator = [AVAssetImageGenerator assetImageGeneratorWithAsset:asset];
[imageGenerator setAppliesPreferredTrackTransform:YES];
NSData *videoData = [NSData dataWithContentsOfURL:asset.URL];
CGImageRef  image = [imageGenerator copyCGImageAtTime:kCMTimeZero actualTime:nil error:&error];

您可以将kCMTimeZero替换为某个实际值，以获得您想要的框架。

之后，您必须将CGImageRef转换为UIImage。

Answer 2

我不确定这是否有任何帮助，但我在处理视频时获取缩略图。为此，我使用

 videoInput --> someMyOperations --> fileOutput
 someMyOperations --> imageOutput //imageOutput is PictureOutput()
 videoInput.start() //that needs to be called!
 imageOutput.saveNextFrameToUrl(coverUrl, format: .jpg) { file in
     //here goes the code what to do with thumbnail
     videoInput.cancel() //quite probably here you want this
 }

那个猜测 - 我没有看到你的代码，但对我来说这很有效。

从GPUImage视频帧生成UIImage

2 个答案: