基于流处理器的YUV视频模糊度分析方法

Blur analysis method for YUV video based on stream processor

视频在传输和编码转换过程中可能会出现质量下降,因此需要对传输后或者编码转换后的视频进行质量判断,以此来确定传输或者编码转换的效果。传统使用肉眼进行主观判断的方法存在误差较大的问题,因此需要一种更加精确的评价模型。在无参考评价模型的基础上,给出了一个基于流处理器的模糊度分析模型,它能够并行地对大量YUV视频进行逐帧的模糊度分析,可以实现多个视频流的横向对比,并且给出比较精确的模糊度评价指标。实验表明,该模型可以在2 s以内完成对1 000帧YUV图像的分析。

The video quality will descend during transmission and code conversion,so it is need to evaluate the video quality after these two processes to judge the loss caused by transmission and code conversion.Since traditional method of visual observation will introduce greater error into the evaluation,more accurate evaluation model is needed.The results propose a blur analysis model based on stream processors over the no-reference analysis model and it cans analysis blur effect of massive YUV videos frame by frame and achieve lateral comparison of multiple video streams.Experiments show the model can analysis 1 000 YUV images within 2 seconds.