目的 针对视频序列相邻帧图像之间的相关性， 提出了基于帧间差分的背景图像的视频图像重构算法。方法 利用前2帧的重构图像求得背景图像， 作为下一帧图像重构的先验知识。结果 明显提高了重构图像的质量， 同时也减少了重构图像所需时间。结论 将帧间差分法与压缩感知重构算法相结合得到的视频重构算法， 在视频图像重构的质量和速度上都有明显优势。

Objective According to the correlation between inter frames of video sequence images, a video image reconstruction algorithm based on inter-frame difference was proposed. Methods The background image obtained by the first two reconstructed frame image was used as the prior knowledge for reconstruction of the next frame image. Results The method obviously improved the quality of the reconstructed image and reduced the reconstruction time. Conclusion Simulation results showed that the method had obvious advantages in image quality and reconstruction speed by combining the inter-frame difference method and the compressive sensing reconstruction algorithm.