Sand-dust weather seriously reduces the effectiveness of computer vision equipment acquisition. To solve this problem, a fast sand-dust video quality improvement method based on adaptive dynamic guided filtering and an interframe detection strategy is proposed in this paper. First, an adaptive gamma correction with weighting distribution and color balance (AGCWDCB) and adaptive dynamic guided filtering (ADGIF) are used to perform color correction, contrast enhancement and detailed information enhancement on the first frame of the video. Second, the interframe detection model is constructed based on the normalized mean square error information between video frames. Finally, each frame after the first frame of the sand-dust video is processed according to the interframe detection strategy until a sand-dust video with improved quality is obtained. Through qualitative and quantitative comprehensive experiments on sand-dust images and videos, the experimental results are compared with the existing methods, the results of processing sand-dust images using our improved frame method have the best visual effect and the highest total scores in quantitative analysis. The results of interframe detection strategy show average 2.65 × speed up as compared with the frame-wise quality improvement method.

沙尘天气严重降低了计算机视觉设备采购的有效性。针对这一问题，本文提出了一种基于自适应动态引导滤波和帧间检测策略的快速沙尘视频质量改进方法。首先，使用具有加权分布和色彩平衡的自适应伽马校正（AGCWDCB）和自适应动态引导滤波（ADGIF）对视频的第一帧进行色彩校正、对比度增强和细节信息增强。其次，基于视频帧间的归一化均方误差信息构建帧间检测模型。最后，根据帧间检测策略对沙尘视频第一帧之后的每一帧进行处理，直到得到质量有所提高的沙尘视频。通过对沙尘图像和视频的定性和定量综合实验，将实验结果与现有方法进行对比，我们改进的帧法处理沙尘图像的结果视觉效果最好，在定量分析中总分最高. 与逐帧质量改进方法相比，帧间检测策略的结果平均提高了 2.65 倍。