基于改进ANE的视频强弱光消除算法

Algorithm for Elimination of the High-low Intensity Light in Video Based on Improved AHE

针对监控杭州湾大桥南接线高速公路隧道的摄像头存在受强弱光影响较严重的问题,提出了改进的自适应直方图算法。该算法首先根据实际视频确定滑动窗口的尺寸和直方图均衡化公式参数,以获得最佳画面效果为准,然后以此滑动窗口对视频某一通道图像作卷积,根据改进的直方图均衡化公式计算每个像素点的灰度值,并最终对其余的视频通道分别作卷积。通过该算法能消除强弱光,增强画面质量,便于后续的目标分类和跟踪。

Cameras used to monitor the south line expressway tunnel of the Hangzhou Bay Bridge are heavily influenced by the high-low intensity light. This paper proposes improved adaptive histogram equalization （AHE） to address the problem. First, the sliding window size and AHE parameters are set according to the actual video effect. Then the sliding window is convolved with one of the video channels using improved histogram equalization equation. Finally, the other 2 channels are further convolved by this method. Results show that this algorithm can effectively eliminate the high-low intensity light, enhance the image quality, and facilitate further object tracking and classification.