一种基于暗原色先验的低照度视频增强算法

Low lighting video enhancement algorithm based on dark channel prior

在低照度条件下进行拍摄时,得到的视频对比度低,细节信息很难体现出来。提出了一种改进的基于暗原色先验的低照度视频增强算法来增强这类视频。传统基于暗原色先验的增强算法会放大视频噪声,为了克服这个缺点,利用时域累积信息对低照度视频进行去噪预处理。在增强过程中,采用快速算法逐帧计算,增强后的视频相比于传统方法具有更高保真度。实验结果表明,算法可有效抑制视频中的噪声、提高视频的对比度、突出细节信息,从而获得更加清晰、自然的高质量视频。

The video captured in low lighting condition has low contrast,and the details in it are hidden.In this paper,an improved low lighting video enhancement algorithm based on dark channel prior is proposed to enhance this kind of video.Traditional enhancement method is based on dark channel prior enlarges video noise.To overcome this drawback,the proposed algorithm exploits temporal accumulation information to preprocess the low lighting video and remove video noise.During enhancing process,fast method is applied on every frame,and the enhanced videos＇ fidelity is higher than the fidelity of videos enhanced by traditional method.Experimental results show that the proposed algorithm can effectively suppress video noise,improve the contrast and highlight details.The enhanced results are clearer and more natural with high quality.