基于暗原色先验的图像去雾优化技术

由于雾霾等大气散射作用会使战场环境获取的图像质量退化,图像去雾可以大幅度提升视觉效果,方便火控系统目标识别与跟踪。基于暗通道先验理论提出一种改进的中值滤波算法,针对不同区域使用不同大小的滤波窗口,可以快速有效地对透射率细化,避免边缘效应。并且还提出了天空区域透射率优化算法,去除了天空区域的纹理和斑块效应。实验结果表明,该算法具有速度快并且效果好的特点。

Optimized Image Dehazing Using Dark Channel Prior

Atmospheric scattering effect can make image quality degradation. Image dehazing can improve visual effect and facilitate target recognition and tracking. This paper proposes improved median filter algorithm based on dark channel prior,using different sizes of filter window with different region. It can effectively refine transmissivity and avoid holo effect. Also,this paper proposes sky area transmissivity refine algorithm. It wipes off veins of sky and patch effect. Experimental results show that this algorithm is high-efficient to dehaze the image.