基于图像分割的去雾算法

Rapid Defogging Algorithm Based on Image Segmentation

针对雾霾条件下,所得的单幅图像出现降质现象,因而对视觉效果要求高的图像进行复原具有必要性;基于图像分割的去雾算法以暗通道先验模型为基础对大气光矢量A值的求取和透射率t(x,y)的处理方法实现改进;首先对单幅图像进行阈值分割找到天空区域,在所获取的天空区域部分结合skyline算法,可以找到精确的大气光矢量A值;进而对初始透射率t(x,y)采用改进的约束最小二乘方滤波进行优化,得到优化透射率t1(x,y),最后将所得的大气光矢量A值和优化透射率t1(x,y)利用大气光传输物理模型复原;改进算法的去雾结果具备保留细致的边缘细节,同时具有高效的去除图像噪声能力;实验结果表明,与he方法相对比,改进去雾算法的处理时间大程度缩短的同时,图像效果得到了提升。

Directed at the haze conditions, the resulting single image degradation, therefore, the images that require high visual effects need to restore. The optimized algorithm based on dark channel prior improve the way of getting the atmospheric light vector A and the transmittance t （x, y）. First using single threshold segmentation find the sky region, then combine with skyline algorithms can locate the precise atmospheric light vector A; then adopting improved Constrained least--square filter optimize the initial transmittance t （x, y） and get optimized transmittance t1 （x, y）, and finally restore degrade image by physical model otc atmospheric with the resulting atmospheric light vector A and optimal transmission t1 （x, y）. the improved defogging algorithm not only retained meticulous edge detail, but also has removed image noise efficiently. Experimental results show that compared with the he method, the improved method short defogging processing time, and improve the image effect simultaneously.