欢迎访问《图学学报》
分享到：
摘要：针对目前去雾算法实时性较差，对天空等区域的处理不理想以及去雾后的图像视
觉效果较差等问题，提出一种新的基于半逆法的快速单幅图像去雾算法。首先从大气散射模型
出发，利用改进的半逆算法得到大气整体光照值；其次，基于大气散射光特性，以图像边缘信
息为合成条件融合图像的边缘信息和场景深度信息，准确估测大气光幂；然后，根据大气散射
模型得到初步复原无雾图像；最后，经过色调调整和细节增强处理，得到一幅真实感强烈的无
雾图像。对于深度发生突变或者远景像素点，消除了光晕效应。与其他算法相比，本算法能很
好保持色彩和细节信息，具有较好的实时性和鲁棒性。
Abstract: In order to overcome the defects of the existing algorithms, such as the poor real-time#br# performance, bad effect in sky area and poor visual effects to dehaze image. A new algorithm for fast#br# single image dehazing is presented based on the semi-inverse method from the point of view of#br# atmospheric scattering model. The improved semi-inverse algorithm is firstly used to obtain the#br# global atmospheric illumination value. Secondly, based on the characteristics of atmospheric#br# scattering light, the atmospheric veil is accurately estimated fusing the image edge information and#br# scene depth information. Then, the restore image eliminates the halo effect for depth uncontinuous#br# pixels or distant pixels according to the atmospheric scattering model. Finally, the obtained restore#br# image has a strong sense of reality by using of the hue adjustment and detail enhancement. Compared#br# with the other algorithms, our algorithm can maintain color and detail information and ensure the#br# real-time performance.
