改进遗传算法在雾天图像清晰化中的应用  

Application of improved genetic algorithm to misty image clearness

作　　者：叶青[1] 洪霞飞[1] 

机构地区：[1]长沙理工大学电气与信息工程学院,湖南长沙410004

出　　处：《长沙理工大学学报（自然科学版）》2011年第1期65-69,共5页Journal of Changsha University of Science and Technology:Natural Science

基　　金：湖南省自然科学基金资助项目(07JJ6123)

摘　　要：将改进遗传算法应用于雾天图像的区域分割,从而使图像清晰化.该方法首先用遗传算法求出近景和远景的分割阈值将图像分割出来,然后通过移动模板对整个图像进行相应的清晰化处理,防止区域边界效应的产生,最后对获取的图像进行信息融合,进一步提高图像质量.通过试验证实了该算法能有效改善雾天图像的退化现象,提高图像的清晰度.The paper applies the improved genetic algorithm to threshold segmentation of the misty images and makes the image clearness.The method seeks the threshold value using genetic algorithm which segment the establishing shot and the close shot.At the same time,the moving template processing method is adopted to resolve the edge effect in the flat region,and images are fused to improve the quality of image.The experiment shows that the method proposed above is very effective to decrease the degradation of images and can enhance the clearness of images.

关 键 词：图像增强 遗传算法 图像分割 阈值 

分 类 号：TN911.73[电子电信—通信与信息系统]