摘要
利用前人提出的一种雾天图像清晰化算法能得到较清晰的雾天降质图像.该算法也不可避免地存在着一定的缺陷,如处理图像所需要的时间过长,处理后的图片会出现增强过度等情况.基于上述不足之处,本文主要研究的是在此基础上给出具有针对性的解决方案.首先,在具体操作的过程中,我们需要将这两大进程——对天空部分的判断和移动模板直方图的均衡化相分离,要把天空部分率先分离出来,必须要保证模板内已经没有属于天空部分的点,这样之后才能执行直方图均衡化这一操作.其次,我们将移动模板的步长拉大.为了验证算法改进的有效性,我们在MATLAB平台上做了大量的实验,根据实验结果可以看出,改进算法不仅使图片的清晰化程度提高了,而且处理图像花费的时间也略有减少.
The fog image sharpening algorithm which was put forward by predecessors succeeded in making part information of depth of field suited through the way of movable platen.Thus,we can obtain clearer fog degraded images.This algorithm also has some defects inevitably,for exzample, disposing images needs a lot of time,excessive enhancement and other phenomena may appear in disposed images.In accordance with the defects we have mentioned before,based on the predecessors' algorithm,the thesis mainly research a solution which has pertinence.First,in the process of concrete operations,we need make the two courses——the judgement on the area of sky and the process of histogram equalization for movable platen apart,we should separate the area of sky first,we must ensure that there is no point belonging to the area of sky in the formwork,then we can operate histogram equalization.Second,we amplify the step length of the movable platen. In this thesis,in order to vertify the effectiveness of the modified algorithm,we have conducted lots of experiments on MATLAB table,based on the experimental result,we can see that the modified algorithm can not only improve the clearness degree of the image,but disposing image spends less time.
出处
《河北建筑工程学院学报》
CAS
2013年第3期70-74,共5页
Journal of Hebei Institute of Architecture and Civil Engineering
