结合OSTU阈值法的自适应DCP图像优化算法

Study on Adaptive DCP Image Optimization Algorithm Combined with OSTU Threshold Method

图像是获取信息的重要途经,对于雾霾天气导致的可见度低的图像,DCP理论算法能进行去雾处理,但结果会出现光晕现象,缺少部分特征信息,存在一定不足。为了解决该问题,提出一种融合OSTU阈值法的自适应DCP图像优化算法,利用改进的OSTU阈值算法分割图像的前景与背景;基于心理灰度公式估计大气光值,利用最小值滤波与中值滤波融合最佳阈值估计透射率,得到自适应的透射率估计值,利用双指数滤波器进行精细化处理;最后基于DCP理论得到去雾后的图像。结果表明算法的去雾效果明显,能有效地保护边缘信息,恢复图像的细节特征,算法处理效果具有较大优越性。

Image is one of the important ways to obtain information.For the low visibility image caused by fog weather,DCP theory algorithm can make defogging,but halo phenomenon will appear,lack of some feature information will occur,there are some shortcomings.In order to solve this problem,an adaptive DCP image optimization algorithm based on Ostu thresholding method is proposed.Firstly,the foreground and background of the image are segmented by using the improved Otsu thresholding algorithm;secondly,the atmospheric light value is estimated based on the psychological gray formula;then the transmittance is estimated by using the minimum filter and the median filter to fuse the optimal threshold,and the adaptive transmittance estimation value is obtained.The dual-exponential filter is used to make refinement.Finally,the defogging image is obtained based on DCP theory.The results show that the defogging effect of the algorithm is obvious,and it can effectively protect the edge information and restore the detailed feature of the image.The algorithm has great superiority in processing effects.