薄雾天气下图像的复原与边缘检测研究

在薄雾天气下拍摄的图像用常规的复原方法达不到较好的复原效果，从而导致图像的边缘检测效率低下．首先通过物理方法分析这种图像的成像原理，提出基于景点深度检测的物理复原方法；然后用三次B-样条函数作为光滑函数定义小波函数，通过小波变换对复原后的图像进行边缘检测．实验结果表明，在同样的薄雾天气条件下，文中方法比基于滤波的图像复原方法更有效．对采用这两种方法复原后的图像分别做边缘检测实验，结果表明文中方法的边缘检测效果明显优于基于滤波的图像复原方法．

Study on Restoration and Edge Detecting of Image in Haze Weather

By the conventional image restoration technique, it is hard to achieve a satisfactory restoration effect to the images shot in haze weather. To solve this problem, by analyzing the photo-generation principle of this kind of image in physics, a physically based restoration method based on scene point depth detection is proposed. Afterwards, a wavelet function is defined by a 3rd order B-spines as the smooth function, and the edge of the image that has been restored is detected through the wavelet transform. The experiment result shows that the method proposed in the paper is more efficient than that based on filtering for image restoration, and the edge detection result also shows the superiority of the restoration method proposed in the paper.