Deblurring Texture Extracted from Digital Aerial Images by Reforming “Step Edge” Curve

A method, based on reforming the “step edge” curve which clearly explains how the diffraction of sunlight makes digital aerial images blur, is proposed to deblur the texture extracted from digital aerial images and thus to improve its visual effect. First, histogram analysis and edge detection are integrated with each other to automatically find out the“step edge” curve for each line feature in texture. In this step, highly precise Canny algorithm is used to locate the central point of “step edge” curve and the two boundary points of “step edge” curve are determined by tracing the intensity histogram around located central point of “step edge” curve. As a result, “step edge” curve could be automatically precisely gained. After that, the selected “push” function anti-tangent is used to adjust the intensity of every the “step edge” curve and make it more steep around the located central point. Last, several textures extracted from digital aerial image are tested by the proposed strategy and algorithm, and the experimental result is encouraging.