Research On Technique In Edge Detection

Edge is the local inconsecutive character of an image. it belongs to the luminance (color or texture)changed more intense place, located between object and background, object and object, background and background, element and element. In a word, it's the fundamental character of an image. A lot of information was contained on the edge, so edge detection is the difficult point and hotspot in research on Image analysis and pattern recognition as the key step.The article start from introducing some basic theory and available algorithm on edge detection, analyzed the difficult point of edge detection, according to the problem around current edge detection algorithm, the study job mainly embodies in the several following aspect:In order to solve the problem between filtering and reserving the detail of the edge, we use the filtering algorithm based on grey correlative degree. Besides in order to obtain an ideal Location of edge, the Krisch operator was used to correct edge position, improved the detecting effects.According to visual perception the HSV color space was defined renewably, so this new HSV color space was more close to human visual perception. In addition, the article modified the conversion relationship between RGB color space and HSV color space to make it run faster.The part of color image edge detection, firstly, we introduced two mainstream methods and analyzed their Defects. In order to make full use of color, we put forward a new color image edge detection Thoughts that is use the three color components, improved the Classical color image edge detection algorithm, moreover, we expanded the edge detection algorithm based on gray correlation degree to color image. At last the new color image edge detection algorithm based on gray correlation degree and human visual perception was put forward.