Multi-scale Images Edge Detection Method Based on Gap Operatorof Weighted Neighborhood
多尺度加权邻域Gap算子图像边缘检测方法

Extracted the method called "Gap statistic" proposed by Hastie T. and Tibshirani, brought forward a con- cept called regular weighted neighborhood. Relying on the gray-level distribution in the relative half-plane of image pixels with weight, a multi-scale edge detection model based on Gap operator was established. The paper took the weight probability density function of regular weighted neighborhood, analyzed the character of Gap operator, opti- mized tbe edge detection algorithm based Gap statistic model, analyzed the correlation between the Gap operator and two the Sobel operator, discussed the anti-noise and multi-scale property of the edge detection model, and an investi- gation is taken to compare the difference of edge detection under the circumstance of different scales.