小波模极大值边缘检测算法研究  

Wavelet Modulus Maximum Algorithm in Image Edge Detection

作　　者：祝强[1] 徐瑾[2] 张铎[1] 张建华[3] 

机构地区：[1]西北工业大学航天学院,陕西西安710072 [2]武汉理工大学机电工程学院,湖北武汉430070 [3]西北工业大学机电工程学院,陕西西安710072

出　　处：《武汉理工大学学报（信息与管理工程版）》2008年第6期849-853,共5页Journal of Wuhan University of Technology：Information & Management Engineering

基　　金：教育部博士学科点专项科研基金资助项目(20020699014)

摘　　要：边缘包含图像的本质信息。通过分析具有不同特性的小波模极大值边缘检测结果,给出了小波高通滤波器序列的构成、长度以及对称性对小波细节系数在方向、强度和形状方面产生的影响;提出了用于模极大值边缘检测的小波高通滤波器应具有单调方向性和偶对称性。对小波模极大值边缘检测算法中的阈值选择、多尺度边缘融合进行了改进,仿真测试验证了该方法的有效性。Edge involves the essential information of image. Based on the analysis of the different wavelets in image edge detection, the construction and the length of the wavelet high - pass filter coefficient were presented. The effect of the symmetries to the wavelet high - pass filter coefficient in directions, strength and shapes were also provided. The idea was proposed that the wavelet high - pass filter used in image edge detection should possess monotonicity and even symmetry. The threshold selection and the multi - scale edge incorporation in image edge detection were improved. The simulation tests verify the validity of the proposed method.

关 键 词：边缘检测 模极大值 阈值 滤波器 对称性 

分 类 号：TN274.2[电子电信—物理电子学]