一种图像边缘检测算法的改进和实现  

Improvement and Implementation for an Image Edge Detection Algorithm

作　　者：汪昡紫 孙宪坤[1] 刘锴[1] 

机构地区：[1]上海工程技术大学电子电气工程学院,上海201620

出　　处：《计算机技术与发展》2014年第9期108-111,136,共5页Computer Technology and Development

基　　金：上海市重点创新项目(14zz156);上海工程技术大学学科建设专项基金(xkcz1212);上海工程技术大学"十二五"内涵建设项目(B-8932-13-0124)

摘　　要：边缘检测是图像分割和模式识别的必要工作。首先分析了传统的导数算子Sobel和Canny的检测原理及其优缺点。然后针对图像边缘检测的特点,从模糊聚类角度出发,提出一种改进的蚁群算法。根据图像灰度和梯度特征设置初始聚类中心,改进启发式函数,将蚁群算法得到的聚类中心作为模糊C均值聚类的初始中心,再进行FCM聚类,实现基于目标函数的模糊聚类。最后对文中提到的各种算法的实验结果进行比较与分析,结果表明文中改进算法是有效的。Edge detection is necessary for image segmentation and pattern recognition. In this paper, analyze the detection principle and advantages and disadvantages of traditional operators like Sobel and Canny. Then, aiming at the features for image edge detection, propose an improved ant colony algorithm from the aspect of fuzzy clustering. Set the center of initial clustering according to the image gray and gradient feature for improving heuristic function, taking the clustering center obtained by ant colony algorithm as the initial center of FCM to cluster, realize the fuzzy clustering based on the object function. At last, make a comparison and analysis for the results of different algorithms, which proves the efficiency of the improved ant colony algorithms.

关 键 词：蚁群算法 模糊聚类 导数算子 边缘检测 特征提取 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]