一种图像边缘检测算法的改进和实现

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。

Improvement and Implementation for an Image Edge Detection Algorithm

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.