基于粒子群模糊聚类算法的边缘检测仿真

Simulation on a New Algorithm Based on PSO Fuzzy Clustering for Image Edge Detection

将粒子群优化算法与模糊C-均值(FCM)聚类算法相结合,并应用于图像边缘检测,以期解决标准FCM算法在图像边缘检测中对初始值敏感及容易陷入局部极小的两大缺陷.首先,基于数学测度概念构造一个描述边缘点信息的特征向量,将灰度图像中的每一个像素点看成是一个数据样本,将该点灰度值处理后构成其边缘点信息特征向量,形成具有三维特征的数据集;然后对这个数据集应用粒子群模糊聚类算法进行分类,自适应地检测出图像的边缘点,达到提取边缘的目的.仿真实验表明,此算法具有良好的抗噪性能,能够得到较好的边缘效果,提高了边缘定位的精度.

The PSO （particle swarm optimization） and fuzzy C-Mean （FCM） algorithms were combined together to form a new algorithm and it is applied to image edge detection, thus overcoming the two shortcomings of standard FCM algorithm, i.e., sensitive to initial value and easy to fall to local minimum. The new algorithm is developed the way an eigenvector is constructed on the basis of measure theory to describe an edge point information, and each of the pixel points in a gray scale image is regarded as a data sample. The eigenvector of the information on an edge point is constructed by processing the gray level of the pixel point, and a 3-D data set is thus given. Then, the data set is classified by PSO fuzzy clustering algorithm to adaptively detect the image edge points so as to extract an edge. Simulation results showed that the new algorithm is highly antinoise and able to get better image edges with improved precision in edge positioning.