基于改进的ncspso-afsa对svm参数的优化及其应用 冯哲 陈云凤 .docVIP

基于改进的NCSPSO-AFSA对SVM参数的优化及其应用 冯哲 陈云凤 周宇 云挺 邓玉和 (南京林业大学信息科学技术学院 江苏 南京 210037) (南京林业大学材料科学与工程学院 江苏 南京 210037) 摘 要：为了找到支持向量机(SVM)最佳的分类参数，用以构建适合纹理图像分割的SVM分类器，本文将基于小生境和交叉选择算子的粒子群算法(NCSPSO)引入变异算子和族外竞争机制加以改进后与人工鱼群算法(AFSA)混合，提出了一种改进的NCSPSO-AFSA混合算法优化支持向量机参数，并分别与AFSA算法，粒子群算法(PSO)，NCSPSO算法在图像分割准确率、参数寻优时间、图像分割时间等方面进行对比和分析，实验表明本文的算法能够更好的获得适用于纹理图像分割的SVM参数，在缩短图像分割时间的同时提高了图像分割准确率，相比较其他算法，本文算法稳健性更好。将本文方法应用到电镜及超声纹理图像分割中能较好的提取出目标区域，图像边缘部分的分类也很清晰。 关键词：NCSPSO算法；人工鱼群算法；支持向量机；图像分割 Parameter Optimization and Application of SVM Based on Improved NCSPSO and AFSA Feng Zhe Chen Yunfeng Zhou Yu Yun Ting Deng Yuhe (College of Information Science and Technology, Nanjing Forestry University, Nanjing 210037, Jiangsu, China) (College of Materials Science and Engineering, Nanjing Forestry University, Nanjing 210037, Jiangsu, China) Abstract: To find the best parameters of SVM and construct the SVM Classifier which is suitable to be applied to texture image segmentation, the paper improves niche and cross-selection operator PSO where are introduced into mutation mechanism and group competition mechanism, and combines with artificial fish-swarm algorithm (AFSA), the parameter optimization algorithm for Support Vector Machine (SVM) is proposed. Compared with AFSA, particle swarm optimization and NCSPSO algorithm on accuracy and time of image segmentation and parameter optimization time, the results show that the proposed algorithm can effectively find the parameters of SVM, it cuts time and improves accuracy of image segmentation, at the same time, its stability is better than other algorithms. When it is applied to the electron microscope image and ultrasonic image segmentation, the method can extract the target area and the image edges are classified quite clear. Keywords: NCSPSO; AFSA; Support vector machine; Image segmentation 0 引言 图像分割是图像处理与计算机视觉领域中最为基础和重要的问题之一，图像分割的效果将直接影响到后续分析、识别和解释等处理。纹理是图像的重要特征，普遍存在于各类图像当中，由于纹理图像自身的复杂性，使得纹理图像的分割显得尤为困难。 支持向量机(Support Vector Machine, SVM)是一种基于统计学习理论