| 改进的粒子群算法用于图像分割 |
| |
| 引用本文： | 苏彩红,吴菁,朱学峰. 改进的粒子群算法用于图像分割[J]. 佛山科学技术学院学报(自然科学版), 2007, 25(3): 15-18 |
| |
| 作者姓名： | 苏彩红 吴菁 朱学峰 |
| |
| 作者单位： | 佛山科学技术学院,自动化系,广东,佛山,528000;华南理工大学,自动化科学与工程学院,广东,广州,510640 |
| |
| 摘 要： | 提出了一种将保证收敛粒子群算法与最大类间方差法相结合的快速阈值分割法。该方法根据最大类间方差法的原理以分离度大小作为判断粒子优劣的准则,即分离度越大粒子就越好,并采用粒子群算法对图像进行多目标优化搜索。实验表明,该算法在继承标准粒子群算法易于实现、实时性好等优点的同时,还避免了标准PSO算法存在的早熟收敛问题,具有更强的寻优能力。
|
| 关 键 词： | GCPSO 最大类间方差法 阈值分割 |
| 文章编号： | 1008-0171（2007）03-0015-04 |
| 修稿时间： | 2007-03-27 |
Image segmentation with Otsu based on improved PSO |
| |
| SU Cai-hong,WU Jing,ZHU Xue-Feng. Image segmentation with Otsu based on improved PSO[J]. Journal of Foshan University(Natural Science Edition), 2007, 25(3): 15-18 |
| |
| Authors: | SU Cai-hong WU Jing ZHU Xue-Feng |
| |
| Affiliation: | 1. Department of Automation Engineering, Foshan University, Foshan 528000 China; 2. South China University of Technology,Guangzhou 510640,China |
| |
| Abstract: | Threshold method is one of the most basic and important techniques for image segmentation because of its simple-realization and good result.This paper proposes a rapid threshold method which combines the GCPSO with Otsu.This method uses the separation level to judge whether a particle is good or bad based on Otsu principle,and the GCPSO is designed to optimize the decision variables.The experiment shows that this algorithm retains the simpleoperation and good real time of standard PSO,and also solves the premature convergence problem,thus it has better search capacity. |
| |
| Keywords: | GCPSO Otsu thresholding segmentation |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
