摘要
测距图像的特征提取和类别划分是计算机视觉的热点问题之一。以2D测距图像为研究对象,提出了一种加权的模糊聚类算法-wFCA算法来进行特征提取。为了自主确定准确的聚类数目,利用多种有效性索引函数对不同聚类算法的有效性进行计算评估,选取一种适合于测距图像有效性分析的索引函数。同时,为了解决聚类算法中局部最优问题,提出一种改进的IVGA遗传算法。通过相关算法的性能比较,所提方法的有效性均得以验证。
Feature extraction and classification division of ranging images is a key problem in computer vision. Considering 2D ranging images as research object, a weighted fuzzy clustering algorithm （wFCA） was proposed. To determine the accurate clustering number automatically, many validation index functions were used to estimate the validity of different clustering algorithms so as to select the most appropriate one for ranging images. At the same time, an improved genetic algorithm IVGA was proposed to solve the local optimum of clustering algorithm. By the comparison with other algorithms, the effectiveness of the algorithms is demonstrated.
出处
《系统仿真学报》
EI
CAS
CSCD
北大核心
2008年第1期76-80,共5页
Journal of System Simulation
基金
国家自然科学基金(60234030)
河南省教育厅自然科学基金(2007520019)
苏州大学江苏省计算机信息处理技术重点实验室开放基金(KJ50715)
河南理工大学博士基金(B050901)
关键词
2D测距图像
特征提取
加权模糊聚类
有效性索引
遗传算法
2D ranging images
feature extraction
weighted fuzzy clustering
validation index
genetic algorithm
