基于改进型模糊聚类算法的植物病斑检测  

Study on detection algorithm of plant disease spot based on improved fuzzy clustering algorithm

作　　者：冯登超[1] 杨兆选[1] 乔晓军[2] 

机构地区：[1]天津大学电信学院,天津300072 [2]国家农业信息化工程技术研究中心,北京100089

出　　处：《计算机工程与应用》2007年第24期203-204,245,共3页Computer Engineering and Applications

基　　金：农业科技成果转化资金项目(No.05EFN211100002);北京市科技计划(the Key Technologies R&D Program of Beijing City;China under Grant No.No.Z0005190040831)联合资助

摘　　要：针对植物病害图像成分复杂、病斑排列无规则等特点,提出了一种改进型模糊聚类的病斑检测算法。该算法采用Markov随机场与模糊聚类算法耦合策略,能够有效解决植物病斑检测时的模糊性和随机性问题。仿真实验表明,改进后的算法能够实现植物病斑的自适应检测,鲁棒性较好。然而,对于Markov与模糊聚类算法的最佳耦合方式及对于如何减少算法的运算量仍需作深入的研究。According to the characteristics of complex components of plant disease images and random alignment of disease spot, this paper introduces a new detection algorithm with improved Fuffzzy clustering algorithm.By adopting the coupled method of Markov random field and fuzzy clustering algorithm,this algorithm avoids the fuzziness and randomness during the plant disease detection.Simulation experiment shows that the improved algorithm has better robust and can realize the self-adaptive detection of plant disease spot.However,how to reduce the operating quantum of the method and get the optimal coupling method between Markov random field and fuzzy clustering algorithm need further research.

关 键 词：植物病斑模糊C均值聚类Markov随机场隶属度函数 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]