基于K-means聚类算法的叶螨图像分割与识别  

A method for image segmentation and recognition of spider mites based on K-means clustering algorithm

作　　者：刘国成[1] 张杨[1] 黄建华[2] 汤文亮[3] 

机构地区：[1]广州铁路职业技术学院,广州510430 [2]江西省农业科学院植物保护研究所,南昌330200 [3]华东交通大学软件学院,南昌330013

出　　处：《昆虫学报》2015年第12期1338-1343,共6页Acta Entomologica Sinica

基　　金：江西省科技支撑计划农业项目(20122BBF60103);江西省科技支撑计划农业项目(20132BBF60083);广东省自然科学基金培育项目(GTXYP1310)

摘　　要：【目的】叶螨（spider mite）是为害多种农作物的主要害虫,叶螨识别传统方法依靠肉眼,比较费时费力,为研究快速自动识别方法,引入计算机图像分析算法。【方法】该方法基于K-means聚类算法对田间作物上的叶螨图像进行分割与识别。【结果】对比传统RGB彩色分割方法,K-means聚类算法能够有效地对叶片上叶螨图像进行分割和识别。K-means聚类算法平均识别时间为3.56 s,平均识别准确率93.95%。识别时间T随图像总像素Pi的增加而增加。【结论】K-means聚类组合算法能够应用于叶螨图像分割与识别。[ Aim ] The spider mites are the main pests of many crops. Traditional recognition methods for spider mites relied on the naked eyes, which wasted a lot of time and energy. In order to study the fast automatic recognition method for spider mites, a method using computer image analysis algorithm was developed. [ Methods ] The method based on the K-means clustering algorithm realized the segmentation and recognition of the spider mite images which were obtained from fields. [ Results ] In contrast to the traditional RGB color segmentation method, the K-means clustering algorithm method was able to separate the images of spider mites from leaf background effectively. The average recognition time based on the K- means clustering algorithm was 3.56 s, and the recognition accuracy was 93.95 %. The recognition time （T） increased as the pixels of tested image （Pi） increased. [ Conclusion ] The method can be applied to the segmentation and recognition of spider mite images.

关 键 词：叶螨 图像 K-MEANS算法 图像分割 图像识别 像素 

分 类 号：Q969[生物学—昆虫学]