基于改进FCM的作物病害图像分割方法研究  

Research on crop disease image segmentation based on improved FCM

作　　者：曹晓丽[1] 齐国红[1] 井荣枝[1] CAO Xiaoli;QI Guohong;JING Rongzhi(SIAS Intemational University,Zhengzhou University,Xinzheng Henan 451150,China)

机构地区：[1]郑州大学西亚斯国际学院,河南新郑451150

出　　处：《智能计算机与应用》2018年第5期51-53,59,共4页Intelligent Computer and Applications

基　　金：河南省科技厅基础与前沿技术研究计划项目(182102210546);郑州大学西亚斯国际学院校级科研项目(2017YB12;2017YB13);河南省教育厅第九批河南省重点学科(检测技术与自动化装置(教高[2018]119号))

摘　　要：针对复杂背景下作物病害叶片分割问题,提出一种改进模糊C均值聚类(Fuzzy C Means,FCM)的作物病害图像分割方法。该方法综合考虑图像的像素点的局部空间信息和灰度信息,计算出更为准确的局部空间信息,减少噪声的同时更好地保留了图像细节,从而使图像分割效果更为精确。选取黄瓜病害叶片图像进行本文算法的验证,并与其它分割方法进行比较实验。实验结果表明,本文提出的方法分割效果更好,其分割正确率为97.81%。Focused on the problem of the lowsegmentation accuracy,a disease segmentation method is proposed based on improved FCM（ Fuzzy C Means）. The local spatial information and gray level information of the pixels of the image are considered,and more accurate local spatial information is calculated,which could reduce noise,better preserve the details of the image,thereby make the image segmentation effect more accurate. By Choosing the cucumber disease leaves to verify the algorithm of this paper,and compared with other segmentation methods,the experimental results showthat the proposed method is effective and its segmentation accuracy rate is 97.81%.

关 键 词：模糊C均值聚类 局部空间信息 分割正确率 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]