摘要
文章主要针对玉米病虫害的检测方法开展研究,利用聚类算法和中值滤波方法对玉米叶片图像进行预处理,去除阴影、光线等外界干扰因素,通过色彩空间聚类分割方法分离叶片的病斑区域,并保存病斑周边区域颜色等特性,利用颜色、纹理、形状等信息进行特征提取,以支持向量机对玉米叶片病虫害差异进行诊断,并利用网络化数据传输方式搭建数据库进行远程采集及监测。
In this paper,the detection methods of maize pests and diseases are studied.The image of maize leaves is preprocessed by clustering algorithm and median filtering method.The external interference factors such as shadows and light are removed.The lesion areas of leaves are separated by color space clustering segmentation method,and the color characteristics of the surrounding areas are preserved.The information of color,texture and shape is used to carry out the detection.Feature extraction was used to support vector machine to diagnose the difference of maize leaf diseases and insect pests,and a database was built by means of network data transmission for remote collection and monitoring.
作者
王超
张勇
王淼
陈龙
Wang Chao;Zhang Yong;Wang Miao;Chen Long(Changchun Guanghua University,Changchun 130033,China)
出处
《江苏科技信息》
2020年第5期27-28,共2页
Jiangsu Science and Technology Information
基金
吉林省教育厅“十三五”科学技术研究规划项目,项目编号:JJKH20181376KJ
2019年-长春光华学院青年科研基金项目,项目编号:QNJJZD2019003。
关键词
机器视觉
玉米病虫害
聚类算法
支持向量机
machine vision
corn diseases and insect pests
clustering algorithm
support vector machine
