基于可拓神经网络的农作物叶部病害图像识别  

Crop leaf diseases image recognition based on extension neural network

作　　者：张飞云[1] 陈静[1] 

机构地区：[1]许昌学院电气(机电)工程学院

出　　处：《电子世界》2017年第12期123-124,127,共3页Electronics World

基　　金：河南省高等学校重点科研项目计划(15A510042;16A510031);2016年度许昌市科技计划项目

摘　　要：以小麦、玉米、大豆等农作物作为研究对象,分别提取每个农作物叶部病害的颜色、形状及纹理特征参数,组成多特征参数,采用可拓神经网络对农作物叶部病害进行分类诊断识别。通过实验表明,该系统对小麦、玉米、大豆等农作物叶部病害的识别率可达到90%以上。The wheat, corn, soybean and other crops were taken as the research object, and the color, shape and texture parameters of leaf diseases were extracted respectively to combine multi feature parameters, then the extension neural network were used to recognize crop leaf disease. Experimental results showed that the recognition rate of the algorithm on wheat, corn and soybean leaf diseases could reach more than of 90%.

关 键 词：农作物病害 可拓神经网络 模糊熵 多阈值图像分割 图像识别 

分 类 号：S436.36[农业科学—农业昆虫与害虫防治]