图像识别处理技术在农业工程中的应用  

Application of image recognition processing technology in agricultural engineering

作　　者：周清松[1] 唐秀忠[1] ZHOU Qingsong;TANG Xiuzhong(School of Technology,Puer University,Puer 665000,China)

机构地区：[1]普洱学院理工学院,云南普洱665000

出　　处：《现代电子技术》2017年第4期107-110,共4页Modern Electronics Technique

基　　金：国家"863"计划基金资助项目(2014-01Z338);云南省教育厅基金(2014Y509)

摘　　要：研究一种基于图像识别处理的粮虫检测方法,将图像识别处理技术应用于农业工程。首先使用灰度化、二值化、平滑以及锐化技术对粮虫图像进行预处理,使得图像更容易进行边缘检测和图像特征提取。之后使用四种边缘检测方法实现图像中待识别粮虫的边缘检测。使用粮虫图像的8种区域描述子特征作为粮虫识别模型的输入特征。最后选取常见的玉米象、拟谷盗和锯谷盗三种粮虫为研究对象,使用基于RBF神经网络的识别模型对三种粮虫图像的几何形态特征进行识别。A grain insect detection method based on image recognition processing is studied. The image recognition processing technology is applied to the agricultural engineering. The gray processing,binaryzation,smoothing and sharpening techniques are used to preprocess the grain insect image to make it easier to conduct the edge detection and image feature extraction. The commonly?used four edge detection methods are adopted to detect the edge of the grain insect image under recognition. Eight re?gional descriptor features of the grain insect image are taken as the input characteristics of the grain insect recognition model.Three familiar grain insects of maize weevil,tribolium and coleoptera are selected as the research objects,and the identification model based on RBF neural network is used to recognize the geometric feature of three grain insect images.

关 键 词：粮虫检测 特征提取 RBF神经网络 图像识别 

分 类 号：TN911.73-34[电子电信—通信与信息系统] S24[电子电信—信息与通信工程]