基于图像内容的常见蔬菜病害诊治研究  

Diagnose of Common Vegetables Disease Based on Image Content

作　　者：相林[1] 侯潮[2] 孙敏[3] 

机构地区：[1]淮阴工学院计算机与软件工程学院,江苏淮安223003 [2]淮阴工学院数理学院,江苏淮安223003 [3]淮阴工学院生命科学与食品工程学院,江苏淮安223003

出　　处：《淮阴工学院学报》2016年第3期1-5,共5页Journal of Huaiyin Institute of Technology

基　　金：淮安市科技支撑计划(农业)项目(SN13045;SN13046);江苏省高等学校大学生创新创业训练计划项目(201511049037Y)

摘　　要：研究计算机视觉技术在蔬菜病害诊治中的应用,构建黄瓜、番茄等常见蔬菜的病害图片库,对图片库中的图片和待检测图片进行预处理,并使用PCA算法对图像提取特征,得到蔬菜病害的特征库;提取待检测的病害图片的特征,使用余弦相似度与病害特征库比较,检测出病害图像。实验结果表明,将PCA算法应用到蔬菜病害的图像识别,有较好的识别效果。A computer visional technology was applied to diagnose vegetable diseases in this paper. Firstly, an image database was constructed, which was consisted of common vegetables such as tomatoes, cucumbers. Secondly, PCA algorithm was used to extract image features so as to obtain a database of disease features after images in database were preprocessed. Lastly, we extracted the feature of image to compare with that in disease database based on cosine similarity to recognize the disease image. The experimental results showed that PCA was effective to recognize images of vegetable diseases.

关 键 词：病害识别 特征提取 PCA 余弦相似度 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]