基于颜色和形状特征的黄瓜霜霉病自动识别研究  

Automatic recognition of cucumber downy mildew based on color and shape feature

作　　者：王雪[1] 马卓[1] 王欣[1] 姜丹[1] 刘小溪[1] 贾志雷[1] 

机构地区：[1]吉林省科学技术信息研究所,长春130021

出　　处：《安徽农业大学学报》2013年第6期1071-1075,共5页Journal of Anhui Agricultural University

基　　金：吉林省世行贷款农产品质量安全项目(2011-Z20)资助

摘　　要：利用计算机图像处理技术,对农作物病害的自动识别做了实验性研究。以黄瓜霜霉病为例,采集了黄瓜霜霉病的数字图像作为试验样本,进行了灰度化处理、二值化处理、平滑化处理和边缘检测处理等,从病斑的形状和颜色两方面提取特征参数,利用k近邻法对样本进行分类识别,识别率达到95%。In this paper, automatic identification of cucumber downy mildew was conducted using computer image processing. The digital image of the cucumber downy-mildew was collected, and then treated with gray processing, binary processing, smoothing processing and edge detection processing. Characteristic parameters were extracted from the two aspects of shape and color of the lesion, and then k-nearest neighbor method was used for classification and identification. The recognition rate was up to 95%.

关 键 词：计算机图像处理 霜霉病 K近邻法 模式识别 

分 类 号：TP391.411[自动化与计算机技术—计算机应用技术]