基于图像分析的病虫识别研究  

Research on Insects Recognition Based on Image Analysis

作　　者：宗精学[1] 杨余旺[1] 赵炜[1] 王磊[1] 莫然[1] 

机构地区：[1]南京理工大学计算机科学与工程学院,南京210094

出　　处：《科学技术与工程》2014年第19期194-200,共7页Science Technology and Engineering

基　　金：江苏省农业科技自主创新基金资助项目(CX(10)221-1);江苏省科技支撑计划基金资助项目(BE2011342-1;BE2012386)资助

摘　　要：为比较应用图像分析技术识别病虫的效果,选取菜蛾、菜粉蝶、玉米螟、中华稻蝗、竹云纹野螟和稻纵卷叶螟共6种病虫的图片进行分类识别。利用图像技术对六种病虫的图片依次进行病虫图像预处理、病虫背景区分和病虫特征提取工作,最后进行两级分类识别。结果表明,经过处理后的病虫图像在两级分类识别中可以达到较高的识别率。同时表明,在农作物病虫害识别与防治领域应用图像分析技术可以提供迅速、可靠、精确和非破坏性的分析诊断识别方面的技术支持。In order to compare the effect of insects recognition based on image analysis,the images of plutellidae,pierisrapaeLinne,pyrausta nubilalis,oxya chinensis,demobotys pervulgalis and rice leaf roller were selected for classification and identification.The images was prepressed,insects was distinguished from the background and extracted inserts＇ feature information from the image.At last,the pests were identificated.Experiments show that,it can achieve high recognition rate on two level identification based on image analysis.Thus,it could be seen that,applying the technology of image analysis can provide efficient,reliable and accurate technical support for crop pest recognition and prevention,in aspects of non-destructive analysis,diagnosis and identification.

关 键 词：图像分析 病虫 分类识别 

分 类 号：S763.301[农业科学—森林保护学]