基于不同背景图像下花椒锈病等级判别的研究  

Investigating the discrimination of zanthoxylum bungeanum rust disease severity under various background images

作　　者：刘延 LIU Yan(Jiangjin Meteorological Bureau,Chongqing 402260,China)

机构地区：[1]江津气象局,重庆402260

出　　处：《中国高新科技》2023年第15期139-141,共3页

摘　　要：花椒锈病作为花椒叶部的主要病害,在人为分辨时无法准确地辨别出严重程度,会造成农户过度打药现象。精准识别花椒锈病的严重程度,进而针对性地科学用药,从而达到农药减施的目的。通过背景优化处理后的Kmeans聚类分割方法对识别花椒锈病等级的准确率达90.62%,其中深色背景下准确率为95.65%,浅色背景下准确率为90.54%,复杂背景下为85.69%,分割效果理想;而加入锈病区域增强方法后,识别的准确率提升了1.41%,该方法为其他植物病害等级识别奠定了研究基础。As the main disease of prickly ash leaf,it can not accurately distinguish the severity,which will cause farmers excessive medicine phenomenon,accurately identify the severity of prickly ash rust,and then targeted scientific drug use,so as to achieve the purpose of reducing pesticide application.The means cluster classification method of 90.62%,and is 95.65%in dark background,90.54%in light background and 85.69%in complex background,and the rust area enhancement method increases by 1.41%,laying a research foundation for the identification of other plant diseases.

关 键 词：花椒锈病 图像识别 阈值分割 Kmeans算法 

分 类 号：S435.73[农业科学—农业昆虫与害虫防治] TP391.41[农业科学—植物保护]