摘要
介绍了储粮害虫自动检测系统的四个部分 :系统硬件组成、粮虫样本图像的处理、特征提取和模糊分类器。详细介绍了模糊分类算法的实现 ,即首先建立各类粮虫目标的模型库 ,并建立隶属函数 ,基于模糊极小极大准则 ,完成待识别粮虫的归类决策。整套系统在粮库现场验证时 ,对七类粮虫分类的正确识别率达 85 %以上 ,分类效果良好。图 3,表 2 ,参 12。
The stored-grain pests automatic detection system is introduced in this paper,which consists of the hardware constitution,the pest sample image processing,character extracting and the fuzzy classifier.the fuzzy classifying arithmetic is discussed in detail,that is ,building the data templets of all sorts of pests and the subject function,realizing the grain pests classifying according to the fuzzy minimums and maximums rule.The whole system is well in the recognition of seven kinds of grain pests,whose correct recognition ration is over 85% in grain depot.
出处
《农业系统科学与综合研究》
CSCD
北大核心
2002年第2期122-125,共4页
System Sciemces and Comprehensive Studies In Agriculture
基金
中国科学院模式识别国家重点实验室开发基金资助 (编号 :NLPR2 0 0 0 )
河南省自然科学基金资助 (编号 :995 10 0 0 8)
关键词
模糊识别技术
储粮
害虫检测
自动检测
特征提取
stoed-grain pests
automatic detection
character extracting
fuzzy recognition
