摘要
水果分类在超市导购中有较为重要的应用价值,应用自动化识别技术能够实现大量水果的高效分类。主成分分析可以从多元事物中解析出主要影响因素,从而利用一个或几个较好的综合指标来概括信息。本研究采用基于主成分分析的识别分类技术对10种外形相似的水果进行预处理、训练和识别,得到了较好的识别效果,各种水果的平均识别率达到93%以上,基本能够满足实际应用。
Fruit classification is important for supermarket shopping guiding,which can be achieved efficiently by automatic recognition technology. The principal component analysis can separate out the main influencing factors from multiple things,and then use one or several better comprehensive indexes to summarize information. In this paper,10 kinds of shape similar fruits were pretreated,trained and recognized by recognition and classification technologies based on the principal component analysis,and better recognition effects were obtained. The average recognition rate of various kinds of fruits were more than 93%,which indicated that this method could meet the actual application.
出处
《山东农业科学》
2015年第8期116-118,共3页
Shandong Agricultural Sciences
基金
河南省科技计划重点项目(112102210333)
平顶山学院青年科研基金(PDSU-QNJJ-2013005)
关键词
水果
分类
主成分分析
外形相似
自动化识别
Fruit
Classification
Principal component analysis
Shape similarity
Automatic recognition
