Seeds recognition method based on image of peanut pods
In order to verify tile feasibility of peanut pods Varieties recognition and DUS test by machine vision and image processing, Based on the measurement of 50 characteristics in three categories of these peanut images,20 peanut varieties of images were collected by Scanner.Comprehensive used statistical characteristics,PCA data optimization, neural networks, support vector machine, clustering analysis, and other methods, A group of meaningful & optional features to BUS test have been selected.The PCA optimization SVM model is better then neural network, else, and identify effect is stablly.Varieties recognition rate is up to 90%.In addition,the recognition correct rare of different origins is reach 100% by this model.Based on these features, the pedigree clustering tree of 20 peanut varieties was built, and then the paternity of partial varieties has be inspected.The results of this study have some certain reference value to the selection of DUS traits, peanuts varieties and origin of recognition,and peanut pedigree research.
Peanuts Variety recognition Principal component analysis Artificial neural network Support vector machine DUS test
HAN Zhong-zhi ZHAO You-gang LI Yan-zhao
College of Information Science & Engineering; Qingdao Agricultural University, Qingdao266109, Shandong, China
国内会议
第二届全国精准农业与农业空间信息技术学术交流暨农业信息化智能化科技成果推介会
深圳
英文
40-51
2013-05-24（万方平台首次上网日期，不代表论文的发表时间）