基于颜色特征的玉米种子纯度识别  

Purity identification of maize seed based on color characteristics

作　　者：闫小梅[1] 刘双喜[1] 张春庆[1] 王金星[1] 

机构地区：[1]山东农业大学机械与电子工程学院,泰安271018

出　　处：《农业工程学报》2010年第S1期46-50,共5页Transactions of the Chinese Society of Agricultural Engineering

基　　金：2009山东省博士后创新基金(72008)

摘　　要：为准确快速的识别出玉米种子中的杂粒,提高玉米种的纯度,该文提出一种以玉米种子冠部与侧面颜色作为特征向量进行纯度识别的新方法。该方法首先将玉米种子原始图像进行背景分割、单粒提取,然后进行冠部核心区域及侧面RGB、HSV颜色特征向量的提取,最后采用Fisher判别理论将多维特征向量投影到一维空间中,进行K-均值聚类分析。试验结果证明:利用Fisher判别理论在一维空间上进行K-均值聚类分析,玉米种子纯度的识别率高于93.75%。影响玉米种子正确识别率的主要因素是投影方向的选择及正确的冠部核心区域的提取。In order to identify miscellaneous seed from maize seed accurately and rapidly, maize seed purity identification method based on color extracted from the images of both the maize crown and the maize side was proposed for improving maize seed purity. First, segmentation and single extraction were carried on the original image; then the color models RGB and HSV were used to extract various color features from the maize crown and the maize side; finally, multidimensional eigenvectors were projected into one-dimensional space through applying fisher Discriminant Theory and K-means algorithm was carried on the new color space. The experimental results show that K-means algorithm based on one-dimensional space received through Fisher Discriminant Theory can effectively identify maize seed purity, and the recognition rate isover 93.75%. The main factors that affect recognition are projection direction and the correct choice of maize crown center region.

关 键 词：种子 颜色 识别 Fisher判别理论 K-均值聚类 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]