用人工神经网络进行果实颜色分级技术研究  

Color Classification of Fresh Fruits by Neural Network

作　　者：何东健[1] 杨青[1] 薛少平[1] 耿楠[1] 

机构地区：[1]西北农业大学机械与电子工程学院

出　　处：《西北农业大学学报》1998年第6期109-112,共4页Journal of Northwest Sci-Tech University of Agriculture and Forestry（Natural Science Edition)

基　　金：杨凌农业科研基金

摘　　要：以计算机视觉自动检测果实表面着色度并进行分级为目的，在分析果实表面颜色色相分布特性的基础上，提出将果实色相分布曲线作为模式处理，用人工神经网络进行果实表面颜色分级。结果表明，人工神经网络分级与人工分级的一致度在９４％以上。The purpose of this paper is to improve automatic techniques of fresh fruit color inspection and color sorting by computer vision.Based on the analysis of apple hue distribution features and taking the hue distribution curves as patterns, the method of color classification by artificial neural network was developed.The results showed that computer vision classification of test apples agreed with manual grading, both above 94%.

关 键 词：果实 颜色分级 人工神经网络 计算机视觉 

分 类 号：S226.5[农业科学—农业机械化工程]