基于机器视觉的干制红枣大小分级方法研究  

The Method of Automatic Dried Red Jujube Hierarchy Based on Machine Vision

作　　者：李景彬[1] 邓向武[1] 坎杂[1] 田绪顺[1] 谢凡[1] 

机构地区：[1]新疆石河子大学机械电气工程学院,新疆石河子832000

出　　处：《农机化研究》2014年第2期55-59,共5页Journal of Agricultural Mechanization Research

基　　金：国家科技支撑计划项目(2011BAD48B04)

摘　　要：为了实现干制红枣的大小自动分级,介绍了应用机器视觉的干制红枣自动分级方法,利用CCD摄像机获取红枣的样本图像,应用MATLAB软件编程实现了样本图像的灰度化、二值化、图像分割、图像滤波、图像形态学处理、边缘检测和特征量提取等处理,参照干制红枣分级标准完成了红枣自动分级。通过实验数据回归分析得出红枣实测纵径、果质量与识别值当量值之间的数学检测模型,其决定系数分别达0.995 5和0.948 1。实验表明,采用句法模式识别对数据进行红枣大小分级,分级准确率达85%。In this paper, in order to realize the automatic classification of the size of the red jujube. The machine vision method was adopted in automatic red jujube hierarchy. The sample image was shot by a CCD camera and a series oftreatments were proposed, with the sott ol MATLAI5, mcluOmg graying, Dreary treatment ,image segmentation,image filter, image morphology processing, edge detection, the picking up of eigenvalues and so on. According to the standard of the red jujube hierarchy accomplished the automatic red jujube hierarchy. Through testing and analysis, we got the math- ematieal model for the relationship between measured fruit longitudinal diameter, fruit weight and recognition fruit equiva- lent value, and the R-square is 0. 9955 and 0. 9481. The date of red jujube＇ s size was graded by syntactic pattern rec- ognition. The results indicated that accuracy of classification was 85 %.

关 键 词：干制红枣 机器视觉 分级 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] S126[自动化与计算机技术—计算机科学与技术]