基于图像处理技术的玉米数字图像特征值与水分含量相关性研究  

CORRELATION BETWEEN THE CHARACTERISTIC VALUE OF DIGITAL IMAGE AND MAIZE MOISTURE CONTENT BASED ON IMAGE PROCESSING TECHNOLOGY

作　　者：黄南[1] 王若兰[1] 岳佳[1] 

机构地区：[1]河南工业大学粮油食品学院粮食储运国家工程实验室,河南郑州450001

出　　处：《河南工业大学学报（自然科学版）》2016年第6期12-17,共6页Journal of Henan University of Technology：Natural Science Edition

摘　　要：通过Matlab图像处理技术对玉米数字图像进行了滤波平滑处理、图像分割处理,采集玉米籽粒数字图像的24个特征值。通过SPSS分析比较24个特征值与玉米水分含量的相关性,结果显示,玉米的颜色特征模型在3 d内预测精度高于轮廓特征模型,随着储藏期的延长,各模型预测精度均出现不同程度下降,同时轮廓特征值对于颜色空间模型具有较好的修正效果。Moisture detection of maize is the key condition for safe storage, and it is also an important basis for the application of maize in feed, food and Industry. The present study was to examine the correlation between the characteristic value of the digital image and maize moisture content. 24 eigenvalues were collected through filtering smoothing and image segmentation technology, and the relevance of the 24 eigenvalues and maize moisture content was analyzed and compared by SPSS. The results showed that the prediction accuracy of color feature model of maize was higher than the contour feature model within 3 days. Each of prediction accuracy of the feature model was declined with the extension of storage time. At the same time, the contour characteristic value had a better correction effect on the color space model.

关 键 词：图像处理 特征值 玉米水分含量 

分 类 号：TS201.2[轻工技术与工程—食品科学]