改进分量法的麦穗图像灰度-二值化处理  

Wheat spike image gray-binarization processing using improved G-component method

作　　者：吕行军[1] 韩宪忠[1] 陈含[1] 李静 王东坡 

机构地区：[1]河北农业大学信息科学与技术学院,河北保定071001 [2]保定市邮政局经营运行部,河北保定071001 [3]保定市邮政局信息技术局,河北保定071001

出　　处：《河北农业大学学报》2012年第3期112-116,共5页Journal of Hebei Agricultural University

基　　金：国家自然科学基金资助项目(60873236);河北省自然科学基金项目(F2009000653);河北省教育厅科学研究项目(2010251);河北省教育厅科学研究项目(Z2009122);河北省农村信息化工程技术研究中心资助

摘　　要：单位面积麦穗数是小麦产量预测中的一个重要参数,应用图像处理技术识别出麦穗的个数是测产的关键。本文提出改进G分量法,对图像上麦穗像素点的绿色分量值进行增强,然后灰度-二值化处理,使麦穗从背景中分隔开,并与普通分量法、最大值法和加权平均法处理后的图像进行了比较。随机选取50幅麦穗图像样本,分别使用上述几种方法处理,改进分量法比其他3种方法耗时分别降低了51.29%、70.68%和61.01%,同时后续的中值滤波用时也至少降低了11.45%。实验结果表明,改进分量法是有效的。The number of wheat spikes per-unit area is an important parameter in wheat yield prediction. And identifying the spike number using image processing technology is the key step for yield prediction. An improved G-component method to enhance the green component of spike pixel is proposed in this paper, followed by image processing by gray-binarization. Spikes from the background were easy to separate. And the images were compared with that of common component method, the maximum value method, the weighted average method, improved component method, respectively. Afterward, 50 spike images were randomly selected, and processed by the aforementioned methods. Time-consuming by the improved component method was reduced by 51.29%, 70.68% ,and 61.01% respectively compared with the other three methods. And the time used in the subsequent median filtering was also reduced by higher than 11.45%. The experimental results show that the new method is effective.

关 键 词：麦穗 灰度-二值化 改进G分量法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]