作物图像光照亮度补偿方法  

Illumination Compensation Algorithm for Crop Image

作　　者：李锦卫[1] 廖桂平[1] 

机构地区：[1]湖南农业大学信息科学技术学院农业信息研究所,长沙410128

出　　处：《农机化研究》2012年第8期26-29,33,共5页Journal of Agricultural Mechanization Research

基　　金：国家自然科学基金项目(31071328)

摘　　要：图像间光照亮度不一致问题是影响作物叶色-氮营养诊断精度的因素之一。针对此问题,提出基于灰度梯度的光照亮度补偿算法、基于十色模型的光照亮度补偿算法以及组合算法。166幅室内外水稻和油菜叶片与冠层图像集的光照亮度补偿实验结果表明,3种算法均能改善图像间的光照亮度差异;提取的颜色特征与SPAD值间相关分析结果表明,3种算法处理后能使相关性提高0.01～0.11。3种算法是解决光照亮度不一致问题的有效方法。The brightness inconsistences problem among images is the one of concern in crop leaf color-nitrogen nutrition diagnosis.We propose three illumination compensation algorithms,which are called shades of grey,ten color model and combination algorithm,respectively.The results of illumination compensation experiment,for 166 rice and rapeseed images,including indoor and outdoor,showed three algorithms could modify the illumination difference among images.The results of correlation analysis between leaf color features value extracted and SPAD value showed that it could improve the relevance of 0.01~0.11.Three algorithms are effective for improving brightness inconsistences problem among images.

关 键 词：计算机视觉 光照亮度补偿 十色模型 作物图像 

分 类 号：S126[农业科学—农业基础科学] TP391.41[自动化与计算机技术—计算机应用技术]