计算机辅助小麦图像识别应用中颜色特征基本参量的表达

在计算机辅助图像处理中，运用颜色特征进行图像的分类和识别是简便而有效的一种方法。然而，颜色特征的表达和提取是否准确、合理直接决定着分类和识别的可靠性。本文在重点分析RGB、HIS和L*a*b*三种常用颜色模式基本参量含义及相互间关系的基础上，结合小麦图像自身的特点，通过对30幅小麦图像在三种颜色模式下的9个基本参量进行主成分分析，建立了应用于小麦图像识别的颜色特征基本参量表达式，并对这三种颜色模式的9个基本参量进行了分类，提出了确定而有意义的表征小麦颜色特征的主成分指标。结果如下：基于第一主成分的分类指标综合表达出小麦冠层的亮绿色特点，分类结果具有较高的准确性和可靠性；第二主成分指标主要表达小麦冠层黄绿颜色变化的特点，能够形成连续的量化指标空间。第三主成分指标主要表达小麦正常绿色的情况，在图像获取亮度差异较小时可以进行小麦正常绿色值的评价。

Basic Components' Expression of Color Features Used in Computer Assisted Crop Images Identifification

Using color features to identifying and classifying images is a simple and efficient method in the computer assisted image procession. Therefore, the nicety and the properness of expressing and extracting the color features decide directly the result's reliability. The paper mainly discussed nine basic components' meaning from three common color models, RGB, HIS and L*a*b*, and analysed their relations. Further, expressions of color features' basic components were founded based on the PrincipaI Components AnaIysis of nine basic color components from 30 wheat images and wheat images' own characteristics. And the nine basic components of three color models were classified and some new principal components' expressions were put forward to show wheat color features definitely and significatively. Some results were as follows: The classification results were exact and reliable based on the first principal components with light-green characteristics of wheat canopy. The second principal components mainly showed wheat color range changed form green to yellow and formed a measurable and sequential room. The third principal components mainly showed the normal green of wheat and could be used to evaluate wheat greenness value when the change of brightness could be neglected.