温室植物病害的图像处理及特征值提取方法的研究——基于色度的特征值提取研究

Image processing and extracting color features of greenhouse diseased leaf

为了进行温室植物病害的智能化防治,运用计算机图像处理技术对生产中常见的两种黄瓜病作了研究,比较了几种常见的色度学系统,以色调H作为颜色特征参数的研究域.以色调直方图统计特征参数的计算,和百分率直方图区间值特征作为区分病变叶片与正常叶片的重要依据.结果发现色调H偏度可以较为明显的区分不同病变情况,在进一步研究不同色调区域病变叶片和正常叶片的色调分布情况时,发现在色调(48～50)和(45～47)区间,区分正常叶片与病变叶片的效果最好.该研究为后期的模式识别提供了重要的特征参数.

In order to carrying on the intelligent disease prevention and cure for glasshouse plants,this paper(researched) two kinds of familiar cucumber diseases in the method of picture processing.Compared several kinds of hue system using hue H as the color characteristic parameter,and distinguished between normal leaf and(diseased) leaf according to hue histogram statistic parameters and chroma histogram of section value.Studied the chroma distribution of normal and diseased leaves in different hue extent.The results showed that(48~50) and(45~47) were the best extent to distinct normal and diseased leaves.This paper provides important characteristic parameters for the pattern identified.