色差直方图特征改进Otsu的橡胶缺素叶片图像分割研究  

Image segmentation of rubber deficiency leaves based on Otsu algorithm improved with characteristic of RGB dispersion

作　　者：秦小立[1] 李玉萍[1] 张培松[2] 叶露[1] 谢铮辉[1] 

机构地区：[1]中国热带农业科学院科技信息研究所/海南省热带作物信息技术应用研究重点实验室,海南儋州571737 [2]中国热带农业科学院橡胶研究所,海南儋州571737

出　　处：《广东农业科学》2015年第24期179-183,F0003,共6页Guangdong Agricultural Sciences

基　　金：海南省自然科学基金(614245)

摘　　要：针对橡胶叶片颜色随缺素程度不断变化的特点,提出色差直方图特征改进最大类间方差法(Otsu)的橡胶缺素叶片图像分割方法。首先利用图像RGB三分量计算色差图像,保留更多的图像颜色空间信息;然后计算图像的色差直方图,并改进Otsu阈值分割算法;最后利用改进算法对缺素叶片图像进行分割。结果表明,改进算法在缺素叶片图像分割上优于传统Otsu算法。With the degree of nutrient deficiency changing,the rubber leaf color changes. Therefore,the Otsu algorithm improved with characteristic of RGB dispersion for rubber deficiency leaves was proposed. Firstly,the image of RGB dispersion was calculated with the RGB three-component of color image,which retained more image color space information,then calculated the color histogram of image,and which was used to improve the Otsu segmentation algorithm. Finally,the deficiency leaf images were segmented by the improved algorithm. The results and algorithm analysis showed that the improved algorithm was superior to the Otsu algorithm in the image segmentation of rubber deficiency leaves.

关 键 词：图像分割 色差直方图 OTSU 橡胶缺素叶片 

分 类 号：S126[农业科学—农业基础科学]