摘要
油菜叶片图像中主叶脉的去除是进行油菜叶片氮营养诊断的基础。经典的植物叶片叶脉提取算法由于不同程度地利用了叶片表面细节特征,不适合于实时性要求较高的油菜叶片营养诊断处理。为此,提出了一种基于RGB颜色空间的快速去除油菜叶片图像主叶脉的快速截留超绿—超红新算法。实验结果表明,提出的快速截留超绿—超红算法不仅能够准确、有效、完整地清除出油菜叶片的叶脉部分,同时能够有效地去除叶面上的污点、黄化部分以及虫咬痕迹。通过油菜叶色特征与叶片SPAD值相关性分析发现,清除主叶脉后的油菜叶片图像颜色特征对油菜叶片氮营养诊断更有效。
The segmentation of rapeseed leaf vein is the basis of rapeseed leaf nitrogen nutrition diagnosis.Classical leaf vein extraction algorithm are not suitable for real-time processing by using leaf detail feature.Therefore,an algorithm for segmentation of rapeseed leaf vein based on RGB color space is proposed.The experimental results show that the algorithm not only eliminate the vein of leaf image exactly,effectively and integrally,but remove the splash,yellowed part and insect biting mark of rapeseed leaf effectively.Correlation analysis with leaf color features and SPAD value results show that the color feature extracted by new algorithm had better performance for rapeseed leaf nitrogen nutrition diagnosis.
出处
《农机化研究》
北大核心
2012年第9期183-186,共4页
Journal of Agricultural Mechanization Research
基金
国家自然科学基金项目(31071328)
关键词
计算机视觉
图像分割
油菜
叶脉
computer version
image segmentation
rapeseed
leaf vein
