叶面尘土量对大叶黄杨叶片图像的影响

Effects of Foliar Dust Content on Euonymus Japonicus Thunb Leaf Image

采用典型的城市绿化树种大叶黄杨叶片图像为研究对象,用1种将均值偏移与边缘置信度相结合的聚类算法对大叶黄杨叶片图像进行分割,建立线性模型和对数模型对图像信息与叶面尘土量进行拟合,分析叶面尘土量对叶片图像RGB(红、绿、蓝)三分量和HSI(色调、饱和度、亮度)等参量的影响。结果表明:采用均值偏移与边缘置信度相结合的聚类算法能够有效分割出大叶黄杨叶片图像;以色调H为自变量的对数模型可较好地拟合叶面尘土量与叶片图像参数间的关系,在一定范围内,随着叶面尘土量的增加,叶片图像H参数值减小。研究结果可为树木图像校正及叶面尘土对叶片反射光谱的影响等研究提供参考。

Using leaf image of Euonymus Japonicus Thunb, with a typical city greening tree species, as investigated subjects, leaf images were segmented by a clustering algorithm, which was a combination of mean shift and edge confidence, the relationship between image information and dust content were fitted by linear model and logarithmic model to research the effect of foliar dust content on the parameters of three components of leaf image RGB（red, green,blue） and HSI（hue, saturation, brightness） and so on. The results show that with the proposed algorithm the segmentation on Euonymus Japonicus Thunb leaf image is reasonable and effective; the relationship between foliar dust content and leaf images parameters can be fitted better by logarithmic model with color H variables, in a certain range,with the increase of foliar dust content the value of H parameter of leaf image reduced. This research result can offer a reference for the correct of tree images and the effects of foliar dust on leaf reflection spectroscopy.