论文部分内容阅读
针对目前面向对象多尺度分类方法中,最优分割尺度的确定方法不具有普适性或者易受主观性影响的问题,以西藏米林县的Landsat 8OLI影像为数据源,对研究区影像多尺度分类进行研究。首先确定多尺度分类的最优分割尺度,提出基于多尺度分类精度Kappa系数的最优分割尺度函数模型法,在此基础上,利用多尺度分类分别与最近邻分类和阈值分类法相结合的方法,对研究区影像进行分类。结果表明:分割尺度分别为190、150、100、60,多尺度分类法比单一尺度分类精度高;最近邻多尺度分类法比阈值多尺度分类精度高,其总精度分别为86%和72%,Kappa系数分别为0.72和0.69。最优分割尺度函数模型在具有普适性的基础上更具有科学理论性,多尺度分类与最邻近分类结合的方法比与阈值分类结合的方法分类效果好,为后续植被动态变化监测提供了依据。
In order to solve the problem of determining the optimal segmentation scale in the current object-oriented multi-scale classification method, Landsat 8OLI image in Milin County, Tibet is used as the data source for multi-scale image classification in the study area Classification for research. Firstly, the optimal segmentation scale of multi-scale classification is determined and the optimal segmentation scale function model method based on Kappa coefficient of multi-scale classification accuracy is proposed. Based on this, the method of multi-scale classification is combined with nearest neighbor classification and threshold classification, The study area images were classified. The results show that the segmentation scales are 190,150,100 and 60, respectively, and the accuracy of multi-scale classification is higher than that of single scale classification. The nearest neighbor multi-scale classification is more accurate than the threshold multi-scale classification with the total accuracy of 86% and 72% Kappa coefficients were 0.72 and 0.69, respectively. The optimal segmentation scale function model is more scientific and scientific based on its universality. The combination of multi-scale classification and nearest-neighbor classification is more effective than the combination of threshold classification and classification, which provides a basis for the follow-up monitoring of vegetation dynamic changes .