基于监督分类和DEM的山脚线提取方法OA
A method for foothill line extraction based on supervised classification and digital elevation model
针对现有山脚线提取方法存在主观性强、适应性差及效率低等问题,本文提出一种基于监督分类和数字高程模型(DEM)的山脚线提取方法.该方法融合高程、坡度及剖面曲率三类地形因子,构建多通道地形特征影像,分别采用最小距离法、最大似然法与神经网络法三种图像监督分类方法,对特征影像中的山脉与非山脉像元进行识别与划分,从而实现山脚线的提取.结果表明,基于图像监督分类的方法能够有效实现山脚线的提取;三种监督分类方法中,神经网络法的提取精度最高,显著优于最小距离法与最大似然法.
To address the issues of high subjectivity,poor adaptability,and low efficiency in existing foothill line extraction methods,this paper proposed a foothill line extraction approach based on supervised classification and digital elevation model(DEM).The proposed method integrated three terrain factors—elevation,slope,and profile curvature—to construct a multi-channel terrain feature image.Three supervised image classification methods,namely the minimum distance method,maxi-mum likelihood method,and neural network method,were applied to identify and distinguish mountain and non-mountain pixels in the feature image,thereby extracting the boundaries of the foothill line.The results demonstrate that the image-based supervised classification approach can effectively achieve foothill line extraction.Among the three supervised classifi-cation methods,the neural network method achieves the highest extraction accuracy,significantly outperforming both the minimum distance method and the maximum likelihood method.
罗晔;金晓彤;蒋汪洋;王久懿
中国地质调查局哈尔滨自然资源综合调查中心,黑龙江 哈尔滨 150081||自然资源部哈尔滨黑土地地球关键带野外科学观测研究站,黑龙江 哈尔滨 150081自然资源部第二航测遥感院,黑龙江 哈尔滨 150081自然资源部第二航测遥感院,黑龙江 哈尔滨 150081中国地质调查局哈尔滨自然资源综合调查中心,黑龙江 哈尔滨 150081||自然资源部哈尔滨黑土地地球关键带野外科学观测研究站,黑龙江 哈尔滨 150081
天文与地球科学
山脚线地形因子波段合成监督分类数字高程模型(DEM)
foothill lineterrain factorband synthesissupervised classificationdigital elevation model(DEM)
《北京测绘》 2026 (8)
1076-1081,6
中国地质调查局地质调查项目(DD20251300222)东北地质科技创新中心区创基金(QCJJ2023-33).
评论