摘要
进行区域地形的自动分割和地理对象的识别提取,对于现代地理分析具有重要意义。该文运用影像分析方法,以中国黄土高原的典型地貌区为实验区,提出了基于DEM影像的地形自动分割和沟头识别方法。在黄土区域地形分割中,以沟谷分支点为种子点,以地表剖面曲率特征分析结果为生长准则,实现了沟间地、沟坡地、沟底地的区域自动分割;在沟坡地形自动分割的基础上,运用面向对象的目标分类识别思路,结合对象的地形和水文特征,进行了沟头对象的自动识别。实验结果表明:地形自动分割结果精度均大于90%,沟头对象的识别准确度为88.27%,验证了该文方法的有效性和可靠性。
In this paper,by using the image analysis method and taking the Chinese Loess Plateau typical landscape area as the experimental area,the terrain automatic segmentation and gully head recognition methods based on DEM image analysis are pro-posed.In the loess regions segmentation process,the valleys branch points and profile curvature were used as growling seeds and rules of growth respectively to implement region growling.Additionally,the automatic segmentation of three types of terrain which include gully interfluve land,gully slope land and gully groove land were achieved.Besides,on the basis of gully slope land automatic segmentation,the gully heads were identified according to the terrain and hydrological characteristics of the objects by using the object-oriented idea.The results show that the precision of terrain automatic segmentation and gully head recognition results are higher than90%and88.27%respectively,which prove that the proposed methods are scientific and effective.
作者
朱红春
黄伟
赵翼鹏
许晓斌
ZHU Hong-chun;HUANG Wei;ZHAO Yi-peng;XU Xiao-bin(Geometrics College, Shandong University of Science and Technology , Qingdao 266590 ,China)
出处
《地理与地理信息科学》
CSCD
北大核心
2017年第4期74-78,共5页
Geography and Geo-Information Science
基金
国家自然科学基金项目(41601408
41471331)
关键词
DEM
剖面曲率
地形分割
图像处理
沟头识别
DEM
profile curvature
terrain segmentation
image processing
gully head recognition
