基于复杂地貌下的土地利用信息提取技术  

Land Use Information Extraction Technology Based on Complex Landform

作　　者：陈兴芳 尹继鑫 郭德福 CHEN Xingfang

机构地区：[1]西宁市测绘院,青海西宁810000

出　　处：《地理空间信息》2021年第4期63-66,I0006,共5页Geospatial Information

基　　金：西宁市2019年科技计划项目(2019-Y-12)。

摘　　要：选取复杂地貌下的西安市临潼区为例,对其高分辨率遥感影像展开信息搜集与提取,采用eCognition自带分类构建规则集,对于土地利用信息进行重点关注。围绕对象分层网络方法展开研究,进一步提升信息提取的精准度,避免过分割以及欠分割情况造成的负面作用,同时有效解决了传统提取中较为困难的阴影信息提取。研究结果表明eCognition在复杂地貌特征下具有良好的适用性,对于国内外其他类似复杂地貌下的研究具有重要的依据和借鉴意义。In this study,taking Lintong District of Xi’an City under complex landforms for example,we collected and extracted information from its high resolution remote sensing images,and used eCognition to build rule set with its own classification.At the same time,we studied the object hierarchical network method to further improve the accuracy of information extraction and avoid the negative effects caused by over segmentation and under segmentation.Then,we effectively solved the more difficult shadow information extraction in the traditional extraction.This study shows that eCognition has good applicability under the complex landform characteristics,and it has important basis and reference significance for the other similar complex landform research at home and abroad.

关 键 词：信息提取 影像分割 土地利用 ECOGNITION 对象层次网络 

分 类 号：P237[天文地球—摄影测量与遥感]