高分辨率遥感影像的田间道路识别方法  

Field road recognition method based on high-resolution remote sensing image

作　　者：袁翠霞 任艳敏 韩志强[4] 刘玉[2,3] 李淑华[2,3] YUAN Cuixia;REN Yanmin;HAN Zhiqiang;LIU Yu;LI Shuhua(School of Land Engineering,Chang'an University,Xi'an 710054,China;Research Center of Information Technology,Beijing Academy of Agriculture and Forestry Sciences,Beijing 100097,China;National Engineering Research Center for Information Technology in Agriculture,Beijing 100097,China;Zhaodong Agricultural Technology Extension Center of Heilongjiang Province,Zhaodong 151100,China)

机构地区：[1]长安大学土地工程学院,陕西西安710054 [2]北京市农林科学院信息技术研究中心,北京100097 [3]国家农业信息化工程技术研究中心,北京100097 [4]黑龙江省肇东市农业技术推广中心,黑龙江肇东151100

出　　处：《测绘通报》2023年第1期145-148,共4页Bulletin of Surveying and Mapping

基　　金：国家重点研发计划(2020YFD1100200);北京市农林科学院2022年度农业农村部农业遥感机理与定量遥感重点实验室建设(PT2022-24);北京市农林科学院院青年基金(QNJJ202232)。

摘　　要：田间道路是农田建设和管理的重要工程,其走向、宽度及通达性直接影响着农田耕作的便利程度,精准识别田间道路对农田建设效果评估具有重要意义。借助高分辨率遥感影像,本文首先基于分层分类思想,采用面向对象方法对田间道路进行识别;然后引入张量投票法、八邻域断点检测和形态学算法,对田间道路断点进行连接修复,最终得到田间道路识别结果。在河北省定州市典型案例区开展实证分析,田间道路识别的完整率、正确率、质量分别可达91%、74%、69%。Field road is an important project for farmland construction and management,and its direction,width and accessibility directly affect the convenience of farmland cultivation.Therefore,accurate identification of field road is of great significance for evaluation of farmland construction effect.With the help of high-resolution images,this paper adopts object-oriented method to identify the road based on the idea of hierarchical classification.Then,tensor voting method,eight-neighborhood breakpoint detection and morphological algorithm are introduced to repair the breakpoints.Take Dingzhou city,Hebei province as an example,the result of field road recognition is obtained,and the integrity rate,accuracy rate and quality of field road identification are 91%,74% and 69%,respectively.

关 键 词：高分辨率影像 田间道路 识别方法 面向对象 断点检测 

分 类 号：P237[天文地球—摄影测量与遥感]