基于高分辨率遥感影像的道路提取方法  

Road extraction method based on high resolution remote sensing image

作　　者：杨泽运 赵海洋 王强[2] 张玉娟[2] 张加颍[3] YANG Zeyun;ZHAO Haiyang;WANG Qiang;ZHANG Yujuan;ZHANG Jiaying(Department of Scientific Research,Heilongjiang Institute of Technology,Harbin 150050,China;College of Surveying and Mapping Engineering,Heilongjiang Institute of Technology,Harbin 150050,China;College of Foreign Languages,Heilongjiang Institute of Technology,Harbin 150050,China)

机构地区：[1]黑龙江工程学院科研处,黑龙江哈尔滨150050 [2]黑龙江工程学院测绘工程学院,黑龙江哈尔滨150050 [3]黑龙江工程学院外国语学院,黑龙江哈尔滨150050

出　　处：《交通科技与经济》2020年第2期73-76,共4页Technology & Economy in Areas of Communications

基　　金：黑龙江省教育厅科学技术研究项目(12531530)。

摘　　要：分别采用面向对象、最大似然、支持向量机3种分类方法进行实验区高分辨率遥感影像的道路提取,并作分类精度评定。结果表明,面向对象法能够综合利用高分辨率遥感影像的波谱、几何、纹理等多种特征信息,可有效克服影响因素的干扰。本案面向对象法道路提取的总分类精度较高,获得的分类结果较好。Three classification methods,object oriented,maximum likelihood and support vector machine(SVM),are used to extract road information from high resolution remote sensing images in study area,and the evaluation of classification accuracy is performed.The result shows that the object-oriented method can make use of spectral,geometric textureand other characteristics of high-resolution remote sensing images,and effectively overcome the interference of other influencing factors.The classification accuracy of object-oriented method is higher and the classification result is better.

关 键 词：遥感影像 道路提取 分类方法 面向对象 

分 类 号：P751[交通运输工程—港口、海岸及近海工程]