基于GF-2面向对象土地利用分类研究  

Object-oriented GF-2 image land use classification

作　　者：郝旋捷 冯晓天 赵燕伶 张巧玲 Hao Xuanjie;Feng Xiaotian;Zhao Yanling;Zhang Qiaoling(Natural Resources Satellite Application Center of Shaanxi Province,Xi’an Shaanxi 710082,China)

机构地区：[1]自然资源陕西省卫星应用技术中心,陕西西安710082

出　　处：《山西建筑》2024年第7期171-174,共4页Shanxi Architecture

基　　金：秦岭卫星遥感综合监测服务平台建设与应用(陕地调院发[2022]54号)。

摘　　要：高分辨率遥感技术发展迅速,传统技术已经无法满足信息提取的要求,严重影响提取精度和效率。以秦岭北麓长安区与鄠邑区交界处北部部分区域为实验区,高分二号影像为数据源,采用面向对象分类法进行土地利用分类研究,并进行监督分类做对比实验。结果表明,采用面向对象分类法分类结果总体精度为90.05%,Kappa系数为0.857,比监督分类方法精度高出14.55%,Kappa系数高出0.288。面向对象分类方法总体分类效果较好,有效提高了分类精度。With the rapid development of high resolution remote sensing technology,traditional technology has been unable to meet the requirements of information extraction,which seriously affects the accuracy and efficiency of extraction.In this paper,part of the northern area at the junction of Chang’an District and Huyi District at the northern foot of the Qinling Mountains was selected as the experimental area,and the Gaofen-2 image was used as the data source.Object-oriented classification method was adopted for land use classification research,and supervised classification was conducted for comparison experiment.The results show that the overall accuracy and Kappa coefficient of the classification results by object-oriented classification method are 90.05%and 0.857,which is 14.55%higher than that by supervised classification method and 0.288 higher than Kappa coefficient.Object-oriented classification method has good overall classification effect and improves classification accuracy effectively.

关 键 词：高分二号影像 面向对象分类 秦岭北麓 ECOGNITION 土地利用 

分 类 号：TU198.1[建筑科学—建筑理论]