高分六号影像在自然资源调查中的应用研究  

Application of GF-6 Images in Natural Resources Investigation

作　　者：康雄华[1] 田一帆 张毅[1] KANG Xionghua;TIAN Yifan;ZHANG Yi(School of Geodesy and Geomatics,Wuhan University,Wuhan 430079,China;Technical University of Munich,Munich 80333,Germany)

机构地区：[1]武汉大学测绘学院,湖北武汉430079 [2]慕尼黑工业大学,德国慕尼黑80333

出　　处：《地理空间信息》2023年第4期54-61,71,共9页Geospatial Information

基　　金：国家重点研发计划项目(2020YFD1100205)。

摘　　要：以广州市南沙区自然资源调查为例,针对高分六号遥感影像的特点,建立了高分六号影像可以适应的自然资源分类体系。分析了面向对象的遥感影像分类过程中的各个过程,研究了多尺度影像分割在不同参数设置下的分类效果,得出了较为合适的多尺度分割参数,进行了面向对象分类实验和分类精度评价,影像分类精度达到了78.66%。通过对南沙区影像在实验各个流程中的区别及联系进行了横向对比,总结了其中的规律,显示了基于高分六号的面向对象遥感影像分类方法在自然资源调查中的应用价值。Taking the natural resources survey in Nansha District of Guangzhou for example,according to the characteristics of GF-6 remote sensing image,we established a natural resources classification system that GF-6 image can adapt to,analyzed each process in the process of ob-jectoriented remote sensing image classification,studied the classification effect of multi-scale image segmentation under different parameter set-tings,obtained more appropriate multi-scale segmentation parameters,and carried out object-oriented classification experiment and classification accuracy evaluation.The image classification accuracy reaches 78.66%.Through the horizontal comparison of the differences and relations be-tween the images in NanshaDistrict in each process of the experiment,we summarized the laws,which showed the application value of the ob-jectoriented remote sensing image classification method based on GF-6 in the investigation of natural resources.

关 键 词：自然资源调查 高分六号 面向对象分类 分类体系 多尺度分割 

分 类 号：P237[天文地球—摄影测量与遥感]