基于规则面向对象分类法的贵州省山区火烧迹地提取
Extraction of Burned Land in Guizhou Province With Rule-based Object-oriented Classification
刘芸 1廖瑶 1李慧璇 1杨娟 1王伟2
作者信息
- 1. 贵州省生态气象和卫星遥感中心,贵州 贵阳 550002
- 2. 贵州省特种水产工程技术中心,贵州 贵阳 550025
- 折叠
摘要
基于国产高分一号卫星宽幅(WFV)影像,对贵州省复杂地形山区火烧迹地的光谱特征、形状特征、纹理特征等进行了分析,建立了火烧迹地提取规则,完成了研究区基于规则面向对象的火烧迹地提取.利用高分二号卫星1m分辨率影像对提取结果进行精度验证.结果表明,基于规则面向对象分类法的GF1WFV火烧迹地提取总体精度为92.67%,总Kappa系数为0.89,能较好地完成研究区的火烧迹地分类提取,分类质量达到极好水平,为贵州省山区火烧迹地提取提供了一定的参考和理论依据.
Abstract
Based on GF1 WFV images,this paper analyzed the spectral,shape,and texture characteris-tics of the burned land in complex mountainous areas of Guizhou Province,established the extraction rules of burned land,and completed the extraction of burned land with rule-based object-oriented classi-fication.The results showed that the overall accuracy of this method was 92.67%,and the total Kappa coefficient was 0.89,which could well complete the extraction of burned land in the research area,and the classification quality reached an excellent level,providing a certain reference and theoretical basis for the extraction of burned land in complex mountainous areas of Guizhou Province.
关键词
面向对象分类法/高分卫星影像/火烧迹地提取/山区/贵州Key words
object-oriented classification/GF1 satellite image/extraction of burned land/mountainous areas/Guizhou Province引用本文复制引用
基金项目
贵州省科技厅基础研究计划(黔科合基础-ZK[2021]一般193)
出版年
2024
NETL
NSTL
维普
万方数据