基于资源卫星数据面向对象法的红树林提取研究  

Research on Mangrove Extraction from ZY-3 Data Based on the Object-oriented Method

作　　者：阮胜权 王新广[1] 李勃燃 RUAN Shengquan;WANG Xinguang;LI Boran(Academy of Prospecting of Tianjin,Tianjin 300110,China)

机构地区：[1]天津市勘察院,天津300110

出　　处：《测绘与空间地理信息》2021年第1期96-98,共3页Geomatics & Spatial Information Technology

摘　　要：利用无人机在研究区域内采集了红树林分类的训练样本点,基于资源三号高分影像对红树林进行分类。首先采用面向对象的方法,建立决策树对保护区内红树林进行提取。之后基于面向对象法对红树林区域进行种间分类,分析保护区内红树林种群空间分布格局。研究表明,红树林种间分类精度达到76.67%,总体Kappa系数为0.68,结果具有高度一致性。本文为红树林自然保护区的管理提供理论支撑,为红树林湿地保护和生态恢复提供科学依据。In this paper,the training sample points of mangrove classification are collected by UAV in the study area,and the mangrove classification is based on the high-resolution image of ZY-3.Firstly,the decision tree is established to extract mangrove in the reserve by object-oriented method.Then,based on the object-oriented method,we classify the mangrove areas and analyze the spatial distribution pattern of mangrove population in the reserve.The results showed that the accuracy of classification was 76.67%,and the kappa coefficient was 0.68.This paper provides theoretical support for mangrove natural reserve management and scientific basis for mangrove wetland protection and ecological restoration.

关 键 词：面向对象法 红树林 资源三号卫星 

分 类 号：P237[天文地球—摄影测量与遥感]