摘要
以云南省德宏州瑞丽市为研究区,以Sentinel-2多光谱遥感影像和MOD13Q1为数据源,利用支持向量机分类方法识别该区的橡胶林,并对橡胶林空间分布特征及其种植适宜区进行分析。结果表明:1)联合Sentinel-2和MOD13Q1-NDVI数据识别橡胶林的结果明显优于仅用Sentinel-2多光谱影像或MOD13Q1-NDVI时序数据,实现了橡胶林高精度识别,总体精度为91.41%,Kappa系数为0.76。2)截至2019年底,瑞丽市橡胶林种植面积为7693.83 hm^2,主要种植区海拔范围为682~900 m,占总种植面积的83.61%;主要种植区坡度为5~15°,占总种植面积的84.61%。3)瑞丽市有6433.11 hm^2橡胶林种植在适宜区内,有1260.72 hm^2种植在不适宜区内,分别占橡胶林种植总面积的83.61%和16.39%。
Taking Ruili City,Dehong Prefecture,Yunnan Province as the study area,this paper analyzes the spatial distribution characteristics of the rubber forest and its suitable planting areas.Sentinel-2 multi-spectral remote sensing images and MOD13Q1 were used as data sources to identify the rubber forest with the method of supporting vector machine classification.The results show that 1)the identification ability of the combined Sentinel-2 and MOD13Q1 NDVI data for rubber forest was significantly better than using only one kind of Sentinel-2 multi-spectral image or MOD13Q1-NDVI time series data,which achieved high precision identification of rubber forest.The overall identification accuracy is 91.41%and the Kappa coefficient is 0.76.2)By the end of 2019,the rubber planting area in Ruili was 7693.83 hm^2,with the elevation range of the main planting area being 682-900 m,accounting for 83.61%of the total planting area;the slope of the main planting area was 5-15°,accounting for 84.61%of the total planting area.3)At present,6,433.11 hm^2 of rubber was planted in the suitable areas in Ruili while 1,260.72 hm^2 of rubber was not,accounting for 83.61%and 16.39%,respectively.
作者
蒋永泉
史正涛
高书鹏
JIANG Yongquan;SHI Zhengtao;GAO Shupeng(Department of Geography,Yunnan Normal University,Kunming 650500,China)
出处
《林业资源管理》
北大核心
2020年第6期54-60,共7页
Forest Resources Management
基金
云南省水利厅水利科技项目“云南主要人工经济林对区域水资源安全的影响调查研究”(2014003)。
关键词
多源遥感数据
橡胶林识别
空间分布特征
种植适宜区
瑞丽市
multi-source remote sensing data
rubber plantation identification
spatial distribution characteristics
suitable planting areas
Ruili City
