基于高时空分辨率可见光遥感数据的热带山地橡胶林识别

doi:10.11873/j.issn.1004-0323.2018.6.1122

遥感技术与应用

2018, Vol. 33

Issue (6): 1122-1131 DOI: 10.11873/j.issn.1004-0323.2018.6.1122

数据与图像处理

基于高时空分辨率可见光遥感数据的热带山地橡胶林识别

高书鹏¹，史正涛¹，刘晓龙¹，柏延臣^2，3

(1.云南师范大学旅游与地理科学学院，云南昆明 650500;
2.北京师范大学地理科学学部，北京 100875;
3.北京师范大学地理科学学部遥感科学国家重点实验室，北京 100875)

Identification of Rubber Plantations in Tropical Mountainous Region based on High Spatio-temporal Resolution Visible Remote Sensing Data

Gao Shupeng¹，Shi Zhengtao¹，Liu Xiaolong¹，Bo Yanchen^2，3

(1.College of Tourism & Geography Science，Yunnan Normal University，Kunming 650500，China;
2.Faculty of Geographical Science，Beijing Normal University，Beijing 100875，China;
3.State Key Laboratory of Remote Sensing Science，Faculty of Geographical Science，
Beijing Normal University，Beijing 100875，China)

全文: PDF(8110 KB)

摘要： 20世纪90年代以来橡胶林种植面积在西双版纳地区迅猛扩大，对该区域橡胶林种植面积、种植结构变化的精确监测是客观评价该地区橡胶林种植与生态环境变化关系的关键。针对西双版纳热带山地地区植被光谱特征的相似性及地形和气候条件的复杂性问题，结合该地区橡胶林冬季落叶的物候特征，采用时空数据融合算法，分别选取中分辨率的ETM+、OLI、Sentinel-2A数据与高时间分辨率的MODIS数据融合，建立高时空分辨率可见光遥感数据集，并分析不同融合数据源对热带山地环境下橡胶林识别精度的差异。结果表明：①基于时空融合数据提取的橡胶林物候变化特征能够实现较高精度的橡胶林识别，识别精度可以达到89%以上，Kappa高于0.83；②运用10 m分辨率的Sentinel-2A数据进行分类时，能够获取比Landsat数据更高精度的分类结果，表明Sentinel-2A数据在高时空数据融合及热带植被遥感应用中有较好前景。

关键词： 橡胶林识别; 时空数据融合; 热带山地; Sentinel-2A

Abstract: Since 1990s，rubber plantations has been growing rapidly in Xishuangbanna，it is vitally important to the evaluation of the relationship between the rubber plantation and the change of ecological environment in the region.Aiming at the problem of similarity of vegetation's spectral feature and complexity of topography and climatic conditions，combined with the phenological characteristics of rubber plantation fallen leaf in winter in the region，ESTARFM algorithm was used and we selected ETM+，OLI and Sentinel\|2A data to fuse with high temporal resolution MODIS data respectively to establish a high spatial and temporal resolution visible remote sensing data setand analyze the difference of the recognition accuracy of the rubber plantations in tropical mountainous environments by different fusion data sources.The results show that：(1) During the key phenological period of the rubber plantation from January to March，the phenological features extracted using remote sensing data based on high spatial and temporal resolution can remarkably improve the identification accuracy of rubber plantations，the recognition accuracy more than 89% and Kappa higher than 0.83;(2) in the vegetation classification of fragmented mountainous area，10 m resolution Sentinel\|2A data used classification will obtain higher classification accuracy than Landsat data，which indicates that the Sentinel\|2A data is more promising in the high spatial\|temporal data fusion and the tropical vegetation remote sensing application.

Key words: Rubber plantations identification Spatio-temporal data fusion Tropical mountainous area Sentinel-2A

收稿日期: 2018-01-21 出版日期: 2019-01-29

ZTFLH:

TP23710

基金资助: 国家重点研发计划项目课题“全球多时空尺度遥感动态监测与模拟预测”（2016YFB0501502)，云南省青年基金项目“基于多源遥感数据的植被类型精细分类方法研究”(2016FD021)，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)。

作者简介: 高书鹏(1991-)，男，云南广南人，博士研究生，主要从事热带植被遥感分类研究。Email：gaoshuaa@163.com。

	服务
	把本文推荐给朋友
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章

引用本文:

高书鹏, 史正涛, 刘晓龙, 柏延臣. 基于高时空分辨率可见光遥感数据的热带山地橡胶林识别[J]. 遥感技术与应用, 2018, 33(6): 1122-1131.

Gao Shupeng, Shi Zhengtao, Liu Xiaolong, Bo Yanchen. Identification of Rubber Plantations in Tropical Mountainous Region based on High Spatio-temporal Resolution Visible Remote Sensing Data. Remote Sensing Technology and Application, 2018, 33(6): 1122-1131.

链接本文:

http://www.rsta.ac.cn/CN/10.11873/j.issn.1004-0323.2018.6.1122 或 http://www.rsta.ac.cn/CN/Y2018/V33/I6/1122

[1]	康峻,王力,牛铮,高帅,邬明权 . 基于局部空间自相关分析的时空数据融合[J]. 遥感技术与应用, 2015, 30(6): 1176-1181.

Viewed

Full text

Abstract

Cited

Shared

Discussed