测绘通报 ›› 2020, Vol. 0 ›› Issue (8): 13-17.doi: 10.13474/j.cnki.11-2246.2020.0240

• 自然资源监测 • 上一篇    下一篇

GEE云平台支持下的西天山森林遥感监测与时空变化分析

邵亚奎1, 王蕾2, 朱长明1, 方晖3, 张新4, 黄端5, 陶莉1   

  1. 1. 江苏师范大学地理与测绘学院, 江苏 徐州 221116;
    2. 新疆林业科学院现代林业研究所, 新疆 乌鲁木齐 830063;
    3. 中国科学院新疆生态与地理研究所, 新疆 乌鲁木齐 830011;
    4. 中国科学院遥感与数字地球研究所遥感科学国家重点实验室, 北京 100101;
    5. 东华理工大学测绘工程学院, 江西 南昌 330013
  • 收稿日期:2020-03-18 出版日期:2020-08-25 发布日期:2020-09-01
  • 通讯作者: 朱长明。E-mail:zhuchangming@jsnu.edu.cn E-mail:zhuchangming@jsnu.edu.cn
  • 作者简介:邵亚奎(1992-),男,硕士生,主要研究方向为遥感与GIS应用。E-mail:syk227816_gis@163.com
  • 基金资助:
    国家重点研发计划(2017YFB0504201);中国科学院战略性先导科技专项课题(XDA05050202)

Forest survey and spatio-temporal analysis in West Tianshan mountains supported by Google Earth Engine

SHAO Yakui1, WANG Lei2, ZHU Changming1, FANG Hui3, ZHANG Xin4, HUANG Duan5, TAO Li1   

  1. 1. College of Geography and Geomatics, Jiangsu Normal University, Xuzhou 221116, China;
    2. Research Institute of Modern Forestry, Xinjiang Academy of Forestry, Urumqi 830063, China;
    3. Xinjiang Institute of Ecology and Geography, Chinese Academy of Sciences, Urumqi 830011, China;
    4. Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, Beijing 100101, China;
    5. Faculty of Geomatics, East China University of Technology, Nanchang 330013, China
  • Received:2020-03-18 Online:2020-08-25 Published:2020-09-01

摘要: 针对区域大尺度森林遥感调查、精确信息提取和时间序列变化监测过程中存在的数据挑选困难、计算效率较低、提取精度不高等问题,本文基于谷歌云计算平台(GEE)强大的海量遥感数据组织、存储和计算功能,根据新疆干旱区森林资源的空间分布特点,结合多源遥感数据和地理要素数据集,首先构建了光谱+纹理+地形等多维分类特征集;然后在地理国情监测森林地面调查样本数据的协助下建立了西天山森林分类样本数据库;进而采用随机森林分类算法实现了对西天山森林1995、2000、2005、2010、2015和2018年6期自动分类;最后通过云端与本地相结合完成了森林资源遥感分类数据编辑检查、制图与分析。研究结果表明:①1995-2018年西天山森林总体呈动态扩张趋势,森林分布面积从1995年的3 953.6 km2年增加到2018年的4 243.2 km2,增长速率为12.6 km2/a;在结构组成上,西天山森林以针叶林为主,阔叶林、灌木林、针阔混交林较少。②在时间变化过程上,西天山森林的扩张态势呈现缓中增强,2005-2018年增长速率要明显高于1995-2005年。③在空间变化特征上,不同森林类型之间的转化很少,新增林地主要来自非林地,2000年以来非林地转林地的面积约为520 km2。非林地转化为林地区域主要集中在特克斯县分局、尼勒克县分局、昭苏县分局、新源林场、巩留分局、伊宁分局,转入面积分别为111.14、102.19、67.16、56.45、42.76、40.71 km2

关键词: 林业遥感, GEE云计算, 时空变化, 随机森林, 西天山

Abstract: Aiming at the problems of difficult data acquisition, low calculation efficiency, large amount of data and low accuracy in large-scale spatial distribution information extraction and long-time series change monitoring, supported by the Google Earth Engine platform, based on multi-source remote sensing data, a multi-dimensional classification feature set is constructed, and the forest extraction and classification of West Tianshan mountains in 1995, 2000, 2005, 2010, 2015 and 2018 are studied by random forest classification algorithm. Six periods of forest high-precision extraction and mapping analysis are completed by cloud-local combination. The results show that the process constructed in this paper has high accuracy. During the past 23 years, the forest area in West Tianshan mountains has been expanding dynamically, increasing from 3 953.57 km2 in 1995 to 4 243.41 km2 in 2018, with an increasing rate of 12.60 km2/a. In terms of time, the growth rate of forest has been rising steadily in the past 23 years. The growth rate in 2005-2018 is higher than that in 1995-2015. In terms of spatial distribution, there are few conversions between different forest types. The newly added forest land mainly comes from non-forest land. The area of non-forest land converted to forest land has been about 520 km2 in the past 23 years. The Tex county branch, Nileke country branch, Zhaosu county branch, Xinyuan forest farm, Gongliu branch, and Yining branch have transferred areas of 111.14, 102.19, 67.16, 56.45, 42.76, and 40.71 km2.

Key words: forest remote sensing, GEE cloud caculation, spatio-temporal change, random forest, West Tianshan mountains

中图分类号: