基于GEE平台的每年夏冬两期土地覆盖分类方法研究  

Land cover classification in summer and winter based on the Google Earth Engine platform

作　　者：梁顺波 孙林[1] 杜永明 赵祥 LIANG Shunbo;SUN Lin;DU Yongming;ZHAO Xiang(Shandong University of Science and Technology,School of Surveying and Spatial Information,Qingdao,Shandong,China;State Key Laboratory of Remote Sensing Science,Geographical Science College of Beijing Normal University,Beijing,China;Aerospace Information Research Institute,Chinese Academy of Sciences,State Key Laboratory of Remote Sensing Science,Beijing,China)

机构地区：[1]山东科技大学测绘与空间信息学院,山东青岛 [2]遥感科学国家重点实验室,北京师范大学地理科学学部,北京 [3]中国科学院空天信息创新研究院,遥感科学国家重点实验室,北京

出　　处：《北京师范大学学报(自然科学版)》2024年第6期831-842,共12页Journal of Beijing Normal University(Natural Science)

基　　金：第三次新疆综合科学考察资助项目(2022xjkk0405);遥感科学国家重点实验室开放基金资助项目(OFSLRSS202307)。

摘　　要：通过采用Landsat卫星遥感影像数据,利用GEE平台的大数据处理和分析能力,结合随机森林算法,开展了2000−2022年新疆地区的夏季和冬季土地覆盖变化的遥感监测.研究表明:通过输入光谱、光谱指数、地形、纹理、夜光等5个维度的18个特征,得到的新疆地区夏季和冬季土地覆盖分类结果的整体精度分别为95.6%和91.3%,Kappa系数分别为94.6%和88.4%.新疆地区夏季土地覆盖的主要变化发生在耕地、裸地、草地和人造地表之间的转化,而冬季的变化主要是雪地与其他土地类型之间的转变.本研究提供的夏冬两季产品有助于全面了解新疆地区土地覆盖变化的季节性动态变化特征,为农业生产、土地政策管理和资源管理提供数据支持.In this study Landsat remote sensing image data were used and big data processing and analysis capabilities of the Google Earth Engine(GEE)platform were leveraged.Random Forest algorithm was employed in this study to conduct remote sensing monitoring of land cover changes in Xinjiang during the summer and winter seasons from 2000 to 2022.The overall accuracy of land cover classification in Xinjiang during the summer and winter seasons were found to be 95.6%and 91.3%respectively,with Kappa coefficients of 94.6%and 88.4%.Over the past 23 years,average annual growth rate of summer cropland in Xinjiang was 0.04%,urban area was 0.005%,but bare land area decreased by 0.6%annually,grassland decreased by 0.3%annually.Water bodies area show an overall annual growth rate of 0.07%,snow-covered area increased from 2.0%to 2.1%.Shrubland and wetland remained relatively stable.Major changes in summer land cover in Xinjiang occured in transition areas between cropland,bare land,grassland,and artificial surfaces,while winter changes primarily involved transitions between snow-covered areas and other land cover types.Winter snow-covered area exhibited a decreasing trend,while other land cover types showed an increasing trend.Snow-covered area decreased from 29.71%to 25.79%of Xinjiang’s total area,grassland increased by 1.47%,bare land slightly expanded,and cropland experienced a slight growth of 0.62%.The seasonal products provided by this study for both summer and winter seasons contribute to a comprehensive understanding of seasonal dynamic characteristics of land cover changes in Xinjiang.This information can be valuable for agricultural production,land policy management,and resource management by offering data support.

关 键 词：土地覆盖变化 季节性变化 Google Earth Engine(GEE) 新疆地区 随机森林算法 

分 类 号：K901.4[历史地理—人文地理学] TP79[自动化与计算机技术—检测技术与自动化装置]