Land cover classification in summer and winter based on the Google Earth Engine platform
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.
land cover changeseasonal changesGoogle Earth Engine(GEE)Xinjiangrandom forest algorithm
万方数据
维普
