关键词: 土地利用, 谷歌地球引擎, 随机森林, 土地利用动态度, 地学信息图谱, 天山野果林

Abstract: Based on Google earth engine （GEE） remote sensing big data cloud platform, through Landsat series images, using random forest algorithm, combined with spectral index, topographic features for land use classification, through the geographic information mapping, and the spatial and temporal changes of land use of Tianshan wild fruit forest distribution area from 1997 to 2021 were analyzed through geographic information mapping and land use dynamics. The results showed that through GEE platform and random forest algorithm, land use classification could be realized quickly, and the overall classification accuracy was greater than 90.00%, and the Kappa coefficient was higher than 0.85; from 1997 to 2021, the Tianshan wild fruit forest distribution area was mainly arable land, forest land and grassland, and the three accounted for more than 90% of the total area; land use changes were dominated by the transformation of grassland to arable land and the transformation of forest land to grassland; construction land, arable land and water area continued to increase, the change showed the aggregative distribution, forest land decreased first and then increased, and the overall land area decreased seriously; the growth rate of construction land was the largest and the change was the most drastic, the overall reduction rate of unused land and forest land was large and the change was more drastic.

Key words: land use, Google earth engine（GEE）, random forest, land use dynamic index, geoinformatics mapping, Tianshan wild fruit forest