Dynamic Change and Driving Force Analysis of Forest Coverage Rate in Changsha City,Hunan Province
The urban area of Changsha in Hunan province was taken as the study area,and three Landsat series data in 2000,2010,and 2020 provided by the geospatial cloud data source are adopted.Four classifiers,namely Support Vector Machine,Minimum Distance Classifier,Decision Tree,and Random Forest were used to extract the forest cover information of Changsha.The overall accuracy of Support Vector Machine is 79.08%,and the Kappa coefficient is 0.64;The overall accuracy of the Minimum Distance Classifier is 82.86%,and the Kappa coefficient is 0.73;The overall accuracy of the Decision Tree is 91.95%,and the Kappa coefficient is 0.88;The overall accuracy of random forest is 94.95%,and the Kappa coefficient is 0.92.Through precision com-parison,among the four classification methods,the Random Forest classification accuracy of Changsha is the highest.Due to the continuous advancement of urbanization construction and development process,the forest coverage rate in the Changsha urban area was continuously decreasing from 2000 to 2020.The forest land area was reduced from 973.03 km2 to 700.39 km2,and the forest coverage rate dropped from 44.55%to 32.07%.Through this study,more suitable methods and technologies for the extraction of forest cover information in Changsha are provided,and accurate data support for large-scale ecological environment management and gov-ernance are also provided in similar areas to Changsha.
forestry remote sensingLandsatRandom Forestforest coverage rateChangsha City
NETL
NSTL
万方数据
