Investigation on the space distribution of berry tea range extraction based on Sentinel-2 remote sensing images
This study took M ground as the research area. Based on multi temporal Sentinel-2 image data, a deci-sion tree classification model was constructed using the typical vegetation index NDVI and EVI time series change char-acteristics. The study area of berry tea planting was extracted and a map of the berry tea planting area was drew;com-pared classification accuracy using decision tree classification method, maximum likelihood method, and support vector machine. The results indicated that(1)berry tea cultivation in M ground was mainly distributed in T ground, M ground, and G ground in the central and northern regions, with less distribution in D ground and Q ground in the eastern and southern regions;(2)through the comparison of three classification results, it was found that the decision tree classifica-tion method(with an overall accuracy of 97.2％and a Kappa coefficient of 0.963)was the best, followed by support vec-tor machine(with an overall accuracy of 92.9％and a Kappa coefficient of 0.896), and the maximum likelihood method (with an overall accuracy of 91.7％ and a Kappa coefficient of 0.888)had the worst classification performance. This study could provide reference for extracting the planting range of berry tea in terms of data source and temporal feature construction.
berry teacropsmulti-temporalremote sensingvegetation index
万方数据
维普
