基于决策树的大尺度复杂地区夏收作物遥感提取与分析

Extraction and analysis of summer crops in large-scale complex areas based on decision tree

以甘肃省为研究区,基于时序Sentinel-2卫星影像数据,计算归一化植被指数(NDVI)和黄度值,采用决策树(Decision trees,DTs)方法提取研究区的主要夏收作物小麦和油菜种植面积,并对比分析其精度。利用此方法得到甘肃省小麦和油菜作物识别的平均总体精度达到87.4%,其中,甘肃省中部地区平均总体精度为92.4%,河西地区平均总体精度为87.7%,东南部地区平均总体精度为82.0%。研究结果表明,基于Sentinel-2卫星影像数据采用决策树方法进行大尺度复杂区域、高分辨率作物种植面积提取具有可行性。

Taking Gansu province as the research area,the normalized difference vegetation index(NDVI)and yellowness value were calculated based on the time series Sentinel-2 satellite image data.The decision trees(DTs)method was used to extract the planting area of wheat and rape in the study area,and the accuracy was compared and analyzed.Using this method,the average overall accuracy of wheat and rape identification in Gansu province was 87.4%.The average overall accuracy was 92.4%in the central region of Gansu province,87.7%in Hexi region,and 82.0%in the southeastern region.The results show that it is feasible to extract the planting area of high-resolution crops in large-scale complex areas by using decision tree method based on Sentinel-2 satellite image data.