多时相哨兵2号影像提取撂荒地方法

The method extraction of abandoned land from Sentinel 2 images

撂荒地有日渐严重的趋势,快速、精准地掌握撂荒地信息,对保护耕地与粮食安全具有重要意义。本文利用2018—2021年不同季节的哨兵2号卫星影像数据,使用eCognition、ENVI等软件,完成了撂荒地植被指数变化检测法、分类后变化检测法和决策树分类法的提取研究。首先对实验区的影像数据进行多尺度分割、优化分割结果和分类得到耕地范围。接着利用植被指数变化检测法、分类后变化检测法,根据耕地范围的变化和归一化植被指数(NDVI)在夏、秋两个季节的变化提取撂荒地。最后利用决策树分类法对夏季的光谱特征、植被指数与纹理特征进行分析,提取撂荒地。实验表明,植被指数变化检测法提取精度为85%,更适用于撂荒地信息提取。

The trend of abandoned land is getting more and more serious,and the rapid and accurate information of abandoned land is important to protect arable land and food security.In this paper,using Sentinel-2 satellite image data in different seasons for four years from 2018-2021,we completed the extraction research of vegetation index change detection method,post-classification change detection method and decision tree classification method for abandoned land using eCognition and the environment for vsualizing images(ENVI)software.Firstly,multi-scale segmentation,optimization of segmentation results and classification were performed on the image data of the experimental area to obtain the extent of cultivated land.Then,we extracted the abandoned land according to the change of cultivated area and the change of normalized difference vegetation index(NDVI)in summer and autumn seasons by using the vegetation index change detection method and post-classification change detection method.Finally,the spectral features,vegetation index and texture features in summer were analyzed using the decision tree classification method to extract the abandoned land.The experiment showed that the extraction accuracy of vegetation index change detection method is 85%,which was more suitable for the extraction of abandoned land information.