Main Crop Classification Based on Multi-temporal Sentinel-2 Data in Chengdu Plain
According to the need of fast and dynamic monitoring of cultivated land in Chengdu Plain in order to support the implemen-tation of the strictest cultivated land protection policy, this paper has studied the method of crop classification based on multi-temporal Sentinel-2 data. The Principal Component Analysis (PCA) has been introduced in the classification process to reduce redundant in-formation and improve classification accuracy. Using 7 Sentinel-2 multispectral images of Chongzhou city in Chengdu Plain in 2021 as the data source, four classification datasets, including time-series multispectral, time-series principal component bands, time-series vegetation index, and typical time-series multispectral+time-series vegetation index were constructed to carry out research on main crop classification based on support vector machine. The results show that the principal component analysis can effectively improve the user accuracy of main crop and reduce the misclassification rate of crops. The datasets based on typical time-phase multispectral+tem-poral vegetation index achieves the highest overall accuracy.
万方数据
维普
