Comparison of TG-1 and EO-1 hyperion in urban land cover classification

The latest successful launch of the high-spatial-resolution hyperspectral satellite TianGong-1 (TG-1) opens up new possibilities for various remote-sensing applications. This study is a precursor to the TG-1 datasets for urban feature analyses. We assessed the performance of the TG-1 sensor system by comparing data acquired using TG-1 to data acquired using EO-1 Hyperion, where both datasets were acquired at approximately the same time and location in Beijing, China. The pixel-based support vector machine (PSVM), pixel-based maximum likelihood classifier (PMLC), hybrid-based support vector machine (hybrid-SVM), and hybrid-based maximum likelihood classifier (hybrid-MLC) methods were implemented to compare the mapping of urban land cover types. The results showed that high-spatial-resolution images from TG-1 outperformed the medium-spatial-resolution images from EO-1 Hyperion using both the pixel-based and hybrid-based approaches. TG-1 with high spatial and spectral resolution imaging provides an opportunity to improve the accuracy of urban land coverage mapping, which has applications in decision-making policies.