Potential of SDGSAT-1 nighttime light data in extracting urban main roads

Abstract

The Sustainable Development Science Satellite 1 (SDGSAT-1) provides a novel nighttime light (NTL) data product with medium spatial resolution, captured by its unique Glimmer Imager (GLI) sensor. Unlike traditional NTL products, the exceptional resolution of SDGSAT-1 NTL data allows for distinct visualization of urban road networks. Although recent studies have validated the effectiveness of SDGSAT-1 NTL data in supporting various sustainable development goals, their potential for urban road extraction has not yet been thoroughly explored. To address this gap, we propose a novel terrain skeleton-based method for extracting urban main roads from SDGSAT-1 NTL images. This proposed method innovatively uses a terrain analogy, considering SDGSAT-1 NTL data as a continuous terrain surface and urban roads as terrain ridge lines to facilitate road extraction. To validate this approach, we selected nine cities with diverse sizes and complex road networks—six in China and three in the United States. Extensive experimental results showed that the proposed method effectively extracts urban roads with an average accuracy of 85.14 % using red-green-blue (RGB) bands and 83.99 % using panchromatic bands, outperforming previous methods, including the optimal threshold, line segment detector, watershed, and U-Net. The main road types extracted were residential, tertiary, secondary, and primary. Additionally, our findings indicated that SDGSAT-1 NTL data capture over 82 % of city road networks, significantly surpassing the coverage provided by the DMSP/OLS, NPP-VIIRS, and Luojia1–01 NTL data. Overall, this study confirms that the significant potential of SDGSAT-1 NTL data for urban main road extraction, offering valuable insights for improving infrastructure mapping and urban planning.