Suspected coastal reclamation areas detection by spatiotemporal fusion of landsat TM and Sentinel-1A: A case study of Fuzhou, Fujian, China

Abstract

Increasing coastal reclamation activities have been undertaken to alleviate the shortage of land resources, posing significant challenges to the coordination of coastal reclamation, economic development and environmental protection. Over the past few decades, coastal reclamation has occurred extensively in the coastal region of China, especially the east coast, while the value of coastal wetlands and ecosystems has only been well understood and appreciated in the last decade. More recently, there is an important new trend to restore parts of reclaimed land to coastal wetlands for environmental protection purposes, which requires reliable reclamation detection results as a foundation for policy. This study aimed to propose a reclamation detection framework based on the spatiotemporal fusion of Landsat TM and Sentinel-1A. To utilize the complementary information of optical and SAR images, the proposed framework models the common and proprietary information of both types of data separately in a sparse manner. Then, with more reliable and detailed information, reclamation areas can be distinguished by band math and threshold segmentation using the spatiotemporal fusion of Landsat TM and Sentinel-1A images. Finally, taking Fuzhou in eastern China as an example, we provided a detailed analysis of coastal reclamation. The results confirmed that the proposed algorithm was effective for rapidly detecting suspected locations of coastal reclamation and showed great potential for monitoring the changes occurring in the coastal zone.