Remote Sensing Extraction Methods for Drilling Platforms in South China Sea Based on Multi-Source Satellite Image: Taking the waters around the Mekong Basin as an example

Abstract

In this paper, the waters around the Mekong Basin were taken as the study area and the methods for extracting offshore drilling platforms were discussed. These methods utilized multi-temporal and multi-source remote sensing image. Firstly, the study automated the extraction of drilling platforms by using Gaussian filtering method and two-parameter CFAR method, respectively, based on 27 scenes of multi-temporal nighttime lighting data 1992-2018 and two periods of SAR data in 2019. Then, for areas with deterministic information, visual interpretation was used to extract drilling platforms from long time series Landsat-8 OLI and YG24 optical satellite images. The results were used as standard data to test the accuracy of the previous two methods. Results showed that Nighttime lighting data was easy to obtain, low cost, and simple to extract information. However, the accuracy of SAR image extraction results was higher, but the process was cumbersome. In practice, according to the characteristics of multi-source satellites, it was recommended to combine the advantages of each other to optimize the extraction process and improve the accuracy of the results.

Keywords

• south china sea
• drilling platforms
• remote sensing
• multi-source imagery
• mekong basin