The Xiamen Symposium on Marine Environmental Sciences (XMAS 2025)

  The pressing need for accurate mapping of mariculture distribution has been long overdue, as the rapid expansion of mariculture and increasing conflicts over sea use demand science-based marine spatial planning, best facilitated through satellite imagery analysis. Previous research has primarily focused on specific regions with limited spatial coverage, leaving the broader applicability and verification of methods largely unexplored. Here, we propose a novel method, ETCH, designed to extract mariculture infrastructures from publicly accessible satellite images, such as Landsat 8. ETCH employs a top-down optimization approach, leveraging the unique features of various mariculture infrastructure types to refine extraction results from coarse to fine detail. Our method has demonstrated efficient performance in eight small-scale tests and has surpassed conventional OBIA extraction techniques in four case studies across different regions in China. ETCH offers a rapid and accurate solution for extracting mariculture infrastructure and could serve as a reference for similar remote sensing applications.

Remote sensing (RS),Aquaculture