High-precision automated extraction of offshore aquaculture areas based on GF-2 remote sensing imagery

Remote sensing technology plays an important role in studying the impacts of offshore aquaculture on the marine environment and traffic. This study focused on Sandu'ao, Ningde City, Fujian Province, China, using GF-2 remote sensing imagery to automatically extract offshore aquaculture areas with high accuracy. This was achieved through the construction of ratio indices, texture information analysis, and histogram threshold segmentation, combined with the geometric characteristics (area parameter, degree of rectangularity) and spatial distribution characteristics (degree of aggregation) of aquaculture areas. The developed ArcGIS model for automated extraction of aquaculture areas demonstrated that the proposed method can accurately extract both raft aquaculture areas and cage aquaculture areas, achieving a classification accuracy of over 90%. The results showed that raft aquaculture was the dominant type of aquaculture in the study area, occupying six times the area of cage aquaculture, and accounting for 12.92% of the total aquaculture area. This study provides reliable data for assessing the impact of offshore aquaculture on the environment and marine traffic, which will assist in the scientific planning of aquaculture layout and the optimal use of marine resources.