Obtaining spatial distribution information on mariculture in a low-cost, fast, and efficient manner is crucial for the sustainable development and regulatory planning of coastal zones and mariculture industries. This study, based on the Segment Anything Model (SAM) and high-resolution remote sensing imagery, rapidly extracted mariculture areas in Liaoning Province, a typical northern province in China with significant mariculture activity. Additionally, it explored the actual marine ownership data to investigate the marine use status of Liaoning Province’s mariculture. The total area of mariculture we extracted in Liaoning Province is 1052.89 km². Among this, the area of cage mariculture is 27.1 km², while raft mariculture covers 1025.79 km². Through field investigations, it was determined that in the western part of Liaodong Bay, cage mariculture predominantly involves sea cucumbers. In the southern end of Dalian, the raft mariculture focuses on cultivating kelp. On the other hand, around the islands in the eastern region, the primary crop in raft mariculture is scallops, showing a significant geographical differentiation pattern. In the planned mariculture areas within Liaoning Province’s waters, the proportion of actual development and utilization is 11.2%, while the proportion approved for actual mariculture is 90.2%. This indicates a suspicion that 9.8% of mariculture is possibly in violation of sea occupation rights, which could be due to the untimely updating of marine ownership data. Based on SAM, efficient and accurate extraction of cage mariculture can be achieved. However, the extraction performance for raft mariculture is challenging and remains unsatisfactory. Manual interpretation is still required for satisfactory results in this context.

In the face of growing demand for marine nutrition and restrictions on wild fish capture, mariculture offers significant potential to enhance marine food production. The natural environment directly influences the growth of marine organisms, but socio-economic conditions are equally critical for sustainable and efficient development. This study investigated the multifaceted factors shaping mariculture in China, Vietnam, and India. We analyzed the natural environment and social economy qualitatively and quantitatively by adopting zonal statistics, chart analysis, and correlation analysis. Results show that all three countries generally possess suitable marine environments for the growth of cultured species. Meanwhile, China and Vietnam demonstrate how robust socio-economic systems and strategic policy support drive successful mariculture development, whereas India’s comparatively underdeveloped socio-economic foundation appears to constrain its sectoral advancement. These analyses suggest a principle: the natural environment is the necessary condition and the social economy serves as the sufficient condition, together determining the state of mariculture. Our study highlights the joint role of environmental suitability, socio-economic readiness, and policy frameworks, providing valuable insights for identifying potential mariculture sites and informing policy strategies to promote sustainable marine aquaculture globally.

As an important part of the global food production sector, the sustainable development of mariculture urgently requires accurate spatial distribution information of mariculture activities, which is critical for refined spatial management of coastal zones and ecosystem-based marine governance. However, current classification methods based on a single remote sensing data source have limitations in distinguishing specific mariculture species. This restricts the spatial quantitative analysis and refined management of mariculture systems. In contrast, statistical yearbooks provide attribute information of cultured species that cannot be obtained by remote sensing. Based on this, this study selected Sansha Bay, a typical mariculture area in China, as the research region and proposed a classification framework integrating multi-source data. First, subclassification of mariculture facilities was achieved based on remote sensing image features. Then, remote sensing image features, mariculture statistical data, knowledge of water depth ecological suitability, and knowledge of mariculture planning distribution were quantified into specific indicators. On the basis, these quantified indicators were further combined through a linear weighted fusion mechanism to construct the Multi-Rule Constrained Mariculture Species Discrimination Model (MRC-MSDM), which realizes refined discrimination of mariculture species in the study area. It achieved an overall classification accuracy of over 93%, a significant improvement compared with single machine learning models, reflecting the enhancing effect of domain rules in classifying complex scenarios and providing reliable methodological support for large-scale accurate discrimination of mariculture species and refined management of coastal mariculture systems.