Author Description

Data-driven ocean environmental perception refers to the various data-driven methodologies used by researchers to study, understand, and explore the ocean for sustainable marine ecosystem, the blue ocean economy, and other ocean-related activities. Ocean environment perception techniques also pave the way for deep insight and full exploration of ocean resources, such as coastal and marine ecosystem protection, ocean vehicle operation and management, and ocean energy utilization. However, it is hard for traditional data processing techniques to get comprehensive understanding of the ocean environmental information attributed to the complex characteristics of environmental information. The traditional data processing techniques are facing challenges such as: • Unable to fully capture the non-linearity and intrinsic uncertain of complex ocean systems; • Isolated systems are involved and make it difficult to integrate data from various sources; • Overwhelming for handling enormous amounts of data, leading to performance issues and long processing times; • Costly in scalability both in terms of hardware and operational expenses. Advances in data-driven methodologies: • Highly dynamic nonlinearity and uncertainty mapping capability in complex system evolution; • Leveraging vast amounts of data collected from various sources effectively; • Flexibility in handling various types of data, including structured, semi-structured, and unstructured data; • Handling big data efficiently by making use of distributed computing and storage. Data-driven technologies are a set of techniques dealing with complex systems, which can mine valuable potential relationships from measurement data through intelligent computational approaches such as machine learning, data mining, data fusion, reinforcement learning, etc. Data-driven applications on the aspect of Remote Sensing, Autonomous Underwater Vehicles (AUVs),