Ⅰ. Aims and Purposes of the Research
ㅇ Various development projects in the coastal areas have been carried out, and mitigation methods to reduce their impact on the environment have been under development based on the prediction and evaluation.
ㅇ There are many limitations in understanding the impacts of development projects on the marine environment.
ㅇ We intend to develop a marine pollution prediction tool by applying data from the Marine Environment Information System (MEIS), satellite data, and physical data calculated from the numerical model to deep learning technology.

Ⅱ. Domestic and Foreign Status and Case Analysis
ㅇ Deep learning studies applied in the marine field were classified and organized into prediction (including missing value correction) and classification studies.
ㅇ Deep learning research conducted in fields other than the marine field is also briefly summarized.

Ⅲ. Methods
1. Marine environment data
ㅇ Data from the MEIS which provides various data related to the marine environment, marine ecology, marine protected areas, marine environment information map, and marine waste discharge were investigated.
ㅇSatellite data produced through the Communication, Ocean and Meteorological Satellite (COMS) operated since 2010 were reviewed.
2. Numerical model
ㅇ Delft3D, widely used in coastal areas, is described as a numerical model that simulates changes in the marine environment such as hydrodynamics, sediment transport, waves, and water quality.
3. Deep learning model
ㅇ In regard to deep learning models, only the main models are explained because the mod