Banca de QUALIFICAÇÃO: MATEUS RIBEIRO DA SILVA

Uma banca de QUALIFICAÇÃO de MESTRADO foi cadastrada pelo programa.
STUDENT : MATEUS RIBEIRO DA SILVA
DATE: 06/04/2025
TIME: 09:00
LOCAL: Videoconferência
TITLE:

APPLICATION OF MACHINE LEARNING FOR ESTIMATION OF SATELLITE-DERIVED BATHYMETRY IN SHALLOW AREAS OF THE POTIGUAR CONTINENTAL SHELF

KEY WORDS:

Machine Learning; Satellite Derived Bathymetry; Neural Networks.

PAGES: 47
BIG AREA: Ciências Exatas e da Terra
AREA: Geociências
SUMMARY:

The study compares the performance of Artificial Neural Networks (ANN) and Linear Regression models applied to images from the Landsat 8 and Sentinel 2A satellites in an area located between the municipalities of Caiçara do Norte and São Bento do Norte (RN). The results demonstrate that the ANNs outperformed the regression models, with Landsat 8 obtaining the best results (RMSE of 4.1046m, MAE of 2.6992m and R² of 0.9443). The analysis by depth ranges revealed greater accuracy in shallow waters (up to 25m), with errors increasing significantly at depths greater than 40m. The study contributes to the development of low-cost alternative methods for bathymetric mapping, which can assist in the monitoring of coastal areas and planning of offshore wind energy projects in the region

COMMITTEE MEMBERS:
Presidente - 2218779 - HELENICE VITAL
Interno - 2042405 - MOAB PRAXEDES GOMES
Externo à Instituição - PAULO VICTOR DO NASCIMENTO ARAÚJO