진학프로석·박사 채용 접수 플랫폼

Accurate analysis and interpretation of meteorological data obtained through radar are fundamental for weather forecasting and monitoring severe weather events. Radar images, especially those of the PPI (Plan Position Indicator) and SAR (Synthetic Aperture Radar) types, often present significant challenges that compromise their interpretation, including speckle noise, attenuation effects, altitude variations, and artifacts caused by atmospheric interference. Furthermore, the inherent complexity of different radar types and their specific configurations hinders the creation of generalizable solutions for processing these images. This project proposes an innovative approach that combines state-of-the-art diffusion models with physical radar knowledge and advanced transfer learning techniques. The methodology incorporates fundamental radar wave propagation equations into the diffusion process, allowing for more accurate and physically consistent image restoration. The developed framework will be able to transfer knowledge between different radar types (PPI, SAR, Doppler) and adapt to different meteorological conditions through a continuous feedback system based on radar-specific metrics. Please, send your CV (Lattes if you are Brazilian) to the email joao.papa@unesp.br

Skills/Qualifications

Experience in machine learning, math, and programming.