EVENTO
New Deep Learning Strategies for Deforestation Monitoring using Synthetic Aperture Radar Data
Tipo de evento: Seminário de Avaliação - Série A
Deforestation monitoring is essential for the management of natural resources, the conservation of ecosystems and biodiversity. Change detection techniques employing multi-temporal remote sensing data represent one of the most attractive options for monitoring land changes. This approach involves processing a collection of images of the same geographical area at different dates, facilitating the environmental changes supervision. With the continuous advancement of deep neural networks, numerous solutions based on deep learning have been widely used for this application. However, there are still challenges that need to be addressed. For example, the use of multitemporal imagery has been underexplored until the present moment, since bitemporal change detection is usually emphasized in the literature. Also, studies using deep learning methods for temporal change detection based on multi-sensor data are relatively limited. Considering these challenges, the present thesis aims to develop deep learning solutions for deforestation monitoring using multitemporal and multi-source data. To the present moment, the focus was on formulating the change detection architectures. Novel models were developed by combining recurrent and residual learning, and also using attention mechanisms. The results obtained with these developed architectures were compared with state-of-art approaches. Experiments were conducted using bitemporal and multitemporal Sentinel-1 data from two sample sites of the Amazon Forest. The use of temporal distance maps containing information about previous deforestation of the areas was also investigated. The next steps of this research include the use of image fusion techniques for employing multi-source data (SAR and optical) in the deforestation monitoring and the evaluation of the proposed models for predicting future deforestation.Para assistir acesse:meet.google.com/owk-zgvh-dsx
Data Início: 24/09/2024 Hora: 09:00 Data Fim: 24/09/2024 Hora: 12:00
Local: LNCC - Laboratório Nacional de Computação Ciêntifica - Virtual
Aluno: Carla Nascimento Neves - - LNCC
Orientador: Gilson Antônio Giraldi - Laboratório Nacional de Computação Científica - LNCC Raul Queiroz Feitosa - Pontifícia Universidade Católica do Rio de Janeiro - PUC-RIO
Participante Banca Examinadora: Gilson Alexandre Ostwald Pedro da Costa - Universidade do Estado do Rio de Janeiro - UERJ Gilson Antônio Giraldi - Laboratório Nacional de Computação Científica - LNCC José Marcato Junior - Universidade Federal de Mato Grosso do Sul - UFMS Roberto Pinto Souto - Laboratório Nacional de Computação Científica - LNCC
Suplente Banca Examinadora: Bruno Richard Schulze - Laboratório Nacional de Computação Científica - LNCC
