| Grant number: | 23/00782-0 |
| Support Opportunities: | Regular Research Grants |
| Start date: | April 01, 2024 |
| End date: | March 31, 2026 |
| Field of knowledge: | Physical Sciences and Mathematics - Computer Science - Computer Systems |
| Agreement: | MCTI/MC |
| Principal Investigator: | Álvaro Luiz Fazenda |
| Grantee: | Álvaro Luiz Fazenda |
| Host Institution: | Instituto de Ciência e Tecnologia (ICT). Universidade Federal de São Paulo (UNIFESP). Campus São José dos Campos. São José dos Campos , SP, Brazil |
| Associated researchers: | Fabio Augusto Faria ; Fernanda Beatriz Jordan Rojas Dallaqua ; Hugo Resende |
| Associated scholarship(s): | 24/06790-8 - Projeto ForestEyes - Using SAR Radar images to detect deforestation, BP.TT |
Abstract
The conservation of Tropical Forests is a key issue due to their importance in the global ecosystem since they contain great biodiversity, carbon storage, rainfall regulation in adjacent regions, and indigenous peoples sheltering.Unfortunately, millions of hectares of tropical forests are lost each year through deforestation and degradation for different and complex reasons.In this research project, a proposal to enhance a novel monitoring system prototype, called ForestEyes, is presented, which aims to combine citizen science and machine learning in detecting deforestation.Citizen Science allows ordinary volunteers to collaborate by inspecting remote sensing image segments over a forest region, seeking to detect deforested or degraded areas, thus participating in a scientific experiment aimed at improving the planet's sustainability, also acquiring more awareness and involvement with the major issue.Machine Learning techniques will be applied in a forest area by using trained databases with samples labeled by volunteers in the previous phase, expanding the coverage for the detection area.Initial experiments over six official campaigns, demonstrated through solid published scientific works, obtained more than 81,000 contributions from 644 different volunteers, allowing to build of a solid training base for classification models, whose results were compared with the official program for monitoring the Brazilian Legal Amazon (PRODES).The methodology developed proved to be promising, both from the point of view of the volunteer's accuracy and the classification techniques applied.This project proposes improvements to the aforementioned project by expanding the application's database and investigating the use of new remote sensing images from multiple optical and radar sensors (SAR), and different regions and periods of time, which implies scientific and technological challenges in the pre-processing, citizen science and machine learning phase. (AU)
| Articles published in Agência FAPESP Newsletter about the research grant: |
| More itemsLess items |
| TITULO |
| Articles published in other media outlets ( ): |
| More itemsLess items |
| VEICULO: TITULO (DATA) |
| VEICULO: TITULO (DATA) |