Banca de QUALIFICAÇÃO: DANIEL CÉSAR BARROS TORRES

Uma banca de QUALIFICAÇÃO de MESTRADO foi cadastrada pelo programa.
STUDENT : DANIEL CÉSAR BARROS TORRES
DATE: 29/02/2024
TIME: 09:00
LOCAL: Google meet
TITLE:

ESTIMATION OF FOREST BIOMASS USING SATELLITE IMAGES IN MANAGEMENT PLANS AND CAATINGA INVENTORIES

KEY WORDS:

mapping, stock, machine learning, semi-arid

PAGES: 26
BIG AREA: Ciências Agrárias
AREA: Recursos Florestais e Engenharia Florestal
SUMMARY:

Due to high cost and longer time demands, accurate estimation of plant biomass of planted or natural stands in large forest areas is still a significant challenge. The evolution of low-cost geotechnologies is rapidly facilitating this task, but their use to estimate biomass in tropical dry forests is still scarce. Therefore, it is necessary to develop machine learning tools that are capable of monitoring the complex spatio-temporal dynamics existing in these highly variable ecosystems. The present work aims to develop a methodology to estimate above-ground biomass stocks in the Caatinga, native or planted in the semi-arid region of Northeast Brazil, using field data from Inventories, Forest Management Plans and seeking to create an estimation model with the use of neural networks. For this to occur, the selected models will be processed to estimate the biomass in comparison with the data already obtained in the inventories carried out in the forest management plans, hoping that the regression models will have the capacity to predict the biomass of the Caatinga in its different extracts. Using the equations that will be defined and the map of the Annual Land Cover and Use classification (Mapbiomas), the development of a mapping representing the biomass stock is expected. This model provides for mapping in a short time and with low survey costs, thus providing reliable estimates and contributing to the monitoring of biomass dynamics in the semi-arid region of Northeastern Brazil.

COMMITTEE MEMBERS:
Presidente - ***.339.107-** - GETULIO FONSECA DOMINGUES - UFMG
Interno - 1305184 - JHONES DA SILVA AMORIM
Externo à Instituição - GUSTAVO EDUARDO MARCATTI