Amazon Archeological Site Discovery - a Deep Learning Approach
-
Updated
Sep 9, 2025 - Python
Amazon Archeological Site Discovery - a Deep Learning Approach
Satellite image classifier that identifies signs of deforestation and pollution using transfer learning with a pre-trained ResNet50 Convolutional Neural Network (CNN) model.
A cost-efficient, AI-powered tool for discovering archaeological sites in the Amazon basin using satellite imagery, LiDAR data, and historical records. Built for the OpenAI to Z Challenge.
This project uses Google Earth Engine's (GEE) platform to analyze deforestation in the Amazon rainforest. The analysis is based on the Hansen Global Forest Change dataset.
A data science project exploring links between Amazon environmental events and disease-related mortality.
"Uma Análise do Sistema Floresta-Atmosfera Amazônica sob a ótica da Teoria do Caos" (Ilum/CNPEM, 2026), de Caio Ruas e Thalles Cansi, com orientação de Vinicius Wasques.
Co-authored workshops on Machine Learning for Geospatial data at the University of Waterloo
COMP 590 - Data Science for Earth @ UNC-CH Final Project, Google Earth Engine, Amazon Rainforest Classification
A machine learning project developed in the context of COP30 to assess socio-environmental risk in Amazonian development projects using environmental indicators.
Open, cross-border library of original seamless surface patterns inspired by the Amazon rainforest. Production-ready for textile and surface printing. CC BY-SA 4.0.
Ambiente de downscaling termico para desagregacao espacial da Temperatura da Superficie Terrestre (LST) de sensores orbitais utilizando preditores multiespectrais de alta resolucao.
Historical analysis of Funai negative certificates and Indigenous lands rights in Brazil (1968-1990)
Fourier Domain Adaptation + Evidential Deep Learning for source-free cross-domain deforestation detection in the Brazilian Amazon using Landsat-8 satellite imagery. ConvNeXt encoder (9.5M params) with per-pixel uncertainty estimation. PyTorch implementation.
Machine Learning Engineer Capstone Project. Multi-Class Deep Learning Classification (CNN) to classify images in the Amazon Rainforest. Kaggle Competition.
I wondered who counts the Amazon's lost trees, so I tried it myself: a U-Net on Sentinel-2 images maps 8,705 ha cleared in Rondônia (2021 to 2026), checked by hand on 300 random points
Aplicação de Data Science para previsão de desmatamento na Amazônia utilizando Random Forest. Inclui pipeline de engenharia de dados, validação temporal e dashboard interativo em Streamlit.
Scripts for the MA research about Brazil’s parliamentary discourses dynamics on the Amazon rainforest.
Windows adaptation of the PyEO forest change detection workflow for Sentinel-2 monitoring in the Amazon BR-163 corridor.
To associate your repository with the amazon-rainforest topic, visit your repo's landing page and select "manage topics."