Usando dados de satélite para rastrear a pegada humana na floresta amazônica.
Objetivos:
- Desenvolver um programa para calcular o percentual de floresta densa de uma imagem. a. Preferencialmente com segmentação de imagens por cores.
- Desafio: Readequar o programa para segmentar a imagem nas diferentes categorias listadas na aba "Data" da página do Kaggle.
Data Labels
- Primary
- Water
- Habitation
- Agriculture
- Road
Data source: Kaggle - Planet: Understanding the Amazon from Space
- DEDOV, Florian. Image Segmentation with K-Means Clustering in Python. YouTube, 2023. Disponível em: https://www.youtube.com/watch?v=X-Y91ddBqaQ. Acesso em: 9 de junho de 2024.
- BHATTIPROLU, Sreenivas . Labeling images for semantic segmentation using Label Studio. YouTube, 2022. Disponível em: https://www.youtube.com/watch?v=UUP_omOSKuc. Acesso em: 17 jun. 2024.
- YADAV, Bhimraj . How to Implement UNet in PyTorch for Image Segmentation from Scratch?. Bhimraj Yadav, 2023. Disponível em: https://bhimraj.com.np/blog/pytorch-unet-image-segmentation-implementation. Acesso em: 18 jun. 2024.
- GSI Technology: A Beginner’s Guide To Segmentation In Satellite Images: Walking Through Machine Learning Techniques For Image Segmentation And Applying Them To Satellite Imagery
- (Video) MATLAB: Semantic Segmentation of Satellite Images
- Satellite Image Semantic Segmentation - Eric Guérin, Killian Oechslin, Christian Wolf, Benoît Martinez.
- A brief introduction to satellite image segmentation with neural networks - Robin Cole.
- How to do Semantic Segmentation using Deep learning
- UNet
- ResNet
- Global Convolution Network