Repo for machine and deep learning methods for geospatial analysis in real-world applications. Some of the projects intended to be developed are:
- Automatic classification of land covers by means of Convolutional Neural Networks, both pre-trained models and original developments.
- Time series analysis using Recurrent Neural Networks approach.
- Object detection and recognition in drone-based imagery.
The main programming framework is based on Python and all the available resources for Deep Learning implementations and numerical analysis like Tensorflow-keras, Scikit-learn, pandas, numpy, scipy.