Band-Adaptive Spectral-Spatial Feature Learning Deep Neural Network for Hyperspectral Image Classification

codes for TGRS paper: Graph Relation Network: Modeling Relations between Scenes for Multi-Label Remote Sensing Image Classification and Retrieval

A TensorFlow implentation of fixed size kernel CNN

Pytorch code for the paper "The color out of space: learning self-supervised representations for Earth Observation imagery"

An implementation of the neural network described in "Convolution Based Spectral Partitioning Architecture for Hyperspectral Image Classification"

Source code for the paper, "Water Body Extraction from Sentinel-2 Imagery with Deep Convolutional Networks and Pixelwise Category Transplantation".

Open source canopy classification system

In order to map LCLU in french-Guyana, few scripts were developped or adapted to enable either to automaticaly map either to explore cloudless mosaic and even automaticaly detect floodings with Sentinel 1 SAR data.

Landcover classification on sentinel-2 data with Prithvi, EfficientNet-Unet and OSM / CNES Landcover labels.

Rough implementation of the Automated landcover classification using unsupervised classification methods.

The Supervised Land Cover Classification (SLaCC) tool is a Google Earth Engine script created by the Summer 2019 Southern Maine Health and Air Quality Team. It uses NASA Earth observations, the National Land Cover Database, land cover classification training data, and a shapefile of Cumberland County, Maine, USA. The goal of the project was to e…

GEE code for pixel-based land cover classification with Random Forest (RF) algorithm, and for NDVI time series visualization.

generates rasters of Köppen-Geiger land-cover classification

A TensorFlow implentation of fixed size kernel CNN

Land Use /Land Cover Classification using PyTorch with the RGB EuroSat Dataset

This repository contains three different models (ResNet-18, ResNet-50, and ViT-Base-Patch16-224) fine-tuned on the EuroSAT dataset, along with their performance comparisons.

The aim of this project was to create a land cover classification of the area near Surat in India for 3 timesteps (2015, 2018, 2022) using a Random Forest classifier to access the process of urbanization

Future Urban-Wildfire Risk Mapping (FUWRM), pronounced as "form". This repository holds the programming script files and some of the binaries that represent the predictive risk maps for wildfires in urban regions of Southern Victoria (AUS) and Northern California (USA) in 2030 and 2040.

This codebase is actively being developed as part of my master's thesis on Knowledge-based semantic enrichment for semantic image segmentation for the task of land cover classification.

A surface cover product dynamic update system utilizing the CCDC algorithm and Random Forest tree algorithm.