Sat2Graph: Road Graph Extraction through Graph-Tensor Encoding

Label-Pixels is the tool for semantic segmentation of remote sensing images using Fully Convolutional Networks. Initially, it is designed for extracting the road network from remote sensing imagery and now, it can be used to extract different features from remote sensing imagery.

Implementation of the paper "ResUNet-a: a deep learning framework for semantic segmentation of remotely sensed data" in TensorFlow.

This software is a modification of Visualizer from Spacenet competition for road extraction. I made it work for DeepGlobe dataset also.

TensorFlow implementation of D-LinkNet for road extraction.

BTRoadnet

Deep Learning based Road Segmentation using Satellite Imagery