Properties of BV-G structures + textures decomposition models. Application to road detection in satellite images

TL;DR

This paper provides theoretical insights into a structures-textures decomposition model and develops an algorithm for detecting road networks in satellite images based on these insights.

Contribution

It offers new theoretical results on a structures-textures decomposition model and applies these findings to improve road detection algorithms in satellite imagery.

Findings

Theoretical characterization of the BV-G structures-textures model

An algorithm for detecting long, thin objects in satellite images

Successful application to road network detection

Abstract

In this paper we present some theoretical results about a structures-textures image decomposition model which was proposed by the second author. We prove a theorem which gives the behavior of this model in different cases. Finally, as a consequence of the theorem we derive an algorithm for the detection of long and thin objects applied to a road networks detection application in aerial or satellite images.