Mesh Reconstruction from Aerial Images for Outdoor Terrain Mapping Using Joint 2D-3D Learning

TL;DR

This paper introduces a joint 2D-3D learning method for reconstructing local meshes from aerial images, enabling efficient outdoor terrain mapping suitable for environmental monitoring and surveillance.

Contribution

It presents a novel approach that combines 2D image features and 3D mesh refinement to produce detailed terrain models from aerial imagery.

Findings

Effective local mesh reconstruction from aerial images

Improved accuracy over traditional sparse methods

Potential for real-time environmental monitoring

Abstract

This paper addresses outdoor terrain mapping using overhead images obtained from an unmanned aerial vehicle. Dense depth estimation from aerial images during flight is challenging. While feature-based localization and mapping techniques can deliver real-time odometry and sparse points reconstruction, a dense environment model is generally recovered offline with significant computation and storage. This paper develops a joint 2D-3D learning approach to reconstruct local meshes at each camera keyframe, which can be assembled into a global environment model. Each local mesh is initialized from sparse depth measurements. We associate image features with the mesh vertices through camera projection and apply graph convolution to refine the mesh vertices based on joint 2-D reprojected depth and 3-D mesh supervision. Quantitative and qualitative evaluations using real aerial images show the potential of our method to support environmental monitoring and surveillance applications.