DSM Refinement with Deep Encoder-Decoder Networks

TL;DR

This paper introduces a deep encoder-decoder neural network approach to automatically refine noisy and artifact-laden digital surface models (DSMs) derived from aerial images, reducing manual cleanup effort.

Contribution

It proposes a novel loss function combining L1 and feature loss, and a deep residual encoder-decoder architecture tailored for urban DSM refinement.

Findings

Effective noise and artefact removal from DSMs

Preservation of geometric structures in urban areas

Automated refinement reduces manual editing time

Abstract

3D city models can be generated from aerial images. However, the calculated DSMs suffer from noise, artefacts, and data holes that have to be manually cleaned up in a time-consuming process. This work presents an approach that automatically refines such DSMs. The key idea is to teach a neural network the characteristics of urban area from reference data. In order to achieve this goal, a loss function consisting of an L1 norm and a feature loss is proposed. These features are constructed using a pre-trained image classification network. To learn to update the height maps, the network architecture is set up based on the concept of deep residual learning and an encoder-decoder structure. The results show that this combination is highly effective in preserving the relevant geometric structures while removing the undesired artefacts and noise.