Efficient Transceiver Design for Aerial Image Transmission and Large-scale Scene Reconstruction

TL;DR

This paper introduces a deep learning-based transceiver that efficiently transmits aerial images and accurately reconstructs large-scale 3D scenes, reducing pilot overhead and enhancing reconstruction fidelity.

Contribution

It presents a novel end-to-end transceiver design integrating 3D Gaussian Splatting into training, enabling sparse pilot schemes and improved scene recovery in aerial imaging.

Findings

Outperforms existing methods in aerial image transmission and 3D scene reconstruction.

Reduces pilot overhead while maintaining high reconstruction accuracy.

Demonstrates superior performance on real-world aerial datasets.

Abstract

Large-scale three-dimensional (3D) scene reconstruction in low-altitude intelligent networks (LAIN) demands highly efficient wireless image transmission. However, existing schemes struggle to balance severe pilot overhead with the transmission accuracy required to maintain reconstruction fidelity. To strike a balance between efficiency and reliability, this paper proposes a novel deep learning-based end-to-end (E2E) transceiver design that integrates 3D Gaussian Splatting (3DGS) directly into the training process. By jointly optimizing the communication modules via the combined 3DGS rendering loss, our approach explicitly improves scene recovery quality. Furthermore, this task-driven framework enables the use of a sparse pilot scheme, significantly reducing transmission overhead while maintaining robust image recovery under low-altitude channel conditions. Extensive experiments on real-world aerial image datasets demonstrate that the proposed E2E design significantly outperforms existing baselines, delivering superior transmission performance and accurate 3D scene reconstructions.