Towards Robust Drone Vision in the Wild

TL;DR

This paper introduces a new drone vision dataset for image super-resolution, analyzes domain gaps at different altitudes, and proposes methods to enhance super-resolution robustness across varying drone altitudes.

Contribution

It presents the first drone-specific super-resolution dataset, identifies altitude-related domain gaps, and proposes altitude-aware and one-shot learning methods for robust super-resolution.

Findings

Super-resolution performance drops across different altitudes.

Altitude-aware layers improve super-resolution at various altitudes.

One-shot learning enables quick adaptation to new altitudes.

Abstract

The past few years have witnessed the burst of drone-based applications where computer vision plays an essential role. However, most public drone-based vision datasets focus on detection and tracking. On the other hand, the performance of most existing image super-resolution methods is sensitive to the dataset, specifically, the degradation model between high-resolution and low-resolution images. In this thesis, we propose the first image super-resolution dataset for drone vision. Image pairs are captured by two cameras on the drone with different focal lengths. We collect data at different altitudes and then propose pre-processing steps to align image pairs. Extensive empirical studies show domain gaps exist among images captured at different altitudes. Meanwhile, the performance of pretrained image super-resolution networks also suffers a drop on our dataset and varies among altitudes. Finally, we propose two methods to build a robust image super-resolution network at different altitudes. The first feeds altitude information into the network through altitude-aware layers. The second uses one-shot learning to quickly adapt the super-resolution model to unknown altitudes. Our results reveal that the proposed methods can efficiently improve the performance of super-resolution networks at varying altitudes.