Gaining Scale Invariance in UAV Bird's Eye View Object Detection by Adaptive Resizing

TL;DR

This paper presents Adaptive Resizing, a preprocessing technique for UAV bird's eye view object detection that enhances scale invariance and inference speed, tested across multiple datasets and tasks.

Contribution

It introduces Adaptive Resizing, a novel preprocessing step that improves scale robustness and speed in UAV object detection, applicable to various datasets and tasks.

Findings

Consistent performance improvements on UAVDT and VisDrone datasets.

Inference speed increased by 2-3 times on average.

Effective application to height transfer tasks and embedded GPU deployment.

Abstract

This work introduces a new preprocessing step for object detection applicable to UAV bird's eye view imagery, which we call Adaptive Resizing. By design, it helps alleviate the challenges coming with the vast variances in objects' scales, naturally inherent to UAV data sets. Furthermore, it improves inference speed by two to three times on average. We test this extensively on UAVDT, VisDrone, and on a new data set we captured ourselves and achieve consistent improvements while being considerably faster. Moreover, we show how to apply this method to generic UAV object detection tasks. Additionally, we successfully test our approach on a height transfer task where we train on some interval of altitudes and test on a different one. Furthermore, we introduce a small, fast detector meant for deployment to an embedded GPU. Code will be made publicly available on our website.