Towards a High-Performance Object Detector: Insights from Drone Detection Using ViT and CNN-based Deep Learning Models

TL;DR

This paper compares CNN and Vision Transformer models for drone detection, showing ViT's superior accuracy in single-drone detection but highlighting its higher data and computational requirements, with state-of-the-art results in multi-drone detection.

Contribution

It provides a comprehensive comparison of ViT and CNN models for drone detection, including new insights into their performance and training requirements.

Findings

ViT achieves 4.6 times more robustness than CNN in single-drone detection.

YOLO v7 and YOLOS attain 98% and 96% mAP in multi-drone detection.

ViT requires more training data and computational power to outperform CNN.

Abstract

Accurate drone detection is strongly desired in drone collision avoidance, drone defense and autonomous Unmanned Aerial Vehicle (UAV) self-landing. With the recent emergence of the Vision Transformer (ViT), this critical task is reassessed in this paper using a UAV dataset composed of 1359 drone photos. We construct various CNN and ViT-based models, demonstrating that for single-drone detection, a basic ViT can achieve performance 4.6 times more robust than our best CNN-based transfer learning models. By implementing the state-of-the-art You Only Look Once (YOLO v7, 200 epochs) and the experimental ViT-based You Only Look At One Sequence (YOLOS, 20 epochs) in multi-drone detection, we attain impressive 98% and 96% mAP values, respectively. We find that ViT outperforms CNN at the same epoch, but also requires more training data, computational power, and sophisticated, performance-oriented designs to fully surpass the capabilities of cutting-edge CNN detectors. We summarize the distinct characteristics of ViT and CNN models to aid future researchers in developing more efficient deep learning models.