Abstract

Unmanned Aerial Vehicles (UAVs) can serve as an ideal mobile platform in various situations. Real-time object detection with on-board apparatus provides drones with increased flexibility as well as a higher intelligence level. In order to achieve good detection results in UAV images with complex ground scenes, small object size and high object density, most of the previous work introduced models with higher computational burdens, making deployment on mobile platforms more difficult.This paper puts forward a lightweight object detection framework. Besides being anchor-free, the framework is based on a lightweight backbone and a simultaneous up-sampling and detection module to form a more efficient detection architecture. Meanwhile, we add an objectness branch to assist the multi-class center point prediction, which notably improves the detection accuracy and only takes up very little computing resources. The results of the experiment indicate that the computational cost of this paper is 92.78% lower than the CenterNet with ResNet18 backbone, and the mAP is 2.8 points higher on the Visdrone-2018-VID dataset. A frame rate of about 220 FPS is achieved. Additionally, we perform ablation experiments to check on the validity of each part, and the method we propose is compared with other representative lightweight object detection methods on UAV image datasets.

Highlights

  • With the advance of Unmanned Aerial Vehicles (UAVs) technology and the growth of UAV suppliers, UAVs are becoming more cost-efficient

  • Compared with the CenterNet, which has a ResNet18 backbone, our proposed model reduces the computational cost by 92.78%, reduces the parameter size by 86.73%, and improves the mean average precision (mAP) by 2.8 points on the UAV image dataset

  • The model we proposed is compared with the baseline approaches on the Visdrone-2018-VID and UAVDT-DET datasets

Read more

Summary

Introduction

With the advance of UAV technology and the growth of UAV suppliers, UAVs are becoming more cost-efficient. Due to their mobility, being autonomous, and their processing capabilities, UAVs are considered in many intelligent transportation system (ITS) application domains [1], such as traffic state estimation, traffic control, incidence emergency response and so on. The detection of objects of interest from UAV images/videos is the initialization process of traffic state estimation [5], which provides fast and accurate traffic data collection. Online object detection based on UAV onboard platforms is of great importance, to improve the flexibility of UAV

Methods
Results
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call