An Adaptive Deblurring Vehicle Detection Method for High-Speed Moving Drones: Resistance to Shake.

Yan Liu,Tiantian Qiu,Wenting Qi,Jingwen Wang

doi:10.3390/e23101358

Abstract

Vehicle detection is an essential part of an intelligent traffic system, which is an important research field in drone application. Because unmanned aerial vehicles (UAVs) are rarely configured with stable camera platforms, aerial images are easily blurred. There is a challenge for detectors to accurately locate vehicles in blurred images in the target detection process. To improve the detection performance of blurred images, an end-to-end adaptive vehicle detection algorithm (DCNet) for drones is proposed in this article. First, the clarity evaluation module is used to determine adaptively whether the input image is a blurred image using improved information entropy. An improved GAN called Drone-GAN is proposed to enhance the vehicle features of blurred images. Extensive experiments were performed, the results of which show that the proposed method can detect both blurred and clear images well in poor environments (complex illumination and occlusion). The detector proposed achieves larger gains compared with SOTA detectors. The proposed method can enhance the vehicle feature details in blurred images effectively and improve the detection accuracy of blurred aerial images, which shows good performance with regard to resistance to shake.

Highlights

In order to explain the performance of the proposed algorithm, DCNet was used on the constructed test dataset
Figure 6n–p are the vehicle detection results in a blurred case, which shows that the model has good performance in blurred inputs
Due to motion blurring generated by high-speed motion of unmanned aerial vehicles (UAVs) or targets, an adaptive deblurring vehicle detection method for high-speed moving drones called DCNet is proposed, which aims to solve the problem of low vehicle detection rate

Summary

Methods

The Drone-GAN module improves mAP by 0.80% (from 28.63% to 29.43%), which improvesthe To 29.43%), which means that the modulemodule greatly improves detection performance of motion-blurred tarmeans that the greatly improves the detection performance of motion-blurred gets. Aiming at module the problem of target detection that inevitably produces blurred images targets. Aiming at the problem of target detection that inevitably produces blurred images in the UAV scene, a high-precision model was successfully generated. DCNet method achieves good performance in blurred and clear demonstrates that the DCNet method achieves good performance in blurred and clear targets detection. The proposed model has obtained relatively good results, it remains a chalthe proposed model obtained relatively good results, it remains lengeAlthough to accurately detect objects in has the VisDrone datasets with smaller targets and ina challenge to accurately in theaccurate

Results

Discussion

Conclusion