International Conference on Computer and Knowledge Engineering

Home / 12th International Conference on Computer and Knowledge Engineering

Real-Time Vehicle Detection and Classification in UAV imagery Using Improved YOLOv5

Authors :

Mohammad Hossein Hamzenejadi¹ Hadis Mohseni²

1- Department of Computer Engineering, Shahid Bahonar University of Kerman, Kerman, Iran 2- Department of Computer Engineering, Shahid Bahonar University of Kerman, Kerman, Iran

Keywords :

UAV imagery،vehicle detection،YOLOv5،object detection

Abstract :

Recently, Unmanned Aerial Vehicles (UAV) have become useful for various applications. In some real world applications such as UAV based traffic surveillance or disaster management, real-time vehicle detection and classification is an important task. However, vehicles appear too small in aerial UAV imagery which reduces detection accuracy. While modern UAVs are capable of recording high resolution videos with higher spatial information, increasing input size reduces inference speed. So making balance between the accuracy and inference speed is a challenge. In this paper, we address this challenge by proposing an improved version of YOLOv5 single stage object detector, make it suitable for detecting small objects in high resolution images. At the same time, we modified the network width and depth to make it suitable for real-time applications that requires high inference speed. Experiments conducted on VisDrone and CARPK datasets confirm that compared to baseline YOLOv5 models, the proposed model has 3.7\% higher mAP50 and 6.1 FPS higher inference speed on VisDrone dataset while its size is 44.6 MB less than YOLOv5L size. These results confirm the efficacy of the proposed modifications applied on YOLOv5 to make a balance between accuracy and inference time.