Research on automatic driving scene object detection algorithm based on improved YOLOv8

Yunxiang Liu; Xiaoya Ren; Jianlin Zhu

doi:10.1117/12.3055407

12 December 2024 Research on automatic driving scene object detection algorithm based on improved YOLOv8

Yunxiang Liu, Xiaoya Ren, Jianlin Zhu

Proceedings Volume 13439, Fourth International Conference on Testing Technology and Automation Engineering (TTAE 2024); 1343931 (2024) https://doi.org/10.1117/12.3055407
Event: Fourth International Conference on Testing Technology and Automation Engineering (TTAE 2024), 2024, Xiamen, China

Abstract

In response to the challenges associated with detecting occluded and small targets in automatic driving scenarios, as well as the issues of low detection accuracy and a high miss rate caused by complex background interference, an enhanced road target detection algorithm based on YOLOv8 is proposed. A plural lightweight convolutional module (PLConv) is devised to replace the C2f module with the PL-C3 module, thereby reducing network parameters and enhancing the network's feature extraction capabilities. Moreover, a P2 small target detection head is integrated at the model's apex, facilitating improved extraction of shallow features and enhancing the model's performance in detecting small targets. Experimental results demonstrate that the enhanced model has seen increases of 5.8%, 5.2%, and 3.5% in mAP50, mAP50-90, and Recall respectively, thus better aligning with the demands of object detection tasks in automatic driving scenarios.

(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Yunxiang Liu, Xiaoya Ren, and Jianlin Zhu "Research on automatic driving scene object detection algorithm based on improved YOLOv8", Proc. SPIE 13439, Fourth International Conference on Testing Technology and Automation Engineering (TTAE 2024), 1343931 (12 December 2024); https://doi.org/10.1117/12.3055407

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available