Research on target detection algorithms in complex road conditions

Jiabao Zhang; Xihan Zhang; Haolun Sheng; Jun Liu

doi:10.1117/12.3025558

28 February 2024 Research on target detection algorithms in complex road conditions

Jiabao Zhang, Xihan Zhang, Haolun Sheng, Jun Liu

Proceedings Volume 13071, International Conference on Mechatronic Engineering and Artificial Intelligence (MEAI 2023); 130712Z (2024) https://doi.org/10.1117/12.3025558
Event: International Conference on Mechatronic Engineering and Artificial Intelligence (MEAI 2023), 2023, Shenyang, China

Abstract

In complex road environments, traditional algorithms for detecting objects often face issues such as misclassification and omission of distant targets and small objects like pedestrians. To overcome these challenges, a new object detection algorithm called SCK-YOLO has been introduced in this paper. This algorithm is an improvement over the model of YOLOv5s network and includes a tiny object Detection Laye. It replaces the computationally intensive C3 module with a more lightweight C2f module and uses the K-means algorithm for anchor box selection, replacing the original anchor box selection method. Experimental comparisons show that the proposed algorithm performs better on the Kitti dataset compared to the original YOLOv5s algorithm, with an increase of 1.59% in mAP@0.5 and a 3.99 % enhancement in mAP@0.5:0.95.

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

Citation Download Citation

Jiabao Zhang, Xihan Zhang, Haolun Sheng, and Jun Liu "Research on target detection algorithms in complex road conditions", Proc. SPIE 13071, International Conference on Mechatronic Engineering and Artificial Intelligence (MEAI 2023), 130712Z (28 February 2024); https://doi.org/10.1117/12.3025558

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