基于改进YOLOv8的低光行人检测算法

计算机工程与科学 ›› 2026, Vol. 48 ›› Issue (3): 540-550.

基于改进YOLOv8的低光行人检测算法

徐广平,徐慧英,朱信忠，黄晓，王舒梦，宋杰

(1.浙江师范大学计算机科学与技术学院（人工智能学院）,浙江金华 321004；2.浙江师范大学教育学院，浙江金华 321004)

收稿日期:2024-03-01 修回日期:2024-08-05 出版日期:2026-03-25 发布日期:2026-03-25
基金资助:
国家自然科学基金（62376252）；浙江省自然科学基金(LZ22F030003)

An improved low-light pedestrian detection algorithm based on YOLOv8

XU Guangping,XU Huiying,ZHU Xinzhong,HUANG Xiao,WANG Shumeng,SONG Jie

(1.School of Computer Science and Technology(School of Artificial Intelligence),Zhejiang Normal University,Jinhua 321004；
2.College of Education,Zhejiang Normal University,Jinhua 321004,China)

Received:2024-03-01 Revised:2024-08-05 Online:2026-03-25 Published:2026-03-25

摘要/Abstract

摘要： 针对目前主流低光行人检测框架因为此任务中的图像亮度和对比度不足的原因导致性能下降的问题，提出了RetinaHA-YOLOv8算法。该算法通过采用RetinexFormer作为前置处理模块来恢复受损图像，确保后续算法能够从增强后的图像中提取到更加清晰和有用的特征；并利用HAT注意力机制在初始阶段保留关键信息并在特征融合后促进深度融合；最后为平衡额外计算负担并满足实时检测需求，引入在线重参数化卷积技术，以提高推理速度和每秒处理的帧数，同时保持检测精度。实验结果验证了RetinaHA-YOLOv8算法在公开低光行人检测数据集上的有效性，AP分别提升5.4%,11.7%和9.5%，且满足实际应用的实时性要求。

关键词: 低光行人检测, RetinexFormer框架, HAT注意力机制, 在线重参数化卷积

Abstract: In order to solve the problem that the current mainstream low-light pedestrian detection framework has poor performance due to insufficient image brightness and contrast in this task, this paper proposes the RetinaHA-YOLOv8 algorithm. The algorithm uses RetinexFormer as a pre-processing module to restore the damaged image, ensuring that the subsequent algorithm can extract clearer and more useful features from the enhanced image. Additionally, it uses the hybrid attention transformation (HAT) attention mechanism to retain key information in the initial stage and promote deep fusion after feature fusion. Finally, in order to balance the additional computational burden and meet the real-time detection requirements, the online re-parameterized convolution technology is introduced to improve the inference speed and frames per second while maintaining the detection accuracy. The experimental results verify the effectiveness of the RetinaHA-YOLOv8 algorithm on the public low-light pedestrian detection dataset, with AP increased by 5.4%, 11.7% and 9.5% respectively, while meeting the real-time requirements in practical applications.

Key words: low-light pedestrian detection, RetinexFormer framework, HAT attention mechanism, online reparameterized convolution

徐广平, 徐慧英, 朱信忠, 黄晓, 王舒梦, 宋杰. 基于改进YOLOv8的低光行人检测算法[J]. 计算机工程与科学, 2026, 48(3): 540-550.

XU Guangping, XU Huiying, ZHU Xinzhong, HUANG Xiao, WANG Shumeng, SONG Jie. An improved low-light pedestrian detection algorithm based on YOLOv8[J]. Computer Engineering & Science, 2026, 48(3): 540-550.