arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2311.00436cs.CV
收藏

通过RGB-事件融合增强可变光照下的交通目标检测

Enhancing Traffic Object Detection in Variable Illumination with RGB-Event Fusion

Waytous Co. Ltd.(万拓思有限公司);Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所)

Zhanwen Liu, Nan Yang, Yang Wang, Yuke Li, Xiangmo Zhao, Fei-Yue Wang

AI总结:

针对可变光照下传统帧相机交通目标检测信息丢失的问题,提出融合RGB与事件模态的SFNet,构建DSEC-Det数据集,实验证明其检测性能显著优于纯帧方法。

摘要:

可变光照下的交通目标检测颇具挑战性,因为传统帧采样相机的动态范围有限,会造成信息丢失。为解决这一问题,我们引入了受生物启发的事件相机,并提出一种新型结构感知融合网络(Structure-aware Fusion Network,SFNet),该网络从事件流中提取清晰完整的目标结构,通过跨模态融合补偿图像中的丢失信息,使网络能够获得适用于交通目标检测的光照鲁棒表征。具体而言,为缓解固定间隔事件采样方法中因交通目标运动状态多样而产生的稀疏或模糊问题,我们提出可靠结构生成网络(Reliable Structure Generation Network,RSGNet),用于生成速度不变帧(Speed Invariant Frames,SIF),确保目标结构的完整性和清晰度。随后,我们设计了新型自适应特征补全模块(Adaptive Feature Complement Module,AFCM),该模块通过感知图像的全局亮度分布,引导两种模态特征的自适应融合,以补偿图像中的信息丢失,从而生成光照鲁棒表征。最后,考虑到现有基于事件的目标检测数据集中缺乏大规模、高质量的标注,我们构建了DSEC-Det数据集,该数据集包含53个序列、63931张图像,以及覆盖8个类别的超过208000个标注。大量实验结果表明,我们提出的SFNet能够突破传统相机的感知边界,在mAP50指标上比基于帧的方法高出8.0%,在mAP50:95指标上高出5.9%。我们的代码和数据集将发布于https://github.com/YN-Yang/SFNet。

英文摘要:

Traffic object detection under variable illumination is challenging due to the information loss caused by the limited dynamic range of conventional frame-based cameras. To address this issue, we introduce bio-inspired event cameras and propose a novel Structure-aware Fusion Network (SFNet) that extracts sharp and complete object structures from the event stream to compensate for the lost information in images through cross-modality fusion, enabling the network to obtain illumination-robust representations for traffic object detection. Specifically, to mitigate the sparsity or blurriness issues arising from diverse motion states of traffic objects in fixed-interval event sampling methods, we propose the Reliable Structure Generation Network (RSGNet) to generate Speed Invariant Frames (SIF), ensuring the integrity and sharpness of object structures. Next, we design a novel Adaptive Feature Complement Module (AFCM) which guides the adaptive fusion of two modality features to compensate for the information loss in the images by perceiving the global lightness distribution of the images, thereby generating illumination-robust representations. Finally, considering the lack of large-scale and high-quality annotations in the existing event-based object detection datasets, we build a DSEC-Det dataset, which consists of 53 sequences with 63,931 images and more than 208,000 labels for 8 classes. Extensive experimental results demonstrate that our proposed SFNet can overcome the perceptual boundaries of conventional cameras and outperform the frame-based method by 8.0% in mAP50 and 5.9% in mAP50:95. Our code and dataset will be available at https://github.com/YN-Yang/SFNet.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑