基于双分支特征聚合网络的车辆检测算法  

Vehicle Detection Algorithm Based on Dual Branch Feature Aggregation Network

作　　者：吕蒙[1] 毛盛辉 柴亮 高鹏飞 时蕾[1] LYU Meng;MAO Shenghui;CHAI Liang;GAO Pengfei;SHI Lei(Zhengzhou Railway Vocational and Technical College,Zhengzhou 451400,China;Zhengzhou Metro Group Co.,Ltd.,Zhengzhou 450047,China)

机构地区：[1]郑州铁路职业技术学院,郑州451400 [2]郑州地铁集团有限公司,郑州450047

出　　处：《计算机工程与应用》2024年第22期240-250,共11页Computer Engineering and Applications

基　　金：国家自然科学基金(51905071);河南省高等学校重点科研项目(22B460031);河南省科技攻关项目(232102220059)。

摘　　要：车辆目标检测是自动驾驶的重要环节,现有的车辆目标检测算法在特征提取方面没有充分考虑卷积神经网络(convolutional neural network,CNN)和Transformer各自的优缺点,一定程度上限制了网络的整体性能。提出了一种由CNN和Transformer组成的双分支特征聚合网络。在编码阶段,基于CNN和Transformer各自的优势,构建了双分支主干网络来提取原始图像的特征信息;通过设计的多级别空间注意力模块和双支路特征聚合模块,使两个分支间的特征信息相互引导学习;通过构建的双分支注意力模块来进一步减少深层神经网络中特征信息的丢失。在实验部分通过消融实验和对比实验进一步验证了所提算法的有效性,其相比主流的目标检测算法,在mAP(mean average precision)指标上提升了约3.5%。Vehicle target detection is an important part of autonomous driving.Existing vehicle target detection algo-rithms have not fully considered the advantages and disadvantages of CNN(convolutional neural network)and Transformer in feature extraction,which to some extent limits the overall performance of the network.This paper proposes a dual branch feature aggregation network consisting of CNN and Transformer.In the encoding stage,based on the respective advan-tages of CNN and Transformer,a dual branch backbone network is constructed to extract the feature information of the original image.By designing a multi-level spatial attention module and a dual branch feature aggregation module,the fea-ture information between the two branches is guided to learn from each other.Finally,a dual branch attention module is constructed to further reduce the loss of feature information in deep neural networks.In the experimental section,the effec-tiveness of the proposed algorithm is further verified through ablation experiments and comparative experiments.Com-pared to mainstream object detection algorithms,it has improved by about 3.5%in the mAP(mean average precision)metric.

关 键 词：车辆目标检测 卷积神经网络(CNN) TRANSFORMER 双分支 引导学习 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]