基于改进P2PNet的铁路车站人群数量估计算法

doi:10.3969/j.issn.1006-2475.2025.11.007

摘要/Abstract

摘要： 摘要：本文针对铁路车站等场景下人群数量估计问题，提出一种基于改进P2PNet的铁路车站人群数量估计算法。该算法对传统的P2PNet算法进行较大的改动和优化。首先，算法模型采用BiFPN的路径聚合来增强网络对不同尺度特征图的融合能力，解决图像中人员大小尺度相差较大的情况；其次，网络在低层特征图引入A-SPP结构以提高网络的感受野范围，增强网络提取多尺度特征的能力；再次，网络中的OutLayer之前采用了CSAM注意力动态地调整特征图中各个通道和空间位置的重要性，更有效地回归图像中人的位置和分类；最后，在损失函数中，Focal Loss代替了传统的交叉熵，来解决正负样本不均衡和难易样本不均衡的问题。在公开数据集和自有数据集上进行对比实验的结果表明，该算法在平均绝对误差上优于当前同类先进算法。在实际的车站视频场景下，该算法能够准确估计人群数量。

关键词: 关键词：人群数量估计, P2PNet, 多尺度特征, 注意力机制, 特征融合

Abstract: Abstract: An improved P2PNet based algorithm for estimating the number of people in railway stations is proposed. The algorithm has made significant modifications and optimizations to the traditional P2PNet algorithm. Firstly， the algorithm model adopts BiFPN path aggregation to enhance the network’s ability to fuse feature maps of different scales， solving the problem of large differences in personnel size and scale in images. Secondly， the network introduces the A-SPP structure into the low-level feature maps to increase the receptive field range and enhance its ability to extract multi-scale features. Thirdly， the CSAM attention mechanism is adopted before the OutLayer in the network to dynamically adjust the importance of each channel and spatial position in the feature map， which more effectively regresses the position and classification of people in the image. Finally， Focal Loss replaces traditional cross entropy in loss function to solve the problems of imbalanced positive and negative samples and imbalanced difficult and easy samples. The results of comparative experiments on publicly datasets and proprietary datasets show that this algorithm outperforms current advanced algorithms in terms of mean absolute error. In actual station video scenarios， this algorithm can accurately estimate the number of people.

Key words: Key words: , crowd counting estimation； P2PNet； multi-scale features； attention mechanism； feature fusion

中图分类号:

中图分类号：TP391

万成凯1, 安高云2, 崔岚3. 基于改进P2PNet的铁路车站人群数量估计算法[J]. 计算机与现代化, 2025, 0(11): 58-64.

WAN Chengkai1, AN Gaoyun2, CUI Lan3. Crowd Counting Estimation Algorithm of Railway Stations Based on Improved P2PNet[J]. Computer and Modernization, 2025, 0(11): 58-64.

参考文献

［1］余鹰，朱慧琳，钱进，等 . 基于深度学习的人群计数研究综述［J］. 计算机研究与发展， 2021，58（12）：2724-2747.
［2］ KHAN M A， MENOUAR H，HAMILA R. Revisiting crowd counting：State-of-the-art，trends，and future perspectives［J］. Image and Vision Computing，2023，129. DOI: 10.1016/j.imavis.2022.104597.
［3］ ZHANG Y Y， ZHOU D S， CHEN S Q， et al． Single-image crowd counting via multi-column convolutional neural network［C］// 2016 IEEE Conference on Computer Vision and Pattern Recognition. IEEE， 2016 : 589-597.
［4］ LIU L B， QIU Z L， LI G B， et al. Crowd counting with deep structured scale integration network［C］// 2019 IEEE/CVF International Conference on Computer Vision （ICCV）.IEEE， 2019:1774-1783.
［5］ SONG Q Y， WANG C G， JIANG Z K，et al. Rethinking counting and localization in crowds: A purely point-based framework［J］. arXiv preprint arXiv:2107.12746， 2021.
［6］张月坤. 北京地铁客流密度自动检测技术研究［J］. 中国铁路， 2017（4）:96-100.
［7］张亚伟，吕晓军，李明，等. 智能视频监控系统在铁路旅客服务系统的研究与应用［J］. 中国铁路， 2020（1）:109-113.
［8］李瑞，李平，代明睿，等. 基于铁路客站视频多视角特征图投影融合的人群密度估计模型［J］.中国铁道科学，2022，43（3）:182-192.
［9］ SIMONYAN K， ZISSERMAN A. Very deep convolutional networks for large-scale image recognition［J］. arXiv preprint arXiv:1409.1556， 2014.
［10］ TAN M X， PANG R M， LE Q V. EfficientDet: Scalable and efficient object detection［C］// 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition（CVPR）. IEEE， 2020:10778-10787.
［11］石瑞姣，陈后金，李居朋，等. 基于注意力和多级特征融合的铁路场景小尺度行人检测算法［J］. 铁道学报，2022，44（5）:76-83.
［12］ LIN T Y， DOLLÁR P， GIRSHICK R， et al. Feature pyramid networks for object detection［C］// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. IEEE， 2017:936-944.
［13］ LIU S， QI L， QIN H F， et al. Path aggregation network for instance segmentation［C］// Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition. IEEE， 2018:8759-8768.
［14］ HE K M， ZHANG X Y， REN S Q， et al. Spatial pyramid pooling in deep convolutional networks for visual recognition［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2015，37（9）:1904-1916.
［15］ CHEN L C， PAPANDREOU G， KOKKINOS L，et al. DeepLab: Semantic image segmentation with deep convolutional nets， atrous convolution， and fully connected CRFs［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2018，40（4）：834-848.
［16］ KUHN H W. The Hungarian method for the assignment problem［J］. Naval Research Logistics Quarterly， 1955，2:83-97.
［17］ LIN T Y，GOYAL P，GIRSHICK R，et al. Focal loss for dense object detection［C］// Proceedings of the IEEE International Conference on Computer Vision. IEEE，2017：2999-3007.
［18］ IDREES H， TAYYAB M， ATHREY K， et al. Composition loss for counting， density map estimation and localization in dense crowds［C］// European Conference on Computer Vision. Springer， 2018:544-559.
［19］ WANG Q， GAO J Y， LIN W， et al. NWPU-Crowd: A large-scale benchmark for crowd counting and localization［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2021，43（6）:2141-2149.
［20］ MA Z H，WEI X， HONG X P， et al. Bayesian loss for crowd count estimation with point supervision［C］// 2019 IEEE International Conference on Computer Vision. IEEE， 2019. DOI: 10.1109/ICCV.2019.00624.
［21］ XIONG H P， LU H， LIU C X， et al. From open set to closed set: Counting objects by spatial divide-and-conquer［C］// 2019 IEEE International Conference on Computer Vision. IEEE， 2019. DOI: 10.1109/ICCV.2019.00845.
［22］ WANG B Y， LIU H D， SAMARAS D， et al. Distribution matching for crowd counting［C］// Proceedings of the 34th International Conference on Neural Information Processing Systems. ACM， 2020:1595-1607.
［23］ YI J，SHEN Z L，CHEN F，et al. A lightweight multiscale feature fusion network for remote sensing object counting［J］. IEEE Transactions on Geoscience and Remote Sensing， 2023，61. DOI: 10.1109/TGRS.2023.3238185 .
［24］ LIANG L J， ZHAO H L， ZHOU F B， et al. SC2Net：Scale-aware crowd counting network with pyramid dilated convolution［J］. Applied Intelligence， 2023，53：5146-5159.
［25］ WANG M J， CAI H， DAI Y， et al. Dynamic mixture of counter network for location-agnostic crowd counting［C］// Proceedings of the 2023 IEEE/CVF Winter Conference on Applications of Computer Vision. IEEE， 2023. DOI: 10.1109/WACV56688.2023.00025.
［26］ CHEN Q， LEI T， GENG X Z， et al. DLMP-Net：A dynamic yet lightweight multi-pyramid network for crowd density estimation ［C］// Proceedings of the 5th Chinese Conference on Pattern Recognition and Computer Vision. Springer，2022：27-39.
［27］ LIANG D K， CHEN X W， XU W， et al. TransCrowd： Weakly supervised crowd counting with transformers［J］. Science China Information Sciences，2022，65. DOI: 10.1007/ s11432-021-3445-y.

[1]	祝文姬1, 班卫华2, 邹林3, 刘旭3. 基于LightGlue网络改进的输电线路巡检无人机V-SLAM方法[J]. 计算机与现代化, 2025, 0(11): 41-48.
[2]	孙二杰1, 张奇丰2, 王德清3. 基于轻量级YOLOv8的变电站设备缺陷检测[J]. 计算机与现代化, 2025, 0(11): 65-70.
[3]	杨安博1, 钟国韵1, 刘梅锋2, 习超2, 张维1, 丁鹏1. LSGI-YOLOv8:基于轻量化YOLOv8的瓷砖表面缺陷检测算法[J]. 计算机与现代化, 2025, 0(11): 80-88.
[4]	师红宇, 张哲于, 杜文, 李怡. 融合空间信息的YOLOv7交通标志检测[J]. 计算机与现代化, 2025, 0(10): 7-13.
[5]	蒋冬梅1, 杨诺2, 陈仁明2, 董昌武3, 彭成东2, 4. 基于变核卷积的HRNetV2模型对舌下络脉图像分割算法[J]. 计算机与现代化, 2025, 0(10): 14-19.
[6]	周邦湲1, 辛国江1, 梁昊2, 丁长松1. 基于两阶段的多重注意力机制网络的胰腺分割[J]. 计算机与现代化, 2025, 0(10): 67-72.
[7]	刘嵘澂1, 辛国江1, 张杨1, 朱磊2. 基于改进的DeepLabv3+模型的自然环境下舌象分割方法[J]. 计算机与现代化, 2025, 0(10): 32-36.
[8]	郭金豪, 王峰萍, 王浩琦. 面向密集场景的PB-YOLOv7行人检测方法[J]. 计算机与现代化, 2025, 0(09): 14-19.
[9]	罗德艳, 徐杨, 左锋云, 张永丹. 改进YOLOv8的轻量化垃圾分类检测[J]. 计算机与现代化, 2025, 0(09): 35-42.
[10]	林睿姿, 姚达, 戴欣, 沈国誉, 王嘉慧, 万伟国. 基于CycleGAN和注意力机制的人脸素描图像转换[J]. 计算机与现代化, 2025, 0(09): 61-66.
[11]	刘成, 冯广. 基于双重注意力机制的CNN-BiLSTM和LightGBM股票预测[J]. 计算机与现代化, 2025, 0(09): 97-103.
[12]	景清武, 陈洪军, 高翟, 周美美. 目标驱动的面向推荐的对话生成方法[J]. 计算机与现代化, 2025, 0(08): 16-23.
[13]	苏衍森, 牟莉. 改进YOLOv8的课堂行为检测算法[J]. 计算机与现代化, 2025, 0(08): 76-81.
[14]	高莉莎1, 彭国政2, 王翀1, 韩硕1, 项楠1, 张启哲2, 朱跃飞3, 程旭3. 基于YOLOv5的双主干网络绝缘子缺陷检测算法[J]. 计算机与现代化, 2025, 0(08): 82-88.
[15]	蔡博涵, 刘俊. 基于优化Transformer的长短期空气污染物浓度预测[J]. 计算机与现代化, 2025, 0(07): 55-62.