朱强,孙晨,徐潘宇驰,闫云凤
ZHU Qiang,SUN Chen,XU Panyuchi,YAN Yunfeng

• 1:广东电网有限责任公司广州供电局
• 2:浙江大学电气工程学院

摘要(Abstract)：

对于智慧工地作业人员异常行为的检测，现有的经典目标检测算法无法到达理想的检测效果，准确率较低。为此，提出一种基于FCOS（全卷积单阶段目标检测）的二阶段检测算法来实现智慧工地异常行为检测。该算法主要包括两个级联网络，首先通过FCOS对作业人员及异常行为标志物进行识别定位，再使用MLP（多层感知器）完成异常行为的检测分类。最后以相关项目现场采集的12 977张样本图片作为数据集，对检测算法进行实验验证。结果表明，该算法在对各类异常行为的检测中均表现优异，而且检测实时性好、计算复杂度低、模型参数少，在实际项目的部署及应用方面具有明显优势。
The existing classical object detection algorithm cannot satisfactorily detect abnormal behavior of operators at smart construction sites for its low accuracy. To this end, a two-stage detection algorithm based on the FCOS(full convolutional single-stage target detection) is proposed to detect abnormal behavior at smart construction sites.The algorithm, mainly consisting of two cascaded networks, first identifies and locates operators and abnormal behavior markers by the FCOS, and then uses the MLP(multilayer perceptron) to detect and classify abnormal behavior. Finally, the 12,977 sample images collected from the relevant project sites are used as the data set to experimentally validate the detection algorithm. The results show that the algorithm performs well in the detection of all kinds of abnormal behavior, and it has a clear advantage in the deployment and application of practical projects due to its excellent real-time detection, less complex computation, and fewer model parameters.

关键词(KeyWords)： 智慧工地;异常行为检测;FCOS;多层感知器
intelligent construction site;abnormal behavior detection;FCOS;MLP

Abstract:

Keywords:

基金项目(Foundation): 国家自然科学基金项目（U1909201）;; 浙江省自然科学基金探索项目（Q21F030038）;; 浙江省科技“尖兵计划”项目（2022C01056）

作者(Author): 朱强,孙晨,徐潘宇驰,闫云凤
ZHU Qiang,SUN Chen,XU Panyuchi,YAN Yunfeng

DOI: 10.19585/j.zjdl.202304008

参考文献(References)：

• [1]崔志诚，马胜.基于物联网技术的智慧工地[J].电子技术应用，2021,47(2):33-35.CUI Zhicheng,MA Sheng. Intelligent construction site based on Internet of Things technology[J].Application of Electronic Technique,2021,47(2):33-35.
• [2]林其雄，陈畅，闫云凤，等.一种基于特征引导的电力施工场景工装合规穿戴二阶段检测算法[J].浙江电力，2022,41(4):44-50.LIN Qixiong,CHEN Chang,YAN Yunfeng,et al.A twostage detection algorithm for workwear compliance in power construction scenarios based on feature guidance[J].Zhejiang Electric Power,2022,41(4):44-50.
• [3] FANG Q,LI H,LUO X.Detecting non-hardhat-use by a deep learning method from far-field surveillance videos[J].Automation in Construction,2018,85:1-9.
• [4]任丹彤，何赟泽，刘贤金，等.面向智慧工厂的双光融合车间人员行为识别方法[J].测控技术，2022,41(8):9-15.REN Dantong,HE Yunze,LIU Xianjin,et al. Personnel behavior recognition method of dual light fusion for smart factory[J].Measurement&Control Technology,2022,41(8):9-15.
• [5]马莉，王卓，代新冠，等.基于双流CNN与Bi-LSTM的施工人员不安全行为轻量级识别模型[J].西安科技大学学报，2022,42(4):809-817.MA Li,WANG Zhuo,DAI Xinguan,et al.Lightweight unsafe behavior recognition model of construction workers based on two-stream CNN and Bi-LSTM[J]. Journal of Xi’an University of Science and Technology,2022,42(4):809-817.
• [6]谢觉，唐俊.基于视频流和位置流混合的建筑施工人员行为识别研究[J].电子世界，2019(16):49-50.XIE Jue,TANG Jun. Research on behavior identification of construction workers based on video stream and location stream[J].Electronics World,2019(16):49-50.
• [7]仝泽友，丁恩杰.矿井皮带区矿工违规行为识别方法[J].河南科技大学学报（自然科学版），2020,41(2):40-46.TONG Zeyou,DING Enjie. Identification method of miner violation behavior in mine belt area[J]. Journal of Henan University of Science and Technology(Natural Science),2020,41(2):40-46.
• [8]胡志坤，赵超越，王振东，等.基于边缘计算和无人机巡检图像的输电杆塔关键部位隐患智能识别[J].浙江电力，2020,39(10):21-27.HU Zhikun,ZHAO Chaoyue,WANG Zhendong,et al.Intelligent identification of hidden troubles in key parts of transmission towers based on edge computing and UAV inspection images[J]. Zhejiang Electric Power,2020,39(10):21-27.
• [9] TIAN Z,SHEN C H,CHEN H,et al.FCOS:fully convolutional one-stage object detection[C]//2019 IEEE/CVF International Conference on Computer Vision(ICCV).Seoul,Korea(South):IEEE,2020:9626-9635.
• [10]王卓，王玉静，王庆岩，等.基于协同深度学习的二阶段绝缘子故障检测方法[J].电工技术学报，2021,36(17):3594-3604.WANG Zhuo,WANG Yujing,WANG Qingyan,et al.Two stage insulator fault detection method based on collaborative deep learning[J].Transactions of China Electrotechnical Society,2021,36(17):3594-3604.
• [11]于敏，屈丹，司念文.改进的RetinaNet目标检测算法[J].计算机工程，2022,48(8):249-257.YU Min,QU Dan,SI Nianwen.Improved RetinaNet algorithm for object detection[J]. Computer Engineering,2022,48(8):249-257.
• [12]李晓艳，符惠桐，牛文涛，等.基于深度学习的多模态行人检测算法[J].西安交通大学学报，2022,56(10):61-70.LI Xiaoyan,FU Huitong,NIU Wentao,et al.Multi-modal pedestrian detection algorithm based on deep learning[J].Journal of Xi’an Jiaotong University,2022,56(10):61-70.
• [13]史晨晨，张长伦，何强，等.基于改进特征金字塔的目标检测[J].电子测量技术，2021,44(20):150-156.SHI Chenchen,ZHANG Changlun,HE Qiang,et al.Object detection based on improved feature pyramid[J].Electronic Measurement Technology,2021,44(20):150-156.
• [14]齐鹏宇，王洪元，张继，等.基于改进FCOS的拥挤行人检测算法[J].智能系统学报，2021,16(4):811-818.QI Pengyu,WANG Hongyuan,ZHANG Ji,et al.Crowded pedestrian detection algorithm based on improved FCOS[J].CAAI Transactions on Intelligent Systems,2021,16(4):811-818.
• [15] HAN G,DU H,LIU J X,et al.Fully conventional anchorfree Siamese networks for object tracking[J]. IEEE Access,2019,7:123934-123943.
• [16]赵志新，赵宗罗，赵颖，等.基于并行化BP神经网络的配电变压器故障快速诊断方法[J].浙江电力，2021,40(12):82-88.ZHAO Zhixin,ZHAO Zongluo,ZHAO Ying,et al.A fast fault diagnosis method of distribution transformer based on parallel BP neural network[J]. Zhejiang Electric Power,2021,40(12):82-88.
• [17]刘翠翠.基于综合特征和多层感知器的图像分类[J].电子测量技术，2019,42(8):74-77.LIU Cuicui. Image classification based on integrated features and multilayer perceptron[J]. Electronic Measurement Technology,2019,42(8):74-77.