基于深度学习的生产车间智能管控研究

Research on Intelligent Management and Control of Production Workshop Based on Deep Learning

摘要: 为解决大型生产车间人工监控存在的易疲劳、漏检错检及成本高问题，以及满足智能化安全管理需求，论文开展深度学习在生产安全管理领域应用研究。以YOLOv8算法为核心构建智能管控系统，通过优化网络结构、采用自适应锚框与多尺度特征融合策略，精准检测场馆人数，结合分布式摄像头阵列实现人员密度超限预警；融合YOLOv8与目标检测算法，构建人体姿态关键点检测模型，利用轨迹分析识别人员跌倒行为，并通过构建专用数据集、数据增强等方法优化模型。实验表明，该模型精度与稳定性良好。系统实现人员流动和跌倒实时监测预警，为生产车间安全运营提供技术保障，推动管理数字化转型，未来具备优化拓展空间。

Abstract: To solve the problems of easy fatigue, missed or incorrect detection, and high cost existing in manual monitoring of large sports venues, and to meet the requirements of intelligent safety management, this paper conducts research on the application of deep learning in the field of public safety management. An intelligent control and management system is constructed with the YOLOv8 algorithm as the core. By optimizing the network structure, adopting adaptive anchor boxes and multi-scale feature fusion strategies, the number of people in the venue is accurately detected. Combined with a distributed camera array, the early warning of personnel density exceeding the limit is achieved. By integrating YOLOv8 with the object detection algorithm, a human pose key point detection model is constructed. Trajectory analysis is utilized to identify personnel fall behaviors, and the model is optimized through methods such as constructing dedicated datasets and data augmentation. Experiments show that the model has good accuracy and stability. This system enables real-time monitoring and early warning of personnel flow and falls, providing technical support for the safe operation of the venue, promoting the digital transformation of management, and has room for optimization and expansion in the future.