动态客流智能感知与场景化服务推演系统数据集
收藏资源简介:
本数据集面向大型文化遗产场馆的动态客流监测与场景化服务推演研究及智慧场馆建设需求,以故宫围栏为空间边界,结合高德多源时空数据与故宫高精度数字孪生模型,构建分钟级客流感知、热力可视化与推演决策的数据基础。地理坐标相关数据主要来源于高德开放平台及其多模态时空大模型输出接口,在合法授权与隐私合规前提下融合移动终端定位、地图搜索与导航行为、围栏感知等信息生成,并经脱敏、清洗、坐标变换、尺度归一与时序同步后映射至 UE5 等三维引擎坐标体系。数据集内容覆盖入院小时客流、馆内热力分布、重点展馆小时监测、观众来源分析及珍宝馆时点分布预测等模块,同时提供项目开发文件、美术资源与UI界面、代码、软件本体及相关文档与测试数据。采集时间范围为2024—2025年,采集与加工主要在联通数字科技有限公司完成;数据量18GB,发布数据格式为exe、fbx、tif、png、json、jar。
This dataset is tailored to the requirements of dynamic passenger flow monitoring, scenario-based service deduction research, and smart venue construction for large-scale cultural heritage venues. Taking the boundary defined by the fences of the Palace Museum as the spatial scope, it integrates Amap's multi-source spatiotemporal data and the high-precision digital twin model of the Palace Museum to establish a foundational dataset for minute-level passenger flow perception, thermal visualization, and deduction-based decision-making. Geospatial coordinate data is primarily sourced from the Amap Open Platform and its multi-modal spatiotemporal large model output APIs. It is generated by integrating information such as mobile terminal positioning, map search and navigation behavior, and fence perception under the premise of legal authorization and privacy compliance. After undergoing desensitization, data cleaning, coordinate transformation, scale normalization, and time synchronization, the data is mapped to the coordinate systems of 3D engines including UE5. The dataset covers multiple modules including hourly incoming passenger flow, in-venue thermal distribution, hourly monitoring of key exhibition halls, audience source analysis, and time-point distribution prediction for the Palace Museum's Treasure Hall. It also provides project development documents, art resources, UI interfaces, source codes, standalone software, as well as relevant documents and test data. The data collection period spans from 2024 to 2025, and the data collection and processing work was primarily completed by China Unicom Digital Technology Co., Ltd. The total size of the dataset is 18GB, and the released data formats include exe, fbx, tif, png, json, and jar.




