The invention is suitable for the technical field of elevator management, and particularly relates to an elevator people flow prediction and elevator scheduling system based on deep learning, and the method comprises the steps: collecting elevator monitoring data; content identification is conducted on the elevator monitoring data, the personnel flow condition of each floor is determined, and basic people flow data is generated; a deep learning neural network is constructed based on the basic people flow data, a real-time monitoring video is collected, and elevator dispatching criteria are extracted from the real-time monitoring video; and elevator dispatching criteria are analyzed through the trained deep learning neural network, and real-time dispatching is conducted on the elevator based on the analysis result. According to the elevator people flow prediction and elevator scheduling system based on deep learning, the historical operation state of the elevator is monitored, the demand conditions of all floors for the elevator in different time periods are analyzed, and therefore when the elevator is idle, position scheduling is actively conducted, and the elevator scheduling efficiency is improved. The elevator can correspond to the elevator operation requirements of users to the maximum extent, and the intelligent degree is improved.
本发明适用于电梯管理技术领域,尤其涉及一种基于深度学习的电梯人流预测和电梯调度系统,所述方法包括:采集电梯监控数据;对电梯监控数据进行内容识别,确定各个楼层的人员流动情况,生成基础人流数据;基于基础人流数据构建深度学习神经网络,采集实时监控视频,从中提取电梯调度判据;通过训练后的深度学习神经网络分析电梯调度判据,基于分析结果对电梯进行实时调度。本发明实施例提供的一种基于深度学习的电梯人流预测和电梯调度系统,通过对电梯的历史运行状态进行监测,分析各个楼层在不同时间段内对电梯的需求情况,从而在电梯闲置时,主动进行位置调度,使得电梯能够最大程度实现对用户电梯运行需求的相应,提高了智能化程度。
Elevator people flow prediction and elevator dispatching system based on deep learning
一种基于深度学习的电梯人流预测和电梯调度系统
2023-08-15
Patent
Electronic Resource
Chinese
IPC: | B66B Aufzüge , ELEVATORS |
Elevator dispatching method, intelligent robot, elevator and dispatching server
European Patent Office | 2020
|Elevator dispatching method, elevator dispatching device and electronic equipment
European Patent Office | 2023
|