Research On Metro Station Healthy Lighting Control Based On Passanger Action Recognition

With the rapid expansion of urban scale,higher requirements have been put forward for the urban transportation system.Rail transit,especially metro,has become the first choice for many urban residents because of its advantages of large capacity,fast speed and punctuality The lack of natural light brings new challenges to the energy saving and healthy lighting of subway station.Therefore,the research on lighting control of subway station has important practical significance and application value for reducing lighting energy consumption and improving light environment quality.In order to reduce the energy consumption of subway station lighting and improve the lighting quality,combined with the lighting design standard of urban rail transit,a subway station lighting control method based on action recognition is proposed.This method detects the action state of passengers at the station by camera,and adjusts the brightness and color temperature of LED lamps in real time to meet the visual needs of passengers under different actions,so as to achieve the purpose of green and healthy lighting.The main work of this paper is as follows(1)Research on passenger action recognition algorithm.Based on deep learning,a lightweight action recognition network algorithm is established.The algorithm divides the action recognition into two steps: human target detection and specific action recognition.The human body target detection is carried out by using yolov3 tiny lightweight target detection network;Combined with the characteristics of deep separable convolution and densenet network,a deep separable dense connection netwok(dsdnet)is constructed for specific action recognition.A large number of experiments show that the network algorithm can accurately and quickly identify human action.(2)Research on the lighting control scheme of subway station.Based on the analysis of the factors that affect the light environment of subway stations and the characteristics of control methods,according to the urban rail transit lighting design specifications and subway station passenger distribution rules,combined with the relevant requirements of healthy lighting,a dynamic control plan for subway station lighting is given.According to the distribution characteristics of waiting passengers in the platform,the sub-regional subway platform is gridded,and the camera is used to detect and recognize the action of waiting passengers in real time to realize the lighting control in each area.(3)Research on lighting control based on passenger action recognition.The action recognition algorithm is transplanted through the programmable machine vision module-openmv,and the real lighting simulation is carried out by identifying the subway station passengers’ action;In addition,combined with the results of passenger action recognition,with the help of Dialux Evo lighting design software,the energy consumption of a platform is modeled and analyzed.Compared with the traditional group control method,the energy consumption of subway station lighting control scheme based on passenger action recognition is reduced by 6.52% in 5 minutes.