Research On Bus Passenger Flow Statistics System Based On Internet Of Things

With the continuous improvement of people’s living standard,the popularity of private cars has facilitated the residents to go out.However,the rapid increase of private cars also brings huge burden to the urban transportation system and leads to a series of problems,such as traffic jams,frequent traffic accidents and automobile exhaust pollution.As an important part of urban public transportation,bus is an important way to solve urban traffic congestion.Its operation directly affects the operation efficiency of social and economic life and people’s quality of life.The current route planning and bus scheduling of the bus group are based on passenger flow information.Currently,there are various problems in passenger flow statistics,such as the inability to count the number of passengers getting off the bus,or the inability to deal with the complex environment of passengers getting on and off the bus,which leads to thein ability to make reasonable planning and bus scheduling.Therefore,the design and implementation of bus passenger flow statistic system based on Internet of Things communication technology has important theoretical significance and practical value.This paper design the bus passenger flow statistics system by using the Internet of Things technology to construct the sensing layer,network layer and application layer.The system realizes the remote data transmission by means of the network connection function provided by the global mobile communication system,and realizes the remote system of passenger statistics by means of data processing and image processing technology of embedded terminal equipment,which solves the problemof high cost,inconvenient statistics and single statistics.The bus passenger flow statistics system involves hardware and software.The hardware part completes thefunctions of datacollection,data processing and datat ransmission.The software part complete and multi-passenger tracking and obtains passenger flow information.This thesis completes the design and implementation of the software and hardware modules of the bus passenger flow statistics system,which are summarized as follows:(1)Overall scheme design of the system.Investigate the existing passenger flow statistical scheme and analyze the defects.Then design the bus passenger flow statistics system based on the Internetof Things to transmit passenger flow information,embedded development board to process data,convolutional neural network to detect passengers,and correlation analysis method to track multiple passengers.(2)Hardware design.Data acquisition equipment,data processing equipment and data transmission equipment are designed.The design can deal with complex bus conditions,provide clear images for the following bus passenger detection algorithm,and provide sufficient computing resources for real-time computation of the convolutional neural network.(3)Software design.It is difficult to run on the embedded development board because the operation of convolutional neural network requires huge computing resources.Based on the current convolutional neural network technology and optimization technology,this paper designs a real-time convolutional neural network structure for bus passenger detection.According to the characteristics of small passenger target in the image sequence,the correlation analysis method is adopted to track the passenger detected by the convolutional neural network structure with multiple targets.Then according to the characteristics of the bus,the passenger flow statistics.After hardware installation,debugging and software and tuning,the system was tested on video segment in real environment,and achieved 93% statistical accuracy.