Research On Intelligent Street Light Monitoring System Based On Machine Vision Vehicle Flow Detectio

In recent years,under the research of many scholars at home and abroad,the lighting control method of street lamps has changed the clumsy and single traditional street lamp dimming method,and gradually moved towards the combination of environmental monitoring and lighting control methods.However,due to the lack of information fusion,the existing street lamp light control scheme still has some problems,such as single dimming level,low universality in the face of Complex Pavement environment,insufficient consideration of dimming factors and so on.To solve the above problems,this paper takes the intelligent street lamp monitoring system based on machine vision traffic flow detection as the research direction,and studies the specific implementation of the positioning system from four aspects:wireless sensor network scheme design,machine vision traffic flow detection algorithm,lighting control strategy design and system node design.According to the requirements of street lamp monitoring system in node distribution,information transmission distance and transmission efficiency,this paper first selects Zig Bee /IEEE802 15.4 protocol,and then the multi-channel banded clustering network topology is designed.On this basis,the data transmission mechanism and time synchronization algorithm are studied.The experimental results show that the designed network is more suitable for the strip distribution of street lamps in a certain area,reduces network conflicts and improves network throughput.For the intelligent street lamp monitoring system designed in this paper needs high real-time and high accuracy traffic flow detection,this paper proposes a traffic flow detection scheme based on machine vision.After the obtained video image frames are digitized,grayed and binarized,the moving target extraction method combining frame difference method and Surendra method is selected to count the traffic flow.This method greatly improves the detection effect of foreground moving targets.Aiming at the problems that the existing traffic flow analysis dimming strategy may lead to the jump of street lamp brightness dimming level,waste of energy and potential safety hazards,this paper establishes the LSTM traffic flow prediction model,and introduces the sample set design,parameter design and the establishment process of the prediction model in detail.On this basis,a dimming strategy combined with traffic flow prediction is proposed.The results show that the prediction accuracy of the traffic flow prediction model and the dimming level allocation accuracy of the final dimming strategy meet the requirements.In this paper,the hierarchical network nodes of the street lamp monitoring system are designed according to the performance indexes of the intelligent street lamp system,and a complete software and hardware platform is built,including acquisition nodes,cluster head nodes,base station nodes,cloud and host computer.Combined with the previous traffic flow detection algorithm design and lighting control strategy design,finally,a complete set of intelligent street lamp monitoring system based on machine vision traffic flow detection is formed.The test results show that the overall system designed in this paper is possible and practical,and can meet the needs of traffic flow statistical accuracy and dimming energy saving of intelligent street lamp system.This research result has certain reference value,provides a theoretical basis for the further improvement of smart street lamps,and is of great significance.