Research On Intelligent Detection System Of Tunnel Equipment Operation Status

It is the basic guarantee for the operation of tunnels that the normal operation of lighting and information boards.The requirements for brightness are quite strict in tunnels.The lack of brightness will affect the safe driving in tunnels.The information board informs the drivers of weather,road sections and other information,so the safe driving in tunnel needs to ensure the normal operation of the lights and information boards and the operation status of the lights and the information boards should be detected.This article designs the whole tunnel equipment operation status intelligent detection system,mainly includes the system function analysis,system composition and design principles.The detection of the operating status of the lights is mainly through the video transmitted to the monitoring center,the B component of the RGB model is used as the graying result,after the 3*3 median filter,the exponential function is used to perform the gray scale transformation to remove the interference in the tunnel and increase the contrast between the lighting and the background of the tunnel,using the largest cluster-like variance method for binarization,the outline of the light is extracted from the tunnel image,using K-means clustering algorithm separates the left and right lights according to the moment of inertia direction of lights and the direction of the largest Feret diameter and some other parameters,because the lights are evenly distributed on the same road section,the ratio of the area of the light's contour can be used to determine if there is a fault,so that when there are lights in the fault state can be found in time to report to the monitoring center to take measures.The operation status of the information board is processing tunnel image contains information board,graying through the luminance component use the HSL model,3*3 median filter,exponential function to perform the gray level change,binarization method of the maximum between-class variance method,closed operation,filling holes and convex hull operations,deleting the boundary target,deleting the small particles and extracting the outline of the information board,dividing the characters in the information board by the horizontal and vertical dual projection algorithm,and then separating them into a single character and then pass through the layered LBP+HOG feature fusion algorithm after the segmentation,characters are extracted and compared with the features of the characters that should be displayed,when the fonts are incomplete or incorrect,they can be found in time,and then the monitoring center is notified to deal with them in time to avoid misleading information from the driver and affect the normal driving of the tunnel.In this paper,the lighting detection subsystem of the tunnel equipment operation status detection system can identify more than 25 lights,and the detection distance can reach more than 125 meters,which can meet the needs of the project.After the intelligence board is detected,the accuracy of character recognition of hierarchical LBP+HOG algorithm can reach 98.8%.Therefore,the methods of the tunnel equipment detection system proposed in this paper is feasible and effective.