Research And Application Of Deformation Corrugated Guardrail Detection For Highway Under Complex Background

The corrugated guardrail of highway is an important part of the transport infrastructure of the highway,which has the functions of guiding the line of sight of drivers,preventing vehicles from rushing out of the road,protecting vehicles and drivers.The structural integrity of the guardrail is an important basis for its functions.However,in highway scenarios,corrugated guardrails often undergo local bending due to factors such as vehicle collisions,natural weather,and human destruction,resulting in deformation losses and weakened protective functions,thus posing serious safety hazards.To ensure the integrity of the guardrail structure,the highspeed maintenance department needs to monitor the guardrail and timely detect deformation of the guardrail.The traditional monitoring of corrugated guardrails on highways is mainly completed through on-site investigations by staff,which has drawbacks such as high cost,long operation time,and susceptibility to traffic accidents,gradually making it unable to adapt to the rapidly increasing mileage and maintenance needs of highways.To further promote the development of intelligent road inspection waveform technology in China,improve the accuracy and efficiency of guardrail detection,and save road management costs.This article conducts relevant research on intelligent detection of deformation corrugated guardrails on highways,which can be mainly divided into the following aspects:1)Using self-developed road inspection equipment in the laboratory to collect guardrail images,and constructing a corrugated guardrail dataset with rich samples and balanced number of various guardrails through manual screening and data expansion methods.Based on the characteristics of diverse backgrounds and similar features in corrugated guardrail detection,and combined with practical detection needs,YOLOv8 instance segmentation model is selected for corrugated guardrail detection.Finally,transfer learning is carried out on the data set to obtain the highway corrugated guardrail instance segmentation model;2)Propose a method for delineating the detection area of corrugated guardrails based on binocular stereo vision ranging technology,in order to avoid duplicate detection caused by the appearance of the same deformed guardrail in multiple continuous images;3)Use the edge fitting method based on the least squares method to fit and analyze the edge contour features of the guardrail,calculate the relevant indicators of the fitting line,use the threshold based method to classify the guardrail samples,and use the ResNet50 image classification model to remove interference from the guardrail connection points,ultimately achieving accurate detection of deformed corrugated guardrails;4)Based on Python language and PyQt toolkit,the proposed algorithm for detecting deformation corrugated guardrails on highways was systematically developed on the Windows platform,and engineering application tests were conducted to verify the accuracy and robustness of the algorithm proposed in this paper.This article randomly selected three sections of the G15 Shenhai Highway for application testing,with corrugated guardrail detection precision and recall rates of 0.843 and 0.906,respectively.14 out of 16 deformation corrugated guardrail samples were successfully detected,showing high detection accuracy.The experimental results show that the deformation corrugated guardrail detection algorithm proposed in this article has high accuracy and feasibility,can assist manual completion of daily inspection work,reduce inspection difficulty,improve the work efficiency of road maintenance departments,and is of great significance for promoting the intelligent and information-based development of road maintenance management in Fujian Province.