Research On A Method For Detecting Structural Defects And Longitudinal Spatial Positioning Of Drainage Pipe

Sewers run like blood vessels in the human body,intricately distributed throughout the city,forming urban meridians whose health is closely linked to the state of life of the inhabitants.Regularly inspecting the city’s sewers and identifying potential threats is one of the most important measures to keep the city running well.The main components of the sewer’s vulnerability are the various defects located in the main part of the pipe section,and the identification of the type and distribution of defects is a priori condition for the maintenance and restoration of the pipeline section.With the iterative improvement of image processing technology,the current sewer defect detection technology tends to be intelligent,but there is no effective solution for defect spatial location.The traditional way of relying on manual experience to estimate the location of defects leads to low reliability and large errors in the intelligent assessment of the health condition of the sewer.Thus,achieving intelligent defect spatial location is an inevitable part of improving the efficiency of urban sewer detection.Rough manual interpretation distance cannot keep up with the intelligent detection and interpretation of sewer pipe defects.In this paper,an algorithm is proposed to improve the traditional manual interpretation of distances so that the spatial location of defects can be automatically and accurately located.Firstly,we investigated,analysed,and described the imaging characteristics of sewer defects and pipe joints.Secondly,we constructed a monocular vision-based ranging model using the mapping relationship between the imaged points and object points.Then,we propose a defect spatial location algorithm and implement it using the Python programming language.Finally,we make a visual representation of the defects on a pipeline map,which provides important reference material for the maintenance and rehabilitation of urban sewers.The content of this paper is composed of the following parts:(1)We observed CCTV inspection videos and images of pipelines in the main area of Kunming and obtained the generality of structural defects through a sampling and evaluation step to optimise the defect and pipeline joint detection algorithm.(2)We analysed the mapping relationship between object and image points in the CCTV inspection imaging process,proposed a defect longitudinal correction model based on monocular vision,and verified the validity of the model with simulation data and actual engineering data.(3)Based on the basic properties of the pipe section,we have done research on the longitudinal location and representation of defects,and constructed a model for automated defect spatial location.The results of the experiment demonstrate an improved accuracy of this method of defect spatial location,and the robustness and stability of the model of the accurate acquisition of the longitudinal distance based on the monocular ranging between the inspection robot and the pipe defect.