Font Size: a A A

Concrete Bridge Cracks Automatic Detection Method Research Based On Machine Vision

Posted on:2019-06-18Degree:MasterType:Thesis
Country:ChinaCandidate:J ShaoFull Text:PDF
GTID:2392330596494790Subject:Engineering
Abstract/Summary:PDF Full Text Request
Bridge crack is one of main causes of bridge failure,so bridge crack detection is of great significance.Manual inspection requires the location of crack to be found,and only one point on the crack can be measured at a time,which is inefficient and costly.Machine vision inspection has the advantages of non-contact,high efficiency and high precision.This paper has proposed an automatic detection method for concrete bridge cracks based on machine vision to realize cracks automatic recognition and measurement.It can reduce the detection cost and improve the detection efficiency on the basis of ensuring the measurement accuracy.The main contents include:(1)A proportional threshold segmentation method is used to achieve crack segmentation,extract crack image features and identify crack images.Compared with existing target segmenting method,proposed the proportional threshold segmentation method,which has a great improvement in segmentation effect and processing speed.Characteristic matrix of crack is extracted which is include target coordinate variance,circularity,aspect ratio and average length.Artificial neural network classification model is used to identify crack image to verify the correctness of the segmentation and feature extraction.(2)An improved structured light strip separation method is proposed to complete the calibration of crack measurement system.Composition and measurement principle of the system are introduced.segmentation and separating structure light strip,fitting center line of the light strip,and then calibrate the system based on plane target.(3)Extract cracks complete contour,edge and skeleton,and then complete crack size measurement.Double-threshold crack expansion algorithm is used to extract complete contour of cracks.The hyperbolic tangent function is used to fit the edge gray level variation model to extract sub-pixel edge of cracks.Project edge point to target plane,extract cracks skeleton,and finally calculate cracks width.(4)Crack measurement experiments and results analysis.Firstly measure regular targets and analyze measurement results,and then measure concrete crack in field,compare measurement results with crack width gauge.
Keywords/Search Tags:Machine vision, Crack recognition, Edge extraction, Crack measurement
PDF Full Text Request
Related items