Concrete Crack Detection Method And Application Based On Hierarchical Clustering

Concrete bridge is the most common construction type of modern bridge,which is not only durable,high strength,but also low cost.Regular inspection of the bridge,as soon as possible to find threats to the safety of the bridge structure cracks,and timely maintenance,is conducive to the long-term and safe use of the bridge.Traditional manual testing methods are time-consuming,laborious and dangerous,which cannot meet the growing demand of bridge testing.Thanks to the rapid development of computer and embedded technology in recent years,more and more researchers begin to pay attention to and study the non-destructive testing technology of concrete bridge based on digital image processing.Aiming at the noise connected domain in the image after threshold segmentation,a morphological filtering method of connected domain area based on hierarchical clustering is proposed,and an adaptive selection method of connected domain area threshold is designed.By obtaining the area threshold of the single connection distance of the connected domain area sequence,the noise connected domain can be filtered out.This method can filter out a lot of noise and false cracks,obtain clear crack images,and realize the recognition of the crack images.Aiming at the phenomenon of fracture connected domain,this paper proposes a morphological expansion method based on the trend of fracture connected domain endpoint region,and designs an extraction method of fracture end region and its direction.By expanding the endpoint of fracture connected domain along the direction of the endpoint region,the fracture is connected.This method improves the integrity of fractures while retaining fracture shape information..In this paper,a portable embedded system is designed for real-time detection of concrete bridge surface cracks.The system is designed based on the AM3352 core board,and the image is acquired by the industrial camera OV9712,which realizes the functions of image collection,display,crack detection and fracture image storage.Finally,the effectiveness and accuracy of the crack detection system designed in this paper are verified by real-time acquisition of concrete crack images and non-crack images,and the results are analyzed.The test results show that the crack detection accuracy of the proposed method is up to 99.35%,and the detection speed of the method is fast,which can meet the requirements of real-time detection and practical engineering applications.