水工建筑物混凝土裂缝非接触高精度监测设备研发OACSTPCD

At present,crack detection in hydraulic structures primarily relies on manual identification,supplemented by conventional instrument detection,achieving a certain frequency of detection that is inefficient and has blind spots.Taking concrete cracks in hydraulic structures as the research object,a crack image recognition system platform and monitoring equipment were comprehensively developed.Firstly,a non-contact high-precision monitoring device for concrete cracks was designed.For the image processing module,deep learning U-Net technology was used to achieve precise segmentation of crack images.With the help of segmentation processing operations,end-to-end image processing was achieved.Subsequently,based on the image erosion technique of the perpendicular line,the crack width was calculated,and the crack length was measured by removing noise using the Hough transform.According to relevant regulations,concrete components were divided into four categories:slabs,beams,piles and pile caps,blocks,breast walls,and piers.Furthermore,crack monitoring and rating criteria were improved.Finally,the reliability and accuracy of the engineering prototype of high-precision monitoring equipment for cracks in hydraulic structures were tested.On the surface of the results,the non-contact high-precision monitoring equipment for concrete cracks in hydraulic structures achieves an accuracy of 0.08 mm in crack recognition.This systems enables automatic identification,type classification,accurate measurement,rating warnings and other functions of cracks,achieving the intelligent monitoring effect of structural cracking in hydraulic structures.