Research On Detection Methods Of Building Surface Cracks Based On UAV Image Acquisition

The construction industry is an important industry supporting social and economic development.However,the inherent characteristics of labor-intensive and on-site production in the construction industry have resulted in low production efficiency.In order to solve the problems of low production efficiency and extensive management forms in the construction industry,intelligent construction has gradually been recognized by the construction industry as a new project construction model.At present,the intelligent construction management model has become an inevitable trend in the development of the construction industry.The technological change centering on automation and intelligence is an important mean to improve the production efficiency of the construction industry in China.During the construction of the project,the detection of surface cracks in buildings is a very important part.For a long time,the detection of surface cracks in buildings has mainly depended on manual work.The efficiency is not high and the effect is not good.The safety of the inspectors is difficult to guarantee.With the increase in the height of buildings,it has been difficult to timely find and measure cracks on the surface of buildings by human eyes and contact measuring instruments.Therefore,exploring an intelligent,automated method for the detection of surface cracks in buildings is a smart construction and a basic requirement for the modern era.It has important innovation and guiding significance for building quality and safety inspections and the construction of smart construction sites.This thesis is supported by the National Natural Science Foundation of China(51408266)and the Humanities and Social Science Fund of the Ministry of Education(14YJCZH047).In this paper,four-rotor aerial UAV are used to collect images of building surface cracks.Then,digital image processing technology is used to realize the identification of cracks,extraction and measurement of features,and on this basis,improvements and optimization are made to improve the efficiency of detection of surface cracks of buildings.The main work is as follows.(1)This paper analyzed the causes of building surface cracks and a large number of building surfaces cracks images.It pointed out the characteristics of temperature change cracks,uneven settlement cracks and shrinkage cracks,such as their position,shape,grayscale and edge of images.Moreover,the feasibility and shortcomings of the basic algorithm of digital image processing in the treatment of building surface cracks are demonstrated.(2)Based on the features of the building crack images,an adaptive filtering method is used instead of a four-neighborhood smoothing filter to improve the classical SFC method.The experimental results show that the improved SFC method can accurately enhance the crack area,smooth the surrounding isolated convex points and water stains,and there is no significant difference in vertical,horizontal,and oblique crack reinforcement.Using Otsu to automatically find the high and low thresholds,improved the Canny edge detection operator.The experimental results show that the edge of the building surface extracted by the improved edge detection operator has better continuity and single pixel.A method of screening crack images based on shape features was proposed.By extracting the coordinates of the pixel points at the edges of cracks and fitting the features,the existence of cracks was judged based on the calculated curvature of the fitted curve.The experimental results show that this method can effectively determine whether there are cracks in the image and eliminate the straight line of cracks,in order to improve the accuracy of crack measurement.(3)The width,length and area of the cracks were measured on the basis of the extraction of building surface cracks' edge.In the crack width measurement,the horizontal and vertical directions were measured using the scribing method and the minimum distance method of the edge points,respectively,and the oblique cracks were measured using the edge point minimum distance method and the tangent perpendicular line method;The skeleton centerline method and the minimum bounding rectangle method were compared.For the measurement of crack area,the number of pixels in the crack region was calculated by the pixel method of the connected region.The real crack area was calculated after conversion.Matlab programming experiments were used,and the applicability of different methods for different types of cracks was analyzed.(4)In order to realize the identification and the measurement of features of building surface cracks better,based on the above steps and methods,a detection system for building surface cracks based on UAV collected images was constructed,and the architecture,functions,and processes of the system were designed.200 images of the building surface cracks collected by the UAV were used as test objects and the system was tested.Tests show that the system can effectively extract the cracks in the image,and reduce the false detection rate of cracks through screening.The test results of crack measurement show that the measurement error of this system is small,and it can meet the requirements of surface crack feature measurement of buildings.It can provide a reliable basis for building safety and quality assessment,and has practical application value.In terms of system operation experience,the interference factors in the drone image acquisition operations were analyzed,and corresponding countermeasures and suggestions were proposed.At the same time,it also analyzes the deficiencies of the system in terms of intelligence,automation and clarifies the future direction of improvement.