Research On Crack Detection System Of Concrete Surface Based On Android Platform

As a common disease form in reinforced concrete structures,concrete cracks are one of the key indicators that reflect the health status of structures and evaluate the quality of construction projects.Regular monitoring of crack information on concrete surface and being repaired in time can effectively avoid accidents.The traditional crack detection methods have some disadvantages such as low detection efficiency and high cost,and they cannot meet the huge concrete crack detection requirements today.With the rapid development of computer hardware and software performance and mobile internet technology,a new type of non-contact nondestructive damage evaluation technology was developed by combining digital image processing technology and wireless transmission technology of smart phones,and its application in concrete crack detection has outstanding engineering application value and social benefits.Combined with android development technology,the correlation between the basic pixel size of smartphones,phone models,and camera zoom factor through experiments were analyzed in this thesis.In addition,it was applied as a benchmark to develop a non-contact and non-destructive detection technology for cracks on the concrete surface based on the android platform.This technology can automatically complete the calculation of the crack characteristic value through collecting the image of crack area by driving the camera of mobile phones,and using digital image processing technology to accurately identify the crack information in the crack image.Compared with traditional detection methods,not only this new technology has a simple detection process and convenient detection equipment,but also the crack detection results can be transmitted in real time,and the real-time detection was provided basis for off-site expert decision-making through the communication function of smartphones.The research content of this thesis has certain frontiers and pioneers due to it conforming to the future trend of crack detection automation and digitization and following the rapid development of Io T(Internet of Things)technology and big data analysis.Within the scope of this study,the main conclusions can be drawn as follows:(1)The conversion coefficient ? was obtained and a method of mobile phone camera calibration based on the size of basic pixel was proposed by experimentally exploring the conversion relationship between the number of crack pixels in the image and the actual size.The results indicated that the basic pixel size had an exponential function relationship with the mobile phone camera pixel and camera zoom factor.The smaller base pixel size could be obtained by increasing the camera pixels or increasing the camera zoom factor.The error of oblique calibration experiments indicated that the camera shooting angle had little effect on the calibration results when the camera lens plane was parallel to the plane of the target,and the deviations caused by this was in acceptable ranges.But,the greater errors of the calibration results would be brought by tilt angle of the mobile phone,and it can be accepted within 5°.(2)The image features of cracks on the concrete surface were analyzed in this study.The results showed that the image preprocessing algorithm based on adaptive size mean filtering was better for the enhancement of the collected cracks,the crack target information was highlighted.The extraction effects of threshold segmentation method,edge detection operator and flood filling algorithm on crack target were analyzed through experimental research,the results showed that the flood filling algorithm had good robustness and could achieve complete extraction of crack targets.(3)Based on Android Studio development environment,the design and development of the concrete surface crack detection APP was completed in this study.APP had realized the functions such as "crack detection project management","crack image import","crack image processing","crack characteristic value detection calculation","detection result storage",and were the reliability and ease of use.(4)The measurement accuracy of the concrete surface crack detection APP was verified through comparing with the standard crack detection card.The experimental results showed that the APP detection method can meet the accuracy requirements of crack detection in engineering.The detection distance of the APP detection method was restricted by the accuracy requirements.In order to ensure the measurement accuracy,the different camera zoom factor were adopted at different detection distances.Based on the experiment of detection accuracy analysis,the allowable error range was set as the absolute error is less than 0.1mm and the relative error is less than 10%.The applicable range of the detection distance of this method and the recommended camera zoom factor at different detection distances were proposed in this thesis.(5)The influences of basic pixel size,shooting distance and camera zoom factor on the measurement accuracy of crack detection APP width were analyzed through comparing with the actual concrete crack detection results.The experimental results indicated that the measurement accuracy of the crack width of the basic pixel size obtained by linear interpolation can be improved by 5%?10% and the guarantee rate of the measurement error within 10% Increased from 61.67% to 93.33% compared with the basic pixel size obtained by the exponential fitting.It could effectively improve the measurement accuracy by shortening the shooting distance and increasing the camera zoom factor.The APP detection method can meet the accuracy requirements of crack detection in engineering.