Applied Study Of Crack Measurement Methods In The Surface Of Buildings Based On Image Processing Technology

With the accumulation of its disrepair, surface cracks of buildings are becomingmore and more obvious and serious. The cases of affecting the quality of projects andhousing collapse resulted from cracks are common. Features of cracks reflect thecurrent status of working and structure of building facilities. Aesthetics, durability andeven the safety of persons and properties are influenced and threatened if featurevalues surpass the security thresholds. Therefore, the correct assessment to surfacecracks of buildings possesses the important practical significance to the assessment,identification and maintenance of the whole structure.The traditional artificial visual method and acoustic emission method haveexposed some defects including high subjectivity, low accuracy and efficiency whenutilized to measure the crack features, which will be replaced gradually by newmethods. With the continuous development of computational image processingtechnology, its application has depth into the quantitative evaluation of surface cracksof building facilities, furthermore, play strengths of non-contact, relative convenience,intuitiveness and gradual precision.The paper applies image processing technology to process the crack images andmeasure features on them. Image processing and crack feature measurements are thetwo core parts in application research. Image processing mainly comprises imagecapture and calibration, image gray, image preprocessing, image edge detection andextraction, binary processing. The image-gray utilizes the RGB-to-YUV color modelconversion to achieve the lossless transformation from color to grayscale. The imagede-noising applies the wavelet threshold de-noising algorithm based on improvedthreshold function and the image enhancement adopts the gray-stretch algorithmbased on median filter and Laplacian operator. The image edge detection andextraction uses the algorithm of the improved Canny operator gray threshold edgedetection and location, based on traditional differential operators and sub-pixelpositioning algorithms, achieve coarse location, furthermore, the sub-pixel locationalgorithm of Gaussian curve fitting based on neighbor gradient features realizesub-pixel precision positioning. The image binarization introduces the optimalthreshold method to extract the crack region from images. The measurement andcalculation to cracks mainly includes that of width, length and area. The width measurement utilizes the algorithms of horizontal distance and tangent perpendicularmethod. The length measurement adopts the algorithm of central line and the area ismeasured through counting the number of pixel-points. The accuracy of algorithmsproposed by this paper has been verified through experiments.Finally, the image processing toolbox and wavelet transformation toolbox, in theplatform of Matlab R2009b, are utilized to analyze and process the crack images,measure and calculate the crack features. It proves that the accuracy and feasibility ofthe algorithms in this paper are relatively high in the aspects of crack edge detectionand extraction and feature measurement, which play advantages of image processing.The deviation could be controlled within the accepted scale. Therefore, the algorithmspossess important practical significance in assessing, verifying and maintainingsurface cracks of building facilities, also could be introduced into the areas ofnation-defense, aerospace-aviation and machinery manufacturing, etc.