Detection Of The Concrete Bridge Bottom Crack Based On Digital Image Processing

Bridge collapse accidents occurred frequently in the world, and caused seriouspersonal injury and economic losses. According to statistics, more than90%of theconcrete bridge damages are caused by cracks. So cracks are the focus of bridgedetection. Currently, the artificial detection method is widely used in bridge crackdetection at home and abroad. However, there are many serious deficiencies, such ashigh labor intensive, low efficiency, high cost, low security, and so on. In order toimprove the current crack detection status, we researched and proposed a method ofdetecting cracks on the bridge bottom surface using digital image processingtechnique, developed the software to analyze crack images, identified and extractedcracks, calculated their parameter.This thesis first briefly described the background of our research and analyzedthe present research status of bridge crack detection of domestic and foreign,illustrated that there are many advantages on using digital image processing techniqueto detect bridge cracks. It adopted the movement way of combining four-wheelchassis and wheel rail when collecting cracks images. And used the industrial CCDcamera to scan the bottom surface of bridge, which is fixed on the movement device.Then, by the use of image acquisition card for A/D conversion, compression coding,we could transmit the images to the computer through PCI bus. At last, based on thebasic image processing algorithms, it developed the intelligent cracks recognitionsoftware to analyze crack images, identified and extracted cracks, and record theparameters of the crack, to assess the crack damage of bridge.Due to the complex environment and light condition under bridges, imagescollection was affected seriously. In addition to crack, the images may also containother disrupting chemicals, and this will influence the identification result. To solvethose problems, this thesis firstly carried out the pre-processing of original image,including graying, gray scale correcting, median filtering, edge detection, adativethresholding，morphological processing, etc. After these steps, it could get the better crack binary images and could analyze them. Based on the analysis of characteristicvalues of crack, and the combination of characteristics of the cracks, it could achievethe extraction of crack targets. To connected the crack fragments, so that we couldreduce the appearance of cracks, and measure their characteristics. According to thoseresults, we can conveniently assess the degree of crack damage and repair the bridgelately. At last, according to the method proposed in this paper, the software wasdeveloped to identify and measure cracks by means of Visual C++6.0． Theexperimental results showed that the method was effective under certain conditions.Compared with the traditional detection methods, this system improved theefficiency and precision(arrive at0.1mm) of the detection of crack on the bridgebottom surface, and reduced the labor cost, shortened the bridge closure time, andensured the safety of the inspectors, improved the engineering detection quality,promoted the development of social economy.