Recognition And Identification Research On Concrete Crack Characteristic Information

Crack is an important content in inspecting concrete structures, and its characteristic information can reflect the service state of the concrete structures. The traditional crack inspection mainly depends on the artificial mode, but it is characterized of low inspection efficiency, great labor intensity and so on. Accordingly, based on the methods such as image processing and so forth, this paper represents detailed researches on the recognition and identification of concrete crack characteristic information, and a suit of concrete crack characteristic information recognition system is hereby developed, which improves the deficiency of current relevant techniques. The main research work is set out as follows:(1) First of all, the image recognition and identification researches are carried out regarding the concrete cracks, overall analyses are made about relevant crack image preprocessing, extraction and calculation of characteristic information and so on, and relevant algorithms suitable for the crack image recognition in modern algorithms are selected through comparisons.(2) In order to improve the identification effect of the traditional algorithms on the concrete cracks, two segmentation algorithms including two-dimension Otsu method and Markov random field have been improved and introduced on the basis of the original algorithms, and respective application advantages of such two methods are analyzed and confirmed in the field of crack image processing through comparisons of the actual processing effects of crack images.(3) Researches are carried out about the crack classification problem, and the classification thoughts of cracks are proposed in this paper; in VC++, relevant algorithm functions are realized through use of Matlab program codes compiled in the dynamic-link library.(4) Based on the researches above, a suit of crack characteristic information recognition system is designed. Beginning with the frame composition and work principle, relevant hardware supports and software developments are analyzed and discussed. For example, as for the hardware, the devices are researched, including CCD camera, image capture card and so on; as for the software, except for the latest increase of Markov random field segmentation algorithm, etc, new functional modes, such as automatic processing, batch-based processing, crack inspection recognition report output and so on, are hereby researched and realized; finally, the concrete structure state assessment module is introduced integrating the extracted relevant information values and structural design parameters.