Application Study Of Data Fusion To Structural Damage Inspection Of Concrete

Data fusion technology introduces a new idea to the field of engineering structural inspection. In this paper, the framework of data fusion technology applied to the field of structural inspection of concrete is preliminarily established by means of combining infrared imaging and ultrasonic CT technology. The 3D quantitative reconstructions of defects and the qualitative identification of defects types can be obtained by the proposed technology.In the field of digital image processing image fusion is an data fusion technology. In this paper, the image fusion technology is applied to the damage detection of concrete in the first time. The comprehensive and particular image information can be used to determine the interior defects of structures effectively and accurately. Image fusion algorithms based on the multi-resolution wavelet transform and wavelet packet transform are explained in detail at the pixel level. Two kinds of fusion rules, current edge gradient and local variance rules are proved to be improper through theoretical and experimental investigations. Therefore, two kinds of modified fusion rules are proposed. A series of experimental researches are carried out to detect the interior damages of reinforced concrete beams and slabs by infrared thermography. The results show that the proposed regional fusion rules are superior to other fusion rules.To overcome the limitation of existing infrared imaging technique, a new developing technology of infrared CT, which combines the infrared imaging and the heat conduction numerical simulation, is used to identify and reconstruct the defects by 1D and 2D physical models. The artificial neural network is employed in this analysis. On the other hand, ultrasonic CT technology is taken to detect interior damage of concrete to realize 3-D image reconstruction.On the ground of the two kinds of inspection methods, i.e. the infrared CT simulation and the ultrasonic CT, multi-sensor data fusion technology is utilized to realize quantitative analysis of defects sizes and qualitative description of defectstypes at pixel level, feature level and decision level. To acquire fused 3-D images pixel level fusion is performed by the evidential theory to integrate infrared CT simulation into ultrasonic CT. The two features, that is heat conduction coefficient extracted from infrared imaging and ultrasonic velocity from ultrasonic CT, are brought up to fuse at feature level. Then cluster analysis based on the principle of minimum-distance is carried out to evaluate the types of the defects. At decision-making level, the evidential theory is put forward to perform the global decision to acquire the types of the defects.