Ultrasonic CT Image Analysis And Processing Based On Concrete Non-destructive Testing

The research in this paper is supported by the NSFC's Funding: Smart Concrete and its Structure( Program No:50238040).Based on the concrete non-destructive testing technology, the principles and experimental methods of ultrasound CT are explored, thereby launching a series of image analysis and post-processing. The main conclusions are:1 Discussion on characteristics of concrete Ultrasonic CT images: To a certain extent, CT images can reflect the existence and location of internal flaws, but can not accurately represent the size and shape. While CT images to the size change are more sensitive, but it is very difficult to identify the shape of concrete flaws.2 CT image quality analysis of the effects of numerous factors: For example, the combinations of flaws, the relationship between the number of survey points and the number of imaging modules, inverse model, calculation accuracy. And given the corresponding solutions, as far as possible eliminate the CT images of "pseudo-like."3 Establishing a task-based CT image quality evaluation system: From the resolving power, readability and reliability of the CT image, the quality of CT image is discussed, then giving a semi-quantitative criterion.4 Respectively, we use median filter, clustering analysis and variance cutoff techniques to process the images, and compare three methods. Integrated technology for the experimental results dealing with, and enhance the reliability of the images, making flaws region and backgrounds region differences apparent.5 We use traditional methods to divide CT images, finding it difficult for the border determination of the concrete flaws. Image partition technology based on the border transition zone is introduced, the methods used for the flaws and the background region dividing, identifying concrete internal flaws, and the results are more satisfactory.