Improving The Flaw Sizing Using Ultrasonic C-scan Imaging Based On Deconvolution Method

The ultrasonic C-scan imaging is the importan method for industrial non-destructive evaluation.But Imersion planar transducer is likely to cause the edge scattering in ultrasonic C-scan imaging,resulting in inaccurate defect sizing.Focused on the above problem,the dissertation used the deconvolution algorithm to improve the sizing accuracy of the ultasonic C-scan images,and combined the totoal-variant regularization Richadson-Lucy algorithm with the Distance-Amplitude correction cruve to determine the final interation for the deconvolution,finally used the algorithm proposed above to improve the sizing accuracy of the ultrasonic C-scan images,the main highlights of this dissertation are as follows:(1)In order to analysis the propagation characteristics of acoustic beam in immersion test,the Multi-Gaussian beam model is built to simulate the ultrasonic propagation and sound field in immersion test.And using Multi-Gaussian beam model and experimented system function to build the ultrasonic measurement model,which can predict effectively the defect response from different transducer,types of the defects and depth.The simulation makes a foundation for studying the point spread function and Distance-Amplitude Correction Curve.(2)According to linear system theory,the ultrasonic C-scan imaging also can be obtained by the 2-D convolution between the point spread function and the object function,and obtain the point spread function analytical expression;Substituting the Multi-Gaussian beam model into the point spread function analytical expression,explaining the edge-blur problem of the ultrasonic C-scan images from the size of point spread function,and provide the known point spread function for the deconvolution.(3)The ultrasonic C-scan images were simulated by the linear system theoretical model,compared with the experiment resluts,the simulation is valid.Using the RL-TV algorithm restored the simulated ultrasonic C-scan images and built the rational relationship between the final iterations and the ratio of flaw size and wave number,from the above rational relationship,the sizes of the flaws are the key parameters,proposing to use the Distance-Amplitude Correction Curve to predict the sizes of flaws and the DAC curve was verified by experiments,and using the predicted sizes of flaws to calculate the final iteration by substituting into the rational relationship.And Using the RL-TV algorithm to restore the experimented ultrasonic C-scan images according to the calculated final iteration,which is not only determine the final iteration,but also restore effectively the blur edge of the ultrasonic C-scan images and improve the sizing accuracy.