Study And Software Implementation Of Image Postprocessing Based On Time Of Flight Diffraction

Ultrasonic TOFD flaw detection is one of the important methods in the field of industrial non-destructive testing,with the advantages of high defect detection rate,accurate defect location,and high defect detection efficiency.To reduce the impact of near-surface dead zone and arc tail on TOFD defect detection,improved algorithms and software implementation for Lateral Wave(LW)processing and Synthetic Aperture Focusing Technique(SAFT)are focused in this thesis,and the design and implementation of a TOFD software running on the Android plarform is introduced,which improves the software function of portable TOFD instrument.The LW processing algorithm studied in the thesis includes three parts: automatic calibration,correction and suppression.Aiming at the problem that the LW calibration algorithm based on peak capture is sensitive to noise,the thesis introduces an improved Hough search algorithm before peak capture to improve the accuracy of results and the efficiency of data processing,and on this basis,the algorithms of LW correction and LW suppression are improved.According to the results of MATLAB experiments,the efficiency of calibration algorithm for LW has been increased by about 10 times,and the improved algorithms of LW straightening and LW suppressing perform better compared with the original ones.A SAFT algorithm combined with amplitude apodization and sparse aperture is introduced to enhance signal amplitude and reduce time of processing.According to experimental results on MATLAB,the algorithm can effectively enhance signal amplitude and shorten time of processing by nearly 1 times,when the Hanning window is selected as the amplitude apodization function and the effective aperture ratio is 0.25.In order to design an ultrasound TOFD software with image post-processing functions,the functional requirements of software are analyzed,and the overall design of software running on Android platform is formulated,which uses a layered architecture and divides each layer into different functional modules.A synthesis module for TOFD images based on Open GL ES is designed,and images synthesis in offline analysis mode is focused;the implementation for LW processing and SAFT are introduced based on JNI and multi-threaded mode to improve the efficiency of software.The testing results show that functions of the software work normally,and the LW calibration can replace manual calibration,reducing the error by 50%,the LW suppression can reduce the near-surface dead zone of a 20 mm workpiece from 8mm to 5mm,and the SAFT can reduce the average error of defect length quantification to 0.5%,which improve the accuracy of defect measurement;the software CPU usage is about 74%,and the memory space occupied is between 60 MB and 90 MB,which can meet the performance requirements of the system.