Welding is a major task for any component fabrication in every industry. Irrespective of the number of years of the science and art of welding, defects may continue to occur in its lifetime or during welding. The welds are expected to be tested and evaluated using codes of practice. Nowadays industries extensively used ultrasonic testing to detect and evaluate the defects/flaws in the weldments. Ultrasonic testing has extensive developments in the last two decades because of the advancement in sensor and signal analysis technologies. Time of flight diffraction is the advanced technique which has better probability of detection of linear defects. A major irritant during application of ToFD especially for the inspecting of austenitic stainless steel weldments is the presence of noises. Such noises have been reduced utilizing different approaches and each has its own merits and demerits. This paper focuses on the application of discrete wavelet transform (DWT) decompositions with lower-order wavelet filters for de-noising ultrasonic ToFD B-scan images. Analysis clearly indicates that this approach also gives greater SNR improvement with less computational time.

coiflet, daubechies, discrete wavelet transform, symlet, time of flight diffraction.