Improvement of microassemblies ultrasonic images using adapted signal processing techniques

This paper deals with two techniques for ultrasonic images understanding and interpretation with image and digital signal processing first with a specific test sample. A simulation allows us to evaluate the effective Point Spread Function (PSF) of a reflection scanning ultrasonic system working in the 10-100 MHz frequency bandwidth and successfully implemented. From this PSF, we can deduce lateral resolution both in the coupling medium and in a material under test. We present the great interest of the PSF knowledge for image deconvolution on a Lucy-Richardson algorithm applied on this test sample. With this method, lateral resolution is strongly improved and real dimensions of an object can be reached. The second technique concerns ultrasonic images with different nature of defects. A better understanding of these images can be realised by means of time of flight detection with Short Time Fourier Transform (STFT) analysis to determine defect depth with strong accuracy.