Técnicas de simulación y filtrado de ruido en señales de espectroscopía RMN mediante transformadas wavelets

Abstract

In this article is shown the development of simulation programs for studying the techniques of noise filter in NMR (Nuclear Magnetic Resonance) signals, known as FID (Free induction decay). This signals are generated by exciting resonant nucleus with radio frequency pulses in an exterior magnetic field, presenting a complex exponential wave with gaussian noise. The computed programs may generate FID arameters signals definable by the user, add gaussian oise with a SNR determined.The signals with a low SNR may be treated by the traditional method of accumulations. Another considered alternative is the noise suppression through wavelets. This innovated technique promises to be quick and efficient. The created programs areused for the study and further development of these techniques, making more easy the results evaluation through the MonteCarlo method. The calculus is easier in different cases of the relation signal to noise residual in dB, establishing the difference signal without noise and with noise,also for the filtered method using accumulations so WT and WPT. From these programs and the associated methodology the investigations of better filtration variables for noise suppression in the FID signals is easier.