Algoritmo de processamento digital de sinais baseado em matlab, aplicado a auscultas pulmonares de estetoscópios digitais

Abstract

The pulmonary auscultation process is very important for obtaining a pre-diagnosis for pulmonary diseases, being a non-invasive and easy-to-answer test, it can detect the beginning or the occurrence of pulmonary diseases. For this reason, this practice must be performed with maximum efficiency, in a non-noisy environment and with instruments that have a good response to the sounds of a lung. The main instrument for lung auscultation is called a stethoscope. This device has the purpose of filtering the sounds of the lung, and sending this sound to the ears of the specialist doctor. There are two types of stethoscopes on the market: analog and digital. The analog stethoscope does not have electronic devices, and is just a device with mechanical tubes that conduct sound through vibrations to the specialist's ears, noting that there is no filtering of the heart signal. The digital model, on the other hand, has electronic devices and several analog and/or digital filters to try to focus on the sound of the heart or lungs. This work aims to present a digital signal processing algorithm which can filter only the lung sound, eliminating the heart sound, which for this application will be treated as noise, in addition to other existing noises. For this purpose, digital signal processing techniques were used, such as windowing as a filter, in addition, to reduce some noise, the 6 sigma quality control method was used, with regard to the sigma level as tolerance and elimination of outliers . With these techniques, this treatment should be exposed by means of tests on an auscultation bank obtained properly.