Automatic detection of characteristic points and form of optical signals in multiparametric capillary sensors

The time series sequence of data readings are the input for computer aided analysis of signal from the multiparametric optical capillary sensor. The time series signals have characteristic points and forms. Their analysis by trained human operators is time consuming and sometimes lacks of precision because of the presence of signal noise. The noises can be mostly rejected with advanced electronic signal processing, but the output analog signal is often modified by the electromagnetic environment and by the noise generated by electronic elements. We propose and analyze an algorithm that can be used as an automatic detector of characteristic points and form of the time series signals that are produced by the measuring head and the analog electronic units of the biodiesel fuel quality test sensor.