Systems and methods for predicting volume solids measurement of paints based on non-invasive photoacoustic sensing

A system comprising:

• a memory storing instructions;
• one or more Input/Output (I/O) interfaces;
• and
• one or more hardware processors coupled to the memory via the one or more I/O interfaces, wherein the one or more hardware processors are configured by the instructions to:
• receive a plurality of wet paint samples in a transparent sample container of a predefined width and a plurality of volume solid (VS) labels associated to the plurality of wet paint samples;
• provide an excitation signal to each of plurality of wet paint samples using a pulsed laser diode of a predefined wavelength and a predefined optical power;
• obtain a first type of time domain photoacoustic (PA) signal corresponding to each of plurality of wet paint samples using a sensor device coupled to the transparent sample container through a coupling gel, wherein the time domain photoacoustic (PA) signal corresponding to each of plurality of wet paint samples is generated based on the excitation signal provided to each of plurality of wet paint samples;
• generate a second type of time domain photoacoustic (PA) signal corresponding to each of plurality of wet paint samples by performing a plurality of operations on the first type of time domain photoacoustic (PA) signal corresponding to each of plurality of wet paint samples;
• extract one or more features from the second type of time domain photoacoustic (PA) signal corresponding to each of plurality of wet paint samples using a signal analysis technique;
• determine a velocity parameter of the second type of time domain photoacoustic (PA) signal corresponding to each of plurality of wet paint samples based on the one or more features and the predefined width of the transparent sample container;
• and
• predict a volume solids measurement for the plurality of wet paint samples by applying a trained regression based prediction model on the velocity parameter, wherein the trained regression based prediction model is based on one or more trained velocity parameters of the second type of time domain photoacoustic (PA) signal obtained during training.

One or more non-transitory machine-readable information storage mediums comprising one or more instructions which when executed by one or more hardware processors cause:

• receiving a plurality of wet paint samples in a transparent sample container of a predefined width and a plurality of volume solid (VS) labels associated to the plurality of wet paint samples;
• providing an excitation signal to each of plurality of wet paint samples using a pulsed laser diode of a predefined wavelength and a predefined optical power;
• obtaining a first type of time domain photoacoustic (PA) signal corresponding to each of plurality of wet paint samples using a sensor device coupled to the transparent sample container through a coupling gel, wherein the time domain photoacoustic (PA) signal corresponding to each of plurality of wet paint samples is generated based on the excitation signal provided to each of plurality of wet paint samples;
• generating second type of time domain photoacoustic (PA) signal corresponding to each of plurality of wet paint samples by performing a plurality of operations on the first type of time domain photoacoustic (PA) signal corresponding to each of plurality of wet paint samples;
• extracting one or more features from the second type of time domain photoacoustic (PA) signal corresponding to each of plurality of wet paint samples using a signal analysis technique;
• determining a velocity parameter of the second type of time domain photoacoustic (PA) signal corresponding to each of plurality of wet paint samples based on the one or more features and the predefined width of the transparent sample container;
• and
• predicting a volume solids measurement for the plurality of wet paint samples by applying a trained regression based prediction model on the velocity parameter, wherein the trained regression based prediction model is based on one or more trained velocity parameters of the second type of time domain photoacoustic (PA) signal obtained during training.