Abstract
In this work, a set of photoacoustic detection system of blood was established to identify the true blood and fake blood. The time-resolved photoacoustic signals and peak-to-peak spectra of blood samples were obtained in the wavelength from 700nm to 1064nm. In experiments, five kinds of blood in total of 150 groups were used, where three kinds of blood were the animal true blood, two others were the fake blood. The experimental results demonstrated that the true and fake blood can be easily and accurately identified from the time-resolved photoacoustic signals or peak-to-peak spectra due to the overlapping of signals or spectra. To accurately identify the true and fake blood, back propagation (BP) neural network was used to supervised train the peak-to-peak values of training blood sample. The correct rate of identifying true and fake blood based on BP is 76.7%. To improve the correct rate, the particle swarm optimization (PSO) was employed to optimize the parameters of BP including weights and thresholds. Moreover, the effects of neurons number, learning rate factor, inertia weight, two acceleration factors, iteration times and training times on the correct rate were all investigated and compared with BP. Under the optimal parameters, the correct rate of BP-PSO algorithm was improved to 96.7%. Therefore, the photoacoustic spectroscopy combined with BP-PSO algorithm has the potential value in the identification of blood.

This publication has 7 references indexed in Scilit: