In this paper, to identify the true and false bloods by using the photoacoustic spectroscopy, a photoacoustic detection system was established. The cow bloods and props bloods was used as the true bloods and false bloods. Based on the established photoacoustic detection system, the real time photoacoustic signals of cow blood and props blood samples were captured with 128 times average. Moreover, the photoacoustic peak-to-peak values of true and false blood samples at 700nm-1064nm were obtained. From the amplitude and profile of the real time photoacoustic signals of cow blood and props blood, the obvious difference between both of them can be found. To achieve the recognition of the true and false bloods, the photoacoustic peak-to-peak values of cow blood and prop blood samples were firstly standardized, then, the interval PLS (iPLS) algorithm was used to choose the characteristic wavelengths of identifying bloods. The photoacoustic peak-to-peak values of blood samples at the chosen characteristic wavelengths (700-715nm) were used, and the 10 true bloods and 10 false bloods were used as the train samples, 5 true bloods and 5 false bloods were used as test samples. The linear discrimination analysis (LDA) algorithm was used to distinguish the true and false bloods. The test results show that under the characteristic wavelengths chosen by iPLS, the correct classification rate (CCR) was improved from 0 to 100%. Results show that the LDA coupled with iPLS, i.e., LDA-iPLS, has good identification effect for the true and false bloods.
|