Non-destructive and rapid detection of blood quality in blood bags based on modified ACO wavelength selection algorithm

摘要

The concentration of free hemoglobin (FHb) in blood bag is an important index to evaluate blood quality. And the chemical method, whichis widely used in clinic, not only wastes precious blood resources, but also consumes time, be labor-intensive, and pollutes the environment. Althoughspectral analysis can make up for the shortcomings of traditional chemical methods, its accuracy also be seriously affected due to unstable the opticalparameters of the blood bag and complexity of blood composition. Therefore, this paper proposed a wavelength selection method based on modifiedant colony optimization (ACO) algorithm. This algorithm firstly calculates the correlation coefficient between each wavelength and the FHbconcentration as the initial pheromone for each wavelength to improve the convergence speed of ACO algorithm; Then, the optimized wavelengths aresorted according to the intensity of the pheromone, the number of optimized wavelengths for modeling is increased in turn, and the root-mean-squareerror of the model is used as an evaluation index to determine the final optimal wavelength. In this study, a small tube in the front of the blood bag wasused as a container, and the transmission spectrum of plasma in the small tube was collected. The regression results of the optimized spectrum werecompared with that of the original spectral. The experimental results showed that the method proposed in this paper can obviously improve the analysisaccuracy of blood quality, which means this method can effectively inhibit the influence of non-target components and the difference in blood bags onthe detection of blood quality.