Intelligent It-based Gold Immunochromatography Test Strip Quantitative Test System

With several advantages, such as ocular, fast, reliability and easy-operation, the colloidal-gold strip has been widely applied in clinic test. Because of many kinds of disturb factors, quantitative tests are extremely difficult and remain in the stage of qualitative test at present, and few tests are half- quantitative. So its application is restrained. The thesis studies the smart quantitative measurement system of immune colloidal-gold strips, specially on the application of median filtering and wavelet in the system. The system may be used in clinic test and POCT. The main research contents of this thesis have the following several respects: (1) Introduce briefly the current development situation of biomedical measurement method and the test. (2) Introduce separately the characteristic of the electric light sources, including LED, gas lights and cold cathode fluorescent lamps, which are adopted in this system. Study both photoelectric principle of the test and construction for the test system. (3) For the characteristic of signal from the test of immune colloidal-gold strips, this paper compares the effect of Fourier filtering, median filtering and wavelet filtering for signal processing and study application of median filtering and wavelet filtering for the quantitative tests. Proposes a new method mixing with median filtering and wavelet filtering and puts it into signal processing in this system, what's more, the good effect has showed. (4) According to the characteristic of genetic algorithms based on probability, a searching strategy is put forward for genetic algorithms, which works in narrow range and then wide one . The algorithms is carried into training connection weights of NN and simulation experiments show the arithmetic can escape local optima and improve learning speed of NN to some extent. A new NN is presented and put into modeling work for the test . The empirical test indicates this kind of NN can perform excellent than traditional modeling methods such as linear interpolation, cubic spline interpolation.