Quantitative detection of the colloidal gold immunochromatographic strip in HSV color space

In this paper, a fast, reliable and accurate quantitative detection method for the colloidal gold immunochromatographic strip(GICA) is presented. An image acquisition device which is mainly composed of annular LED source, zoom ratio lens, and 10bit CMOS image sensors with 54.5dB SNR is designed for the detection. Firstly, the test line is extracted from the strip window through using the H component peak points of the HSV space as the clustering centers via the Fuzzy C-Means(FCM) clustering method. Then, a two dimensional eigenvalue composed with the hue(H) and saturation(S) of HSV space was proposed to improve the accuracy of the quantitative detection. At last, the experiment of human chorionic gonadotropin(HCG) with the concentration range 0-500mIU/mL is carried out. The results show that the linear correlation coefficient between this method and optical density(OD) values measured by the fiber optical sensor reach 96.74%. Meanwhile, the linearity of fitting curve constructed with concentration was greater than 95.00%.