基于金字塔连接与高斯混合模型算法的层析芯片荧光信号检测

Fluorescent Signal Detection of Chromatographic Chip by Algorithms of Pyramid Connection and Gaussian Mixture Model

提出了一种层析芯片荧光信号自动识别的算法。该算法使用基于RGB空间的数学形态学运算方法对图像滤波增强,并采用金字塔连接算法分割图像,然后使用高斯混合模型(GMM)检测荧光信号,最后再计算荧光区域的平均亮度。经实验证明,该算法能将荧光区域完整地分割出来,并能快速准确地检测出荧光信号,实现层析芯片的定量检测。

We proposed a new algorithm for automatic identification of fluorescent signal. Based on the features of chromatographic chips, mathematic morphology in RGB color space was used to filter and enhance the images, pyramid connection was used to segment the areas of fluorescent signal, and then the method of Gaussian Mixture Model was used to detect the fluorescent signal. Finally we calculated the average fluorescent intensity in obtained fluorescent areas. Our results show that the algorithm has a good efficacy to segment the fluorescent areas, can detect the fluorescent signal quickly and accurately, and finally realize the quantitative detection of fluorescent signal in chromatographic chip.