小波变换在基于图像处理的金免疫层析测定中的应用  

Application of wavelet transform in image processing of gold immunochromatographic strip quantitative test

作　　者：姜海燕[1,2] 杜民[3,2] 

机构地区：[1]福州大学电气工程与自动化学院,福建福州350108 [2]福建省医疗器械和医药技术重点实验室,福建福州350108 [3]福州大学物理与信息工程学院,福建福州350108

出　　处：《福州大学学报（自然科学版）》2011年第2期222-226,共5页Journal of Fuzhou University(Natural Science Edition)

基　　金：科技部国际合作资助项目(2009DFA32050);福建省自然科学基金资助项目(2009J012802009J01281);福建省教育厅科研资助项目(JB08019)

摘　　要：采用小波变换软阈值法有效去除试条图像噪声,根据小波分解后的系数重构图形定位GICA的光谱信号突变点的位置,提取试条光谱特性曲线的特征参量.采用BP神经网络对特征参量与浓度之间的多变量非线性函数关系进行建模,HCG试条测定结果表明,该测定方法相对误差符合要求,适用于GICA的定量测定.This thesis uses wavelet soft -threshold denoising to preproeess the strip image. According to the analysis of maximum module of wavelet transform coefficients and their orientation in different scales, the breaking points of signals could be determined accurately by the wavelet transform and the useful characteristic parameters were extracted. By the training for the back -propagation neural net- work, the muhivariables nonlinearity relationship between the HCG concentration and characteristic parameters is established, the result shows that HCG relative error can be decreased effectively by this method; it can be a good method for quantitative test of the gold immunochromatographic strip.

关 键 词：定量测试 小波变换 HCG GICA 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]