荧光免疫层析试条光电信号处理及特征量选取

Fluorescent immune-chromatographic strip signal processing and feature selection

荧光免疫层析定量检测技术是一种正在兴起的即时检验技术。在前期荧光免疫层析定量检测仪的研究基础上,采用小波算法对A/D采样后的荧光信号进行去噪处理,选用中值滤波剔除基线漂移,最后比较了采用测试线峰值和面积作为特征量的定量检测效果。数据显示,去噪后信噪比提高10 d B,相对均方差减小30.4%;采用峰值和面积值作为定量特征量的测量CV值分别为5.1%和0.4%,测量误差分别为9.1%和0.9%。结果表明采用积分面积作为特征量对荧光信号进行定量检测具有更好的准确度和重复性。

Immunofluorescence chromatographic quantitative detection technology is an emerging technique of the point of care test （POCT）. Based on the previous work of immunofluorescence quantitative chromatography detector, the wavelet algorithm was used to denoise the fluorescence signal after the A/D sampling in this paper. Then the baseline drift was removed by the median filter method. Finally, the peak height and the area were sellected as the characteristic value for the quantitative detection. The experimental data shown that the SNR improved 10dB and the SRMS reduced by 30.4% after denoising. In the quantitatively detect step, the CV were 5.1% and 0.4% , and the test error were 9.1% and 0.9% , for the peak hight and area characteristic value, respectively. The results show that the peak area of the test line as the characteristic value has better accuracy and repeatability than the peak height.