摘要
针对荧光检测技术中荧光光谱重叠引起荧光值偏差的问题,提出了一种基于偏差递推最小二乘算法辨识补偿矩阵的方法。首先,在多输入多输出系统(MIMO)下,利用单染色荧光光谱和多染色荧光光谱实际测量的荧光值,通过递推最小二乘法进行迭代运算,推导出参数估计值。然后,在其中引入一个修正项,补偿在采集荧光中过程噪声引起的误差。最后,计算出偏差补偿递推最小二乘法迭代的估计值。理论分析与仿真表明,该算法在参数误差估计中误差率小于1%,相对于递推最小二乘算法性能提高了50%。所用算法能够有效提高估计值精度,同时也能减小噪声产生的影响。
Focused on the intensity deviation issue caused by spectral overlap in fluorescence detection technology, the bias compensation recursive last squares method for compensation matrix was proposed. Firstly, based on the measured single and multiple staining fluorescence values, the parameter estimates were deduced through recursive least squares method in the multi-input multi-output system（MIMO）. Secondly, by introducing a correction term into the estimated values, the errors caused by noise in the fluorescence acquisition process were compensated. Finally, the estimates was calculated iteratively with bias compensation recursive least squares. The simulation results and theoretical analysis show that with this method, the error rate is less than 1% and the performance is improved by50% compared with the recursive last squares algorithm. The proposed method can effectively improve the accuracy of estimates, meanwhile reduce the negative effect of noise.
出处
《价值工程》
2017年第6期114-117,共4页
Value Engineering
基金
广西自然科学基金项目(2015GXNSFAA139299)
关键词
荧光检测
偏差补偿递推最小二乘
荧光补偿
参数辨识
fluorescence detection
bias compensation recursive least squares
fluorescence compensation
parameter identification
