为了解决不同的光学相干层析（OCT）图像预处理方式对皮肤真皮层散射系数计算影响的问题, 提高无创血糖的测量精确度, 提出了一种对OCT图像数据前期处理的最优化方法。在对采集的3维图像数据进行皮肤表面对齐、3维重建、1维平均处理的基础上, 分析了背景噪声和数据是否归一化对血糖预测精度的影响, 并结合临床实验进行了验证。结果表明, 预测误差较预处理之前减小了18.31%。该研究对于提高基于OCT技术的光学无创血糖测量精度具有重要的参考价值。

In order to solve the problem of the influence of different optical coherence tomography (OCT) image preprocessing methods on scattering coefficient calculation in dermis, and to improve the accuracy of noninvasive blood glucose detection, an optimization method of early OCT image data processing was proposed. The 3-D images were first processed by skin surface alignment, 3-D reconstruction and 1-D average, and then the background noise and data normalization effect on the accuracy of blood glucose prediction were analyzed by clinical experiments. The results show that the prediction error decreases by 18.31% compared with that before pretreatment. This study has important reference value for improving the accuracy of optical noninvasive blood glucose detection using OCT.