利用光声技术对血糖浓度无损检测进行了初探，搭建了一套可调谐脉冲激光器激发、聚焦超声探测的光声无损血糖检测系统，采用前向探测模式、前置预放大、GPIB-USB 总线传输和LabVIEW 软件开发平台，实现了葡萄糖水溶液的光声信号实时采集。实验得到了质量浓度为0~300 mg/dL 的葡萄糖水溶液的实时光声信号，并在1300~2300 nm 近红外波段内对不同浓度的葡萄糖溶液进行波长间隔10 nm 的连续扫描激发，得到不同浓度在不同激发波长下的光声峰峰值。利用差分和一阶导数谱方式得到葡萄糖特征吸收波长，对优选的多个特征波长利用最小二乘拟合算法建立了浓度和光声峰峰值数学校正模型，并对多个特征波长下建立的校正模型进行浓度反演对比，以预测均方根误差最小为标准，筛选出两个浓度预测结果相对较好的特征波长。

A preliminary research of noninvasive monitoring of blood glucose concentration is performed using the photoacoustic detection technique, and a noninvasive photoacoustic detection system is established. A tunable pulsed laser is used as the excitation source and a confocal ultrasonic transducer is used to capture the photoacoustic signal of glucose concentrations. The pre-amplifier, data transmission based on GPIB-USB bus and virtual instruments developing software LabVIEW are used in this system. The time-resolved photoacoustic signals for the glucose solutions with concentration from 0 mg/dL to 300 mg/dL are acquired based on the established photoacoustic detection system. The glucose solutions with different concentrations are scanned by the tunable pulsed laser with output wavelength from 1300 nm to 2300 nm and with interval of 10 nm, and the photoacoustic peak-to-peak values are obtained. In order to obtain the characteristic wavelengths of the glucose solution, the difference spectrum algorithm and the first order differential derivative are used. The mathematical correction model between glucose concentrations and photoacoustic peak-to-peak values is established by means of linear least square fitting algorithm for the optimized characteristic wavelengths, and the glucose concentrations are inverted. Two better characteristic wavelengths are chosen based on the minimum root mean square error.