动态光谱幅值检测中的奇异值剔除  

Processing of singularity in detection of dynamic spectrum

作　　者：王焱[1] 吕春玲[1] 马勇[2] 李刚[3] 曾锐利[4] 

机构地区：[1]辽宁工程技术大学电气与控制工程学院,辽宁葫芦岛125105 [2]辽宁工程技术大学理学院,辽宁阜新123000 [3]天津大学精密测试技术及仪器国家重点实验室,天津300072 [4]军事交通学院汽车工程系,天津300161

出　　处：《辽宁工程技术大学学报（自然科学版）》2009年第2期265-268,共4页Journal of Liaoning Technical University (Natural Science)

基　　金：国家自然科学基金资助项目(60174032;60674111);辽宁省博士科研启动基金资助项目(20071099)

摘　　要：为了解决动态光谱检测技术中单波长波形变异引入的误差,提出了基于归一化光谱的标准差奇异值剔除方法,有效克服了单波长信号稳定性判别方法无法判断光谱变异的缺点,动态光谱幅值提取自光电容积脉搏波脉动幅值,可用于人体血液成分检测,而单一波长脉搏波信号的畸变通常是由检测设备因素造成的,与被测个体的生理信息无关。动态光谱信号有效性需以脉搏波基波相对幅值稳定性作为判据,即在对信号进行归一化处理后,再判断其变异程度,以标准差作为判别手段。实验结果表明,该方法能够有效剔出由单波长波形变异引入的误差,有效保障了动态光谱的检测精度。In order to get rid of the error introduced by the variances of single wavelength wave, a kind of standard deviation processing of singularity based on unitary pulse wave is presented. Dynamic spectrum can be extracted fi＇om the pulsatile part of photo-electric pulse wave and can be used to detect the concentration of the component of human blood. The variances of single wavelength wave are always imported by the detect instrument and are independent of the healthy information. The validity of dynamic spectrum should be evaluated based on the relative amplitude of the fundamental of unitary pulse wave. Standard deviation is employed in the presented method. Experimental result shows that the error introduced by the variances of single wavelength wave is eliminated efficiently, thus the accuracy of dynamic spectrum is ensured.

关 键 词：动态光谱 奇异值 归一化脉搏波 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]