【摘要】 为了定量地描述“光电液滴指纹图”鉴别液体的功能,提出一种波形分析法来提取指纹图的特征参数。采用“邻域比较法”或“基于最大值和最小值的极值检测法”确定指纹图的峰谷位置,分别计算表征主峰、次峰和波谷的电容信号和光纤信号的参数,以及指纹图的曲线长度和曲线面积。实验证明,“曲线面积”是最有效的识别参数,不同种类液体的相对分辨率为7.45%;同一种类不同品牌的水、饮料、酒、醋和酱油的相对分辨率分别为1.96%,7.64%,14.39%,0.07%和3.65%。指纹图的“次峰”是相对较弱的特征,相对分辨率仅为0.07%。更多还原

【Abstract】 In order to quantitatively describe the liquid identification function of the Opto-electronic Liquid Drop Fingerprint (OLDF), a method for extracting features of fingerprint based on waveform analysis is proposed. The peak and valley positions of fingerprint are determined by using neighborhood comparison method or extreme detection based on maximum and minimum. The parameters of capacitance signal and fiber signal for characterizing main peak, secondary peak and valley, curve length and curve area of fingerprint, are calculated, respectively. Experiments demonstrate that curve area is the most effective identification parameter and the relative identification ratio of different kinds of liquid is 7.45%. The relative identification ratio of different brands of water, drink, wine, vinegar and soy are 1.96%, 7.64%, 14.39%, 0.07% and 3.65%, respectively. The secondary peak of fingerprint has relative weak feature with relative identification ratio only 0.07%.更多还原