中文摘要：

荧光分析法具有灵敏度高、选择性好、易于设计等优点,是检测水中油类污染物的重要手段。光电探测器产生的噪声会影响荧光检测系统的灵敏度,荧光信号的噪声消除一直是研究的热点问题。由于荧光信号增加了支集长度,dbN族小波能够解决信号的边界问题,通过比较dbN族不同小波基的去噪效果,选择db7为最优小波基,对含噪荧光信号作5层静态小波分解。根据小波熵理论自适应地选择阈值,高频系数经过阈值量化并重构得到纯净的荧光信号。与离散小波变换相比,静态小波变换去噪后信号具有信息完整性和时移不变性。

英文摘要：

Fluorescence analysis is an important means of detecting mineral oil in water pollutants because of high sensitivity,selectivity,ease of design,etc.Noise generated from Photo detector will affect the sensitivity of fluorescence detection system,so the elimination of fluorescence signal noise has been a hot issue.For the fluorescence signal,due to the length increase of the branch set,it produces some boundary issues.The dbN wavelet family can flexibly balance the border issues,retain the useful signals and get rid of noise,the de-noising effects of dbN families are compared,the db7 wavelet is chosen as the optimal wavelet.The noisy fluorescence signal is statically decomposed into 5levels via db7 wavelet,and the thresholds are chosen adaptively based on the wavelet entropy theory.The pure fluorescence signal is obtained after the approximation coefficients and detail coefficients quantified by thresholds reconstructed.Compared with the DWT,the signal de-noised via SWT has the advantage of information integrity and time translation invariance.