海水痕量重金属元素现场自动识别技术  

In-situ auto-analysis technique of trace heavy metal in seawater

作　　者：王银瓶[1] 潘跃峰[1] 李毅[1] 邹绍芳[1] 胡卫军[1] 王平[1] 

机构地区：[1]浙江大学生物传感器国家专业实验室,生物医学工程教育部重点实验室,浙江杭州310027

出　　处：《浙江大学学报（工学版）》2007年第2期230-235,共6页Journal of Zhejiang University：Engineering Science

基　　金：国家"863"高技术研究发展计划资助项目(2001AA635060)

摘　　要：针对自行研制的海水重金属元素自动分析仪提出了一种快速的数据自动识别算法.基于实测数据提出并比较了多种方法.算法采用分位数法剔除实测数据的大误差;利用有限脉冲响应(FIR)滤波器进行数字滤波,采用窗函数进行平滑滤波,很好地去除了信号中的小噪声;通过对3种去基线方法的比较,采用切线识别法进行基线校正;利用标准曲线法和神经网络法对数据进行自动识别,克服了不同金属间互化物对浓度的影响.实验结果表明,该算法适用于海水重金属元素的现场快速分析和自动识别,取得了令人满意的效果.A novel algorithm of fast and automatic data processing and identification utilized in the ship equipped auto-analysis instrument of heavy metal in seawater was put forward. After analysis of practical data from sensors, some improved arithmetics were proposed and compared. The algorithm used tetrasectional quantile for outlier detection and correction, and finite impulse response （FIR） digital filter and window function method for data-smoothing. Also tangent identification was used for base-line correction, and neural network method was used for auto-recognition of the heavy metal concentration from peak signal to reduce the effect of intermetallic compound. Each method was developed according to the feature of practical data, therefore the algorithm could get more accurate heavy metal concentration values. The proposed algorithm was proved to be feasible and effective through practical experiments on the auto-analysis instrument.

关 键 词：重金属检测 数据自动识别算法 信号处理 

分 类 号：O657.15[理学—分析化学]