“最小二乘法变点冰水情数据处理算法”的研究与应用  

Research and Application of Ice Conditions Change Point Least Squares Algorithm

作　　者：陈宁[1] 秦建敏[1] 李国宏[1] 刘永强[1] 张宝森[2] 

机构地区：[1]新型传感器与智能控制教育部重点实验室,测控技术研究所太原理工大学,山西太原030024 [2]防汛抢险研究所黄河水利委员会黄河水利科学研究院,河南郑州450002

出　　处：《数学的实践与认识》2012年第1期108-114,共7页Mathematics in Practice and Theory

基　　金：2009年度高等学校博士学科点专项科研基金(20091402110004);2011年水利部科技推广计划项目(T91115)

摘　　要：将非线性统计学中的变点分析方法应用于基于空气、冰与水的物理特性差异实现冰水情自动检测过程的现场数据处理中,提出了利用被测物质物理参数的动态变点替代经验阈值来确定空气、冰与水界面位置的"最小二乘法变点冰水情数据处理算法",较好地解决了原始采样数据中奇异值引起误判的工程难题,提高了冰情数据分析的准确度.利用该算法对黄河河道现场冰水情采集数据进行分析,得到了较准确的结果,实现了对冰层生消全过程的自动监测.Used the change-point analysis, a widely applied technique in non-linear statistics, into the data processing for automatic detection of ice water field. Using measured material parameters＇ change point substitute for experience threshold to determine the interface of ice and water, ice and air, which is named Ice conditions change point least squares algorithm. It can solve the engineering problem that is easily leading to false result caused by the singular values of raw data, which greatly improved the accuracy of data analysis. Least squares method is used to analyze the data from the Putanguai test sites of Yellow River. It can get more accurate data and can achieve the automatic monitoring of the entire process of the ice＇s generation and disappearance.

关 键 词：冰水情数据处理 最小二乘法 变点分析 冰层厚度 

分 类 号：O241.5[理学—计算数学] P332.8[理学—数学]