一种基于Java平台的近红外光谱实时分析系统设计与开发  

DESIGN AND IMPLEMENTATION OF A NIRS REAL-TIME ANALYSIS SYSTEM BASED ON JAVA PLATFORM

作　　者：林晏清 胡澍芃 刘晶[1] 范爽[1,2] 徐琢频 王琦 赵莉萍[1,3] 吴跃进 

机构地区：[1]中国科学院技术生物与农业工程研究所,安徽合肥230031 [2]中国科学技术大学,安徽合肥230026 [3]曼彻斯特大学计算机科学学院

出　　处：《计算机应用与软件》2017年第12期8-12,106,共6页Computer Applications and Software

基　　金：国家自然科学青年基金项目(31500300);中国科学院战略性科技先导专项(A类)(XDA0804107);合肥研究院院长基金青年"火花"项目(YZJJ201520);院企重大攻关项目(Y19HX14702)

摘　　要：针对某肥料生产企业产品质量评价方式耗时、低效等问题,构建基于Java平台的近红外光谱实时分析系统。结合实际生产需求与近红外光谱技术的原理,对系统整体架构进行设计,并通过需求分析、系统设计、关键技术实现等过程,开发出近红外光谱实时分析系统,进行产品成分预测。可实现光谱数据采集、计算分析和结果实时输出功能,用户界面简洁直观。通过硅胶的吸水性实验,验证了预测分析结果的准确可靠(产品成分预测值对真实值的决定系数为95.13),满足生产要求,具备较大应用前景和商业潜力。In order to solve the problem of time consuming and low efficiency of product quality evaluation method in a fertilizer production enterprise, a real-time analysis system of NIRS （near-infrared reflectance spectroscopy） based on Java platform is proposed. The functions of the system were designed based on production demand and the theory of NIRS. And we designed the overall architecture of the system, and through requirements analysis, system design, implementation of key technologies, etc. We developed a real-time analysis system of NIRS, and then carried out product composition prediction. Designed to implement concentration prediction, this system was able to achieve spectrum acquisition, analysis and results output real-timely. The user interface is low-complexity and intuitive. The experimental results of water absorption of silica gel show that the analysis result is accurate and reliable （ the coefficient of determination of the predicted value of the product composition to the real value is 95. 13 ） , which can meet the requirement of production and has great application prospect and commercial potential.

关 键 词：产品成分 实时分析 JAVA平台 近红外光谱 系统设计 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]