区间值直觉模糊麦克劳林聚合算子的近红外反射波段划分方法  

Near Infrared Reflection Band Division Method Based on Interval-Valued Intuitionistic Fuzzy Maclaurin Symmetric Mean Aggregation Operator

作　　者：任维佳 杜向军 杜玉琴 孙荣禄[1,3] 李学良 REN Wei-jia;DU Xiang-jun;DU Yu-qin;SUN Rong-lu;LI Xue-liang(School of Mechanical Engineering,Tiangong University,Tianjin 300387,China;College of Mechanical Engineering,Tianjin University of Science and Technology,Tianjin 300222,China;Tianjin Key Laboratory of Advanced Mechatronics Equipment Technology,Tiangong University,Tianjin 300387,China;School of Economics,University of Chinese Academy of Social Sciences,Beijing 102401,China)

机构地区：[1]天津工业大学机械工程学院,天津300387 [2]天津科技大学机械工程学院,天津300222 [3]天津市现代机电装备技术重点实验室,天津300387 [4]中国社会科学院大学经济学院,北京102401

出　　处：《光谱学与光谱分析》2024年第6期1731-1739,共9页Spectroscopy and Spectral Analysis

基　　金：国家自然科学基金青年科学基金项目(51205288);天津市自然科学基金项目(17JCYBJC19400);天津市研究生科研创新项目(2022BKY140)资助。

摘　　要：采用不同波段光谱对不同类异性纤维进行识别,可有效剔除异性纤维,提高检出率。针对波段传统划分方法存在的评价指标单一性问题,研究不同异性纤维多种属性间相互影响,结合多属性群决策(MAGDM)方法的优势,提出利用多属性群决策方法实现棉花中异性纤维最佳检测波段的选取,根据不同异性纤维多种属性指标的关联性,综合考虑多种属性指标,确定波间可分度、相关性和ABS指标作为异性纤维属性的评价指标。首先,针对多属性群决策方法中评价准则不准确问题,构建评价准则函数线性方程组,使其增广矩阵的秩等于未知量个数,保证方程组具有唯一解,从而提高决策结果的准确性。接着,使用幂等(PA)算子消除不合理评价信息数据值对决策结果的负面影响,结合麦克劳林对称平均(MSM)算子综合考量多个数据间的相互关系,推导加权区间直觉模糊幂等麦克劳林对称平均(WIVIFPMSM)聚合算子,拓展TOPSIS方法确定异性纤维权重信息,聚合不同异性纤维多种属性的评价信息,根据所建立的评价准则确定决策结果,进而构建基于区间值直觉模糊集(IVIFS)的多属性决策方法,实现多种异性纤维多种属性指标的最佳波段选取。将区间值直觉模糊幂等麦克劳林聚合算子与类间可分性波段选择(ISBC)方法和自适应选择(ABS)方法进行了对比,分析了不同波段划分方法对结果的影响,总结了现有研究中存在的问题和不足。为提高多属性群决策方法的决策精度,分析了参数k对决策结果的影响,证明区间值直觉模糊幂等麦克劳林聚合方法稳定性更好,为复杂环境下异性纤维波段划分研究提供了新思路。最后,通过实验验证近红外波段W 3:780~1100 nm为最佳检测波段。该方法对波段选取领域的理论扩展和多属性群决策方法的应用具有一定的指导意义。Using different wavelength spectra to identify different types of foreign fibers can effectively eliminate foreign fibers and increase the detection rate.Given the problem of a single evaluation index existing in traditional band division methods,the interaction of attribute indexes of different fibers is studied,combined with the advantages of the multi-attribute group decision-making(MAGDM)method,this paper proposes to use the MAGDM method to realize the selection of the optimal detection band for foreign fibers in cotton.According to the relation of attribute indexes of different fibers,the inter-class separability,correlation and ABS index are determined as attribute evaluation indexes.Firstly,to solve the problem of inaccurate evaluation criteria in the MAGDM method,a system of evaluation criteria function linear equations is constructed so that the rank of the augmented matrix is equal to the number of unknowns,ensuring that the equation system has a unique solution,thereby improving the accuracy of the decision result.Next,the power mean(PA)operator is used to eliminate the adverse effects of unreasonable evaluation information values on decision results,combined with the Maclaurin symmetric mean(MSM)operator to comprehensively consider the relationship between input arguments,deriving the weighted interval-valued intuitionistic fuzzy power Maclaurin symmetric mean(WIVIFPMSM)aggregation operator.Then,the TOPSIS method is used to determine the weight information of foreign fibers,the evaluation information of various attributes of different foreign fibers is aggregated,and the decision results are chosenaccording to the established evaluation criteria.Thus,a MAGDM method based on interval-valued intuitionistic fuzzy sets(IVIFSs)is constructed to realize the optimal band selection of various attributes of foreign fibers.Moreover,the WIVIFPMSM aggregation operator is compared with the inter-class separability band selection(ISBC)method and adaptive band selection(ABS)method,the influence of different band div

关 键 词：异性纤维 反射率光谱 WIVIFPMSM聚合算子 多属性决策 波段选择 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]