基于改进万有引力优化的LSSVM模型在标签缺陷检测中的应用.docVIP

基于改进万有引力优化的LSSVM模型在标签缺陷检测中的应用 庄葛巍1，张晓颖1，张维平2，高大智2 国网上海市电力公司电力科学研究院，上海 200051；2. 河北海纳电测仪器股份有限公司，河北 秦皇岛066004） 摘要针对最小二乘支持向量机LSSVM）在缺陷检测过程中的模型参数选择问题，提出了一种改进的万有引力搜索算法（IGSA）对模型参数进行优化，该算法有效地克服了标准GSA易陷入局部最优解且优化精度不高的缺点，显著提高了原算法中物体的探索能力与开发能力。通过利用UCI 数据库的数据进行分类验证，相比交叉验证、标准GSA、遗传和粒子群优化的LSSVM， IGSA-LSSVM分类模型有效提高了分类正确率和模型的泛化能力。最后，把该模型应用于标签缺陷自动检测中，取得了良好的效果。 关键词：万有引力搜索算法；最小二乘支持向量机；分类模型；缺陷检测 中图分类号：TM93 文献标识码：B 文章编号：1001-1390（201）00-0000-00 LSSVM model optimized by improved gravitation search algorithm and its application on label defects?detecting? Zhuang Gewei1, Zhang Xiaoying1, Zhang Weiping2, Gao Dazhi2 (1. Electric Power Research Institute, State Grid Shanghai Municipal Electric Power Company, Shanghai 20051, China. 2. Haina Power Measuring Instruments Co., Ltd., Qinhuangdao 066004, Hebei, China) Abstract: In the light of the problems existed in selecting the parameters of LSSVM model in the process of defect detection, the Improved Gravitational Search Algorithm (IGSA) is brought in and applied to optimize the model parameters of LSSVM. The algorithm overcomes the shortcoming of standard GSA that is easy to fall into local optimum and has low accuracy and effectively improves the exploration ability and development ability of GSA. Experiments are carried out classification validation on the data sets from the UCI database. Compared with cross-validation, standard GSA, LSSVM of genetic algorithm and particle swarm optimization, the classification model of IGSA-LSSVM has the better classification accuracy and generalization ability. Finally, this model is applied to the label defect automatic detection, and has obtained a good result. Keywords: gravitational search algorithm, least squares vector machine, classification model, defect detection 0 引 言 低压电流互感器自动化检定线最后一个环节是利用自动贴标机对检定结果合格的产品粘贴合格证标签。然而在粘贴过程中容易出现贴错标签、上下颠倒、打印文字不清或缺失、甚至是粘贴不上等缺陷，在回库之前必须要剔除。与此同时，许多行业的产品或外包装自动贴标时都存在标签缺陷检测的问题。然而，目前国内对此问题的处理还主要依靠人工目测，不仅效率低还常常出现误检和漏检，因此亟需找到更高效和可靠的检测方法。