基于麻雀搜索算法优化支持向量机的瓶盖装配检测研究

浏览次数： 12

分享到:

摘要：

针对基于支持向量机的瓶盖装配检测算法准确度不高、调参难度大的问题,提出通过麻雀搜索算法(Sparrow Search Algorithm,SSA)对支持向量机(Support Vector Machines,SVM)的关键参数寻找最优解.采集瓶盖部位图像,包括标准、歪斜、铝塑分离、胶塞缺失、高盖5种类型.提取6个典型特征构建数据集,采用二分类支持向量机分类,分别通过遗传算法、粒子群算法和麻雀搜索算法对支持向量机参数进行调节.训练结果表明,麻雀搜索算法优化后的支持向量机模型测试准确率达到98.33%,高于其他几种算法.基于SSA-SVM的瓶盖装配检测模型识别精度高,调参速度快,泛化能力强.

Aiming at the low accuracy of the vial cap assembly detection algorithm,based on support vector machine and the difficulty of parameter adjustment,a sparrow search algorithm(SSA)was proposed to find the optimal solution to the key parameters of supporting vector machines(SVM).Images of bottle cap parts were collected,including standard,skew,aluminum-plastic separation,missing glue plug and high cap.Six typical features were extracted to construct a data set.The two-class support vector machine was used for classification.The parameters of supporting vector machine were adjusted by genetic algorithm,particle swarm algorithm and sparrow search algorithm.The training results show that the test accuracy of the SVM model optimized by the sparrow search algorithm reaches 98.33%,which is higher than that of other algorithms.The SSA-SVM-based vial cap assembly detection model has high recognition accuracy,fast parameter adjustment and strong generalization ability.

作者：

张冬至韩栋星毛瑞源郗广帅

Zhang Dongzhi;Han Dongxing;Mao Ruiyuan;Xi Guangshuai(College of Control Science and Engineering,China University of Petroleum,Qingdao 266580,China)

机构地区：

中国石油大学（华东）控制科学与工程学院

出处：

《betway官方app 学报：自然科学版》 CAS 北大核心 2023年第1期29-38,F0002,共11页

Journal of Henan Normal University(Natural Science Edition)

基金：

国家自然科学基金(51407200).

关键词：

瓶盖装配检测机器视觉图像处理支持向量机(SVM) 麻雀搜索算法(SSA)