基于FPCA和PSOSVM回收塑料瓶分类  

Classification of recycled plastic bottles based on FPCA and PSOSVM

作　　者：吴开兴[1,2] 范亭亭 李丽宏[1] 张琳 WU Kai-xing;FAN Ting-ting;LI Li-hong;ZHANG Lin(School of Information and Electrical Engineering,Hebei University of Engineering,Handan 056038,China;Hebei Engineering Laboratory of Coal Mine Information Technology,Hebei University of Engineering,Handan 056038,China)

机构地区：[1]河北工程大学信息与电气工程学院,河北邯郸056038 [2]河北工程大学煤矿综合信息化河北省工程实验室,河北邯郸056038

出　　处：《计算机工程与设计》2018年第11期3555-3558,3575,共5页Computer Engineering and Design

基　　金：河北省教育厅基金项目(ZD2014081);河北省自然科学基金项目(F2015402150)

摘　　要：为提高回收塑料瓶颜色分类的识别率,提出一种基于FPCA和PSOSVM的分类算法。在HSI模型中,使用快速主成分分析(FPCA)法对图像进行降维处理,提取有效的特征;采用粒子群算法(PSO)对支持向量机(SVM)的参数惩罚因子和核函数进行优化;为避免PSO的计算结果陷入局部极值中,引入惯性权重和收敛因子;构建支持向量机分类模型,将优化后的参数和提取的特征作为输入进行分类识别。实验结果表明,该分类算法的识别率为93.4%,较未使用粒子群算法寻优的分类算法,识别率提高了5.8%,可以进行有效识别。To improve the recognition rate of color classification of the recycled plastic bottles,a classification algorithm based on FPCA and PSOSVM was proposed under the HSI model.Fast principal component analysis method(FPCA)was used to reduce dimensions of the image,and the effective features were extracted.Particle swarm optimization(PSO)was adopted to optimize the parameters of penalty factor and kernel function of support vector machine(SVM).The inertial weight and convergence factor were introduced to avoid PSO’s calculation results in local extremum.The classification model of SVM was constructed,and the optimized parameters and extracted characteristics were classified as input.Experimental results show that the recognition rate of the algorithm is 93.4%,which is higher than that of unused particle swarm optimization algorithm,and the recognition rate is increased by 5.8%,which can be used to do effective recognition.

关 键 词：回收塑料瓶分类 快速主成分分析 支持向量机 粒子群算法 参数寻优 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]