改进ALPSO-SVM的精密加工结构件图像分类算法

Improved ALPSO-SVM Precision Machining of Structural Parts Image Classification Algorithm

针对精密加工结构件在加工过程中人工分拣效率低的问题,提出一种改进PSO对于SVM参数优化的图像分类算法(Adaptive Levy Flight Particle Swarm Optimization Algorithm-Support Vector Machine,ALPSO-SVM)。首先,建立一个SVM分类器用来完成图像分类,然后使用改进的PSO算法(ALPSO)来帮助SVM找到最优参数,以提高分类效果;通过引入自适应步长的Levy飞行算法来调整PSO算法中粒子的位置更新公式,提高PSO的寻优效率,以得到SVM的最优参数;最后,SVM以ALPSO输出的结果作为最优参数进行分类测试。实验结果表明,ALPSO有效地提高了SVM的分类性能,分类正确率与鲁棒性都得到提高,平均正确率达到90%以上。

Adaptive Levy Flight Particle swarm optimization algorithm algorithm-support Vector Machine(ALPSO-SVM)was proposed to solve the problem of low manual sorting efficiency in the process of precision machining structural parts.First,a SVM classifier was established to complete image classification,and then an improved PSO algorithm(ALPSO)was used to help SVM find the optimal parameters to improve the classification effect.The adaptive step Levy flight algorithm was introduced to adjust the particle position updating formula in the PSO algorithm,improve the PSO optimization efficiency,and obtain the optimal pa-rameters of SVM.Finally,the SVM takes the result of ALPSO as the optimal parameter for classification test.The experimental re-sults show that ALPSO effectively improves the classification performance of SVM,the classification accuracy and robustness are improved,and the average accuracy is over 90%.