摘要
为了实现视觉引导装配过程中钣金零件图像的识别,对零件图像进行预处理,提取形状特征,将遗传算法的交叉变异操作引入粒子群算法,形成遗传粒子群算法。采用遗传粒子群算法同时进行支持向量机的参数优化和特征选择。实验表明,将所选用特征由初始的12维降维到3维,测试集识别准确率100%,完全满足零件识别分类的要求。
To realize the image recognition of the sheet metal parts in the visual guided assembly process,the image of the part is preprocessed,the shape feature is extracted,and the cross and mutation operation of the genetic algorithm is introduced into the particle swarm optimization,which is used to form genetic particle swarm optimization.This algorithm is used to optimize parameters and select the feature for support vector machines.Experiments show that the selected features are reduced from the initial 12-dimensional dimension to 3 dimensions,and the test set recognition accuracy is 100%.It fully meets the requirements of part identification and classification.
作者
方舟
程筱胜
崔海华
石诚
韦号
FANG Zhou;CHENG Xiaosheng;CUI Haihua;SHI Cheng;WEI Hao(College of Mechanical and Electrical Engineering,Nanjing University of Aeronautics and Astronautics,Nanjing 210016,China)
出处
《机械制造与自动化》
2020年第5期116-118,122,共4页
Machine Building & Automation
关键词
零件识别
支持向量机
粒子群算法
遗传算法
part identification
suppor vector machine
particle swarm optimization
genetic algorithm
