基于Fisher判别分析的轮毂识别研究

Research on Hub Recognition Based on Fisher’s Discriminant Analysis

针对传统轮毂生产线因人工目检带来的识别问题,提出以机器视觉为识别基础、Fisher判别为分类方法对轮毂进行识别分类研究。在对机器视觉系统获取轮毂图像进行预处理后,对图像进行特征提取操作获得轮毂半径、轮毂辐条数、辐条类型、轮毂宽度的特征数据。进而用Fisher判别法对轮毂样本数据库进行学习分类。由检测结果得知,Fisher判别法对轮毂分类具有较好的效果,且方法简单,具有较高的识别率。

In view of the recognition problem caused by manual eye inspection in traditional wheel hub production line,this paper puts forward the research on recognition and classification of wheel hub based on machine vision and Fisher’s discrimination.After preprocessing the hub image acquired by the machine vision system,image feature extraction operation is carried out to obtain the feature data of hub radius,hub spokes number,spoke type and hub width.Furthermore,Fisher’s test is used to study the hub sample database and classify the unrecognized samples.It is known from the test results that Fisher’s discriminant method has a good effect on classification of hub detection,and the method is simple and has a high recognition rate.