基于贝叶斯分类与角点匹配的工件定位算法研究

Workpiece Localization Algorithm Based on Bayesian Classification and Corner Matching

针对工件环境复杂、光照干扰造成机器视觉定位精度低的问题,提出了基于贝叶斯分类与角点匹配工件定位算法。首先,采集工件图像,对图像进行特征分析,引入角点检测算子,建立角点配准方程,完成待测工件与标准工件的角点配准,为工件定位做好准备。然后,基于贝叶斯分类器,对待测目标进行识别确认,根据贝叶斯分类器来确定目标评分,以准确定位工件。最后,基于C#语言与.NET平台开发定位系统软件,借助底层语言实现算法,由软件系统调用算法模块,实现算法的性能测试。实验测试结果表明,本文算法具有理想的定位准确性,可以实现大批量工件定位,为机器视觉设备的实际应用提供技术基础。

Aiming at low accuracy of machine vision positioning,this paper proposes a workpiece positioning system based on Bayesian classification and corner matching.Firstly,the image of the workpiece is collected and analyzed and the corner detection operator is introduced as well as the corner registration equation is established.Then,based on the Bayesian classifier,the target is confirmed,and the target score is determined and located accurately.Finally,the positioning system software is developed based on C#language and.NET platform to test the possibilities of commercial application and technical confidentiality.The experimental results show that the system is conducive to the positioning of large quantities of workpieces and lays a foundation for the practical application of machine vision equipment.