基于机器视觉的磁铁装配系统

Magnet Assembly System Based on Machine Vision

针对自动化装配中常见的圆形物体定位检测过程中原始图像成像质量不高、感兴趣区域边缘模糊和位姿求取精度低的问题,采用形态学处理提高图像质量,引入同态增晰算法增强边缘特征,用最小二乘法Huber圆拟合算法,用梯度下降方法求出圆心,完成高精度定位。通过现场实际验证测试和数据分析,该算法运行速度快,定位精度高,且运行稳定。

In order to solve the problems of low quality of original image,blurred edge of region of interest and low precision of pose calculation in the process of automatic assembly,morphological processing was used to improve image quality,homomorphic enhancement algorithm was introduced to enhance edge features,and minimum square Huber circle fitting algorithm was used to calculate the center of circle with gradient method to complete high-precision positioning.Through the actual verification test and data analysis,the algorithm runs fast,positioning accuracy is high,and the operation is stable.