摘要
介绍了结构光三维双视觉检测的基本原理 ,提出了标定点的空间坐标和图像坐标的获取方法 ,设计了用于获取标定点的双向光电瞄准装置 ,并建立了双向光电瞄准装置安装误差的补偿方法 ,在此基础上实现了双视觉传感器标定点数据的全局统一。最后 ,利用标定点样本数据建立了结构光三维双视觉 RBF神经网络模型 ,最佳 RBF模型的训练精度为 0 .0 78mm,测试精度为0 .0 84 mm。
The basic principle of 3D double vision inspection based on structured light is firstly introduced. A method of gaining calibration points for 3D double vision inspection system is proposed in detail. In order to gain calibration points with high precision, a double direction photoelectric aiming device is well designed, and a method for compensating the position setting error of the aiming device is also described. The coordinates of all calibration points are precisely unified in world coordinate system. Finally, using the calibration points, the inspection model of 3D double vision based on RBF neural network is successfully established. The training accuracy for the RBF neural network based model is 0.078mm,and its testing accuracy is 0.084mm.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2002年第6期604-607,624,共5页
Chinese Journal of Scientific Instrument
基金
航空科学基金 [991 51 0 0 0 1 ]
北京市科技新星计划 [951 872 0 0 0 ]资助项目
关键词
结构光
三维双视觉检测
RBF神经网络
标定点
Structured light 3D double vision inspection RBF neural network Calibration point Training and testing of network.
