一种基于神经网络的视觉定位系统

A VISION-LOCATION SYSTEM BASED ON NEURAL NETWORK

摘要: 本文提出了一种基于人工神经网络的多线阵相机系统标定与3D定位方法，并应用于基于多线阵相机构成的视觉空间定位系统．该视觉定位系统利用了线阵相机的快速性与高分辨率的特点，非平行空间投影面相交定位的基本原理，实现了这种结构下快速、高精度空间定位．实验表明，人工神经网络的定位方法简化了多线阵视觉定位系统标定与定位计算的复杂性，在定位精度上达到了良好效果．为机器人位置反馈控制提供了有效的技术途径．

Abstract: A new space location vision system based on one dimensional cameras is presented.High speed and high precision are the features of this position measurement system.In this paper,the measurement methods based on artificial neural networks and the structure of the system are introduced.The experimental results are presented to show the efficiency of this system on position measurement of space targets.