摘要
基于双目立体视觉理论建立的三维测量方法具有非接触、大视场、较高精度以及实时性强等优点,因此在工业众多领域得到广泛的应用。立体匹配是实现双目立体测量的关键技术。为了提高匹配精度和鲁棒性,将一种随机编码结构光投射到被测物体表面,使物体表面具有丰富的纹理信息,从而使立体匹配不受自身表面纹理的影响,而且具有较强的抗外界噪音的能力,满足现场测量的要求。该方法仅需拍摄1次,完成一次三维测量仅需十几秒,因此能够应用到动态场景并且实现快速三维测量。
The 3D measurement method based on binocular stereo vision has many advantages, such as contactlessness, large field of view, high accuracy and real time. It has been applied to many industrial fields. Stereo matching is a key technology in binocular stereo measurement. In this paper, in order to improve matching accuracy and robustness, random pattern is projected onto the object surface to make richly textured, as a result, stereo matching will not be influenced by itself texture and environmental noise in site. In this method, 3D measurement needs just single shot and takes about 10 seconds, so it is rapid and suitable to dynamic scene.
出处
《机电一体化》
2013年第3期61-65,共5页
Mechatronics
关键词
双目立体测量
随机编码
立体匹配
三维测量
binocular stereo measurement random pattern stereo matching 3D measurement
作者简介
张鑫 1988年生，硕士研究生。研究方向为机器视觉与影像测量技术。
习俊通 1963年生，教授，博士生导师。研究方向为数字化产品、精密测量及数字辅助医学工程。
