基于共面圆的双目立体视觉分步标定法  

Binocular stereo vision step by step calibration based on coplanar circle

作　　者：王科俊[1] 魏娟[1] 

机构地区：[1]哈尔滨工程大学自动化学院,黑龙江哈尔滨150001

出　　处：《应用科技》2010年第1期36-39,56,共5页Applied Science and Technology

基　　金：国家高技术研究发展计划(863计划)基金资助项目(2006AA04Z248);黑龙江省杰出青年科学基金资助项目(JC200703);哈尔滨市科技创新人才研究专项基金资助项目(2007RFXXG009)

摘　　要：计算机视觉中,在对景物进行定量分析或对物体进行精确定位时,都需要进行摄像机标定,即准确确定摄像机的内外参数.为了快速、有效地进行摄像机的标定,针对常用的带有一阶径向畸变的小孔摄像机模型,提出了一种简单有效的分步标定方法.先用预标定法得到左右摄像机主点坐标参数,然后再用TSAI两步法获得左右摄像机的内外参数,最后以左摄像机光心为世界坐标系原点,通过坐标转换关系,进而得到双目视觉的各种标定参数.该方法实验要求低,不需要移动摄像机.通过实验,验证了该方法能够准确、有效地求出各标定参数.Camera calibration is one of the basic tasks for analyzing the geometry of a scene and determining the shapes and locations of objects in vision. It is important to calculate the internal and external parameters of a camera used when viewing. In order to calibrate camera quickly and efficiently, a simple and effective setp-by-step method of calibration was proposed using a radialorder aberration small camera model. First pre-demarcation was used to get the main point coordinates, and then the internal and external parameters of camera were obtained through two-stage TSAI. Finally, the optical center of left camera was taken as the origin of the world coordinate system, and various parameters were obtained through coordinate transformation relations. This method only requires a coplanar target without camera＇ s motion. Experimental results show that the new approach is very efficient with satisfactory accuracy.

关 键 词：双目视觉 摄像机标定 线性标定 TSAI两步法 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]