摘要
结合立体相机与计算机等硬件设备,本文设计并实现了一套双目视觉测量原型系统。该系统首先利用张正友标定法解算出相机的内部参数及其相对位置关系;然后实时纠正所获取的目标场景立体影像,用半全局匹配(SGM)方法对立体影像进行逐像素匹配,计算目标场景的深度图和三维点云坐标;最后实验验证了该系统的可行性,测试了立体影像获取与目标重建的速度。结果表明,该系统具有一定的实时性与较强的可靠性。
A set of binocular vision measurement prototype system is designed and developed combining stereo cameras with computer and other hardware equipment. Firstly, the Zhang Zhengyou calibration method is used to calculate the intrinsic param- eters of the cameras and their relative position relations. Secondly, stereo images of the target scene are corrected in real time, the SGM method is used to match the images pixel by pixel, and the depth map of target scene and 3D point cloud coordinate are calculated. Finally, an experiment is conducted to verify the feasibility of the system and the speed of stereo image obtaining and target reconstruction is tested. The result shows that the system possesses good reliability.
出处
《测绘科学与工程》
2016年第2期58-63,78,共7页
Geomatics Science and Engineering
基金
国家自然科学基金资助项目(41371436),国家重点基础研究发展计划(973计划)资助项目(2012CB720000).
关键词
双目视觉
标定
半全局匹配
核线影像
实时测量
stereo vision
calibration
semi - global matching
epipolar image
real - time measurement