控制点识别的全自动相机标定算法研究  

Study of automatic camera calibration based on control point identification

作　　者：孙韬[1] 方俊永 李莹 

机构地区：[1]中国科学院电子学研究所/空间信息处理与应用系统技术重点实验室,北京100190 [2]中国科学院遥感与数字地球研究所,北京100101 [3]北京博科鸿图信息技术有限公司,北京102209

出　　处：《测绘科学》2017年第10期32-37,共6页Science of Surveying and Mapping

基　　金：中国科学院科技服务网络计划项目(KFJ-EW-STS-046)

摘　　要：针对现有相机标定环节中控制点需要手动选取,标定自动化程度不高,效率较低的问题。该文在圆心投影点精确定位研究成果的基础上,提出了一种基于角距的控制点自动识别算法,解决了三维高精度控制场下全自动相机标定的关键环节。初步实验表明,在同等条件下,基于该方法的全自动标定精度、效率和稳定性均优于传统人工选取控制点的标定方法。In order to improve the degree of automation and effectiveness of camera calibration, this paper proposed an angular-distance based automatic control point identification method on the basis of the approaches of accurate location of projected circular center. This method avoided manually identification of control points, and solved the key problem of camera automatic calibration under the 3D control field. The preliminary experiment demonstrated that, under the same condition, the accuracy, effectiveness and robust of camera calibration based on our method were greatly better than those based on manually identification of control points.

关 键 词：全自动相机标定 摄影测量 圆心定位 角距 控制点识别 

分 类 号：P208[天文地球—地图制图学与地理信息工程]