双目立体视觉中特征点三维坐标重构校准研究

双目立体视觉是一种商业化较成熟的三维测量技术,左右摄像机内外参数的精确标定是实现三维重构的基础和关键。针对棋盘格和圆点两种标定板图案研究了相应的图像处理技术,实现了标定点的亚像素精度定位及其有效排序。采用基于单映性约束和非线性优化的多视角平面标定算法实现了摄像机光学及空间位置参数求解。用极线约束残差法衡量标定结果的准确度。基于桥式三坐标标准实现标定点三维重构平面度以及多平面空间夹角测量结果的校准,基于光学三坐标标准实现了标定点三维重构空间距离的校准,并分析了校准结果的不确定度。圆标志点三维重构空间距离示值误差为0.029 mm,不确定度U=24 μm（k=2）

Research on Calibration of Three Dimensional Coordinate Reconstruction of Feature Points in Binocular Stereo Vision

Binocular stereo vision is one of the 3D measurement techniques which is more adult in commerce. The accurate calibration of inner and out parameters for left and right camera is basic and important in realizing 3D reconstruction.Corresponding image process techniques to two calibration block pattern with chess corners or circle points are researched，so the feature points are located in sub-pixel accuracy and effectly sequenced.The optical and space parameters are calculated by calibration algorithm of homography constraint and multe-view plane.Accuracy of calibration result is evaluated by epipolar constraint residual error.The calibration of measurement results of flatness and angle between planes of feature points 3D reconstruction is completed by standard of bridge CMM.The calibration of space distance of feature points is completed by standard of optical CMM，and the uncertainty is analyzed.The indication error of space distance of circle calibration points 3D reconstruction is 0.029 mm, uncertainty U is 24 μm(k=2).