摘要
分析了RGB-D传感器的标定模型及标定方法,利用标定棋盘和Matlab工具箱对RGB-D传感器进行了标定。采用摄像机标定方法计算了彩色摄像机和深度摄像机的初始化参数,根据标定棋盘和标定板的共面关系计算了相对位姿的初始估计值,并采用非线性最小化方法对标定参数进行了优化。实验结果表明:标定后的彩色图像和深度图像的重投影误差较小,深度测量不确定性标准差比标定前约减小了1/2,有效地提高了RGB-D传感器的测量精度。
The calibration model and method for RGB-D sensor are analyzed,the calibration experiment of RGB-D sensor is investigated by using a chessboard and Matlab toolbox.The initial parameters of color camera and depth camera are calculated by the calibrating method of camera.Then the initial relative pose position is estimated by the coplanarity of the calibrating chessboard and calibrating board.Finally, the calibrating parameters are optimized by a nonlinear minimum method.The results of the calibrating experiment show that,the re-projection errors of the color image and the depth image are smaller,the standard error of the depth measurement uncertainty after calibrating is one time smaller than before,and the measurement accuracy of RGB-D sensor is efficiently improved.
出处
《装甲兵工程学院学报》
2015年第1期65-70,共6页
Journal of Academy of Armored Force Engineering