目标跟踪与定位中的视觉标定算法研究

目标跟踪与定位中的一个重要步骤就是摄像机标定，其目的是估计摄像机的外部和内部参数。在严格的几何约束关系之上建立准确的数学模型，提出一种快速得到摄像机中心的方法，然后通过合理的求解次序获得其他摄像机参数，保证了标定参数的精度。利用计算出来的标定参数校正失真图像中的各个像素位置以重新得到像素间原来的空间关系，从而产生精确的不失真图像，并且利用摄像机标定参数对标定模板上的点进行位置计算后，再和实际位置比较进行检验。图像的校正效果实验以及精度验算的结果表明：提出的算法得到了准确可靠的标定参数，和其他算法相比有效提高了精度，能够满足运动跟踪时目标特征位置估计的精度要求。

Vision calibration algorithm for object tracking and positioning

An important task for object tracking and positioning is camera calibration, whose purpose is to estimate the internal and external parameters of cameras. A exact mathematical model, which could provide a fast method to obtain the camera centre, was established based on a strict geometric constraint relationship. The other parameters of the camera was acquired in a resonable resolving sequence to guarantee the parameter accuracy of the camera. The pixel position in a distorted image is corrected based on the computed calibration parameters to recover the original space relationship between pixels, then the accurate and undistorted images are obtain. After the position of the spot on the calibration template is computed based on camera calibration parameters, the computed spot position is compared with its real position for verifing its accuracy. The experiments to check the image correction effect and accuracy indicates that the calibration parameters obtained by the algorithm are accrate and reliable, the accuracy is effectively improved, and the accuracy requirement for the position estimation of the target in moving-target tracking can be satisfied.