基于视频动态人物自适应标定技术

Adaptive calibration technique for dynamic character based on surveillance video

为了提高摄像机标定的自适应性和应用灵活性,运用基本光学成像原理,提出一种基于监控视频动态人物自适应标定方法。该方法结合Faugeras标定方法以及卫星定位原理,将经过背景消除的双目视频中人物进行角点检测和轮廓提取,并对双目视图以及相邻帧频图像中对应角点进行匹配,代入三维自适应标定公式,形成摄像机标定参数计算环域,得出视觉传感器的内外参数,使用线性最小二乘法求解畸变参数,并对计算结果进行全局非线性优化,得到最优摄像机内外部参数。最后,根据采集的动态人物特征点像素信息,对整体参数相对测量误差与张正友标定算法进行了对比分析,并求取动态人物的高度值。该方法不需要任何已知的摄像机参数,求解简单方便。实验结果表明该方法具有较高的标定精度和鲁棒性。

In order to improve the adaptability and flexibility of camera calibration,an adaptive calibration method for dynamic character based on video surveillance using the basic principle of pinhole optical imaging.The method combines Faugeras calibration algorithm and satellite positioning principle to perform corner detection and contour point extraction for the character in binocular video after background elimination.Then,the corresponding comer points in the binocular image and adjacent frame images are matched and substituted into the three-dimensional coordinate adaptive calibration formula to form the calculation ring area of the camera parameter calibration and obtain the inside and outside parameters of the visual sensor; the method of linear least squares is used to solve the distortion parameters; global nonlinear optimization is performed on the calculated results and the optimized internal and external parameters of the camera are obtained.Finally,according to the pixel information of the characteristic points of the dynamic character,Zhang Zhengyou calibration algorithm and the proposed method were compared,their relative RMS measurement errors of the overall parameters were analyzed,and the height of the dynamic character was solved.This method does not require knowing any camera parameters,and is simple and convenient.Experiment results show that the method has high calibration accuracy and robustness.