论文部分内容阅读
针对室外环境下的PTZ(pan-tilt-zoom)摄像机,提出了一种全自动运行且精度较高的摄像机标定算法,并将其成功地应用于视频监控系统下人体的身高测量中.具体而言,首先将摄像机模型分解为变焦转换模型与定焦摄像机参数模型.对于前者,提出了一种基于多参考点的变焦转换模型标定方法.对于后者,利用摄像机旋转角度信息,对定焦摄像机模型进行了在线标定.由于将变焦中心与主点位置作为相互独立的变量,在变焦转换模型与定焦模型中分别进行标定,因此确保了模型的完整性与准确性;引入了多个参考点,使得本文方法较好地解决了多倍变焦情况下模型误差较大与特征匹配点过少等问题,提高了标定的精度.基于自标定结果,提出了一种适用于监控环境的人体身高测量方法.实验结果表明本文的自标定算法与身高测量方法具有良好的性能.
Aiming at the PTZ (pan-tilt-zoom) camera in outdoor environment, a fully automatic camera calibration algorithm with high accuracy is proposed and applied to the height measurement of the human body under the video surveillance system The first, the camera model is decomposed into a zoom conversion model and a fixed-focus camera parameter model.For the former, a calibration method based on multiple reference point zoom conversion model is proposed.For the latter, the use of camera rotation angle information, fixed focus camera The model has been calibrated online.Because the center of the zoom and the position of the main point are taken as independent variables, they are calibrated respectively in the zoom conversion model and the focusing model, thus ensuring the integrity and accuracy of the model; introducing a plurality of reference points Makes the method of this paper better solve the problem of large model error and too few matching points in the case of multiple zooming and improve the calibration accuracy.Based on the self-calibration results, a human body height measurement is proposed for monitoring environment The experimental results show that the self-calibration algorithm and height measurement method in this paper have good performance.