重叠畸变区域下PTZ主从摄像机定标方法研究

Research on Master-slave Calibration Method by PTZ Camera Under Overlapping and Distorted Area

针对传统摄像机覆盖面不足、镜头成像畸变等问题,结合PTZ摄像机可灵活控制的特性,采用镜头畸变处理算法,提出一种重叠畸变区域下PTZ主从摄像机系统的定标方法。利用PTZ摄像机的预置位点对场景进行分区域采集,提取区域视图中的特征点进行重构处理,使用镜头的径向校正算法完成区域视图的校正,将主摄像机视图与重构视图依据映射关系函数进行定标,通过预置位点调整来提高精度,并引入混淆矩阵的概念验证定标的有效性。实验结果表明,该方法不用考虑摄像机中心偏差等问题,与传统定标方法相比速度快、精度高,适合应用于居民小区等监控环境。

In view of the questions of insufficient coverage and lens distortion,combining with the flexible teatures ot ran Tilt Zoom（PTZ） and distortion processing algorithms,this paper proposes a method for calibration by the fast and flexible feature of PTZ camera. It collects the areas respectively by using preset points,then uses algorithms to extract the region feature to splice and reconstruct,and tries to correct the distorted view of regional,makes a mapping between main view and reconstruction view. It improves the calibration accuracy by using the preset positions,with a statistical of confusion matrix. The method does not have to consider the center deviation with high speed and precision to living area. Experimental results show that the proposed method can be accurate to do calibratation,with good robustness and accuracy.