为了提高大视场靶标场标志点自动编码的容量和效率,设计自编码标志和普通标志,建立以自编码标志作为编码基准、以普通标志点作为控制点靶标的室内靶标标定场。基于“同名点同编号”的原则,提出自编码标志和普通标志点的自动编码方法。该方法能自动实现同名影像标志点与同名物方控制点标志的自动编码映射,避免人工交互设置影像标志点编号,提高了标定的效率。推导出含有隐式约束条件的多视组合相机标定的间接平差模型的一般形式,并用其标定一种由网络摄像头组成的多视相机。实验结果表明:该方法具有标志点编码容量大(可达65535个)、解码效率高、能自动解码的优点;该标定方法的反投影像点平面方均根误差(RMS)小于0.20 pixel且相机之间的相对定向元素波动较小,能够有效地提高多视相机标定的稳健性和精度;该方法可方便地扩展到其他同类多视相机或者全景相机的标定应用中。

Self-encoding and common markers are designed to improve the capacity and efficiency of automatic encoding of the markers in a large-sized target calibration field. An indoor target calibration field is established, where the self-encoding markers are adopted as the encoding reference and the common markers are adopted as the control-point targets. Based on the principle of “homologous points with same number,” an automatic encoding method is proposed to deal with the self-encoding and common markers. This method can automatically realize the automatic coding mapping between the image markers and the object control-point markers with the same name, avoid setting the number of image markers manually, and improve the calibration efficiency. Subsequently, the general form of indirect adjustment model is derived for the calibration of a multi-view combined camera with implicit constraints. This model is then applied to calibrating a developed multi-view camera comprising webcams. The experimental results denote that the proposed method exhibits the advantages of large encoding capacity (up to 65535), high decoding efficiency, and automatic decoding. Because the relative orientation parameters exhibit less fluctuation and the back-projection root-mean-square (RMS) error in the image point plane is less than 0.20 pixel, the robustness and accuracy of the calibration can be improved using the proposed calibration method. Furthermore, the usage of the proposed method can be easily extended to other similar multi-view or panoramic cameras to obtain appropriate calibration.