大尺寸摄影测量局部参数优化相对定向方法

Relative orientation method for large-scale photogrammetry with local parameter optimization

相机定向是摄影测量中的关键技术之一,然而传统定向方式需要预先精确校准定向靶上特征点坐标且测量过程中定向靶位置固定,测量空间范围受到很大限制。为了克服测量过程中对定向靶标的依赖,扩展测量网络和测量灵活性,提出并实现了一种基于局部参数优化的全局相对定向方法。首先,在测量场中布置编码点和基准尺,提取所有图像中编码点的图像坐标和编码值;然后,在归一化8点算法的基础上,采用局部参数优化方法解算2幅图像外方位参数;最后,建立所有图像间的外方位参数转换链,并全局定向出各个图像在预定世界坐标系下的外方位参数。实验结果表明,在6 m×4 m测量范围内该方法的平移向量测量精度优于0.8 mm,旋转角度测量精度优于0.02°,相比于传统定向靶的定向方式,定向精度提高了一个数量级,可以满足大尺寸摄影测量系统对定向精度和效率的要求。

Camera orientation is one of the key technologies in photogrammetry.Unfortunately,traditional orientation approach requires that the feature point coordinates are pre-calibrated precisely and the Exterior Orientation （EO） device is fixed during the measurement process.As a result,traditional close range photogrammetry is highly limited in measurement volume.In order to overcome the dependence of the measurement on EO device,expand the network configuration and measurement flexibility of industry close range photogrammetry,a global relative orientation method based on local parameter optimization is presented.Firstly,coded targets and scale bars are deployed in the measurement volume,which are recognized and measured in every image.Secondly,based on normalized eight-point algorithm,the local parameter optimization method is adopted to relatively orient the image pairs.Finally,exterior orientation parameters among all the images are calculated by establishing a relative orientation conversion chain so that all the orientation results are constructed in a predefined reference frame.The experiment on measuring real camera exterior orientation parameters was performed in a measurement volume of 6 m × 4 m.With the proposed method,the translation vector measurement accuracy is better than 0.8 mm and the rotation angle measurement accuracy is better than 0.02° ; compared with traditional orientation approach,this method improves the orientation accuracy by an order of magnitude.The method can satisfy the requirements of large-scale photogrammetric systems for accuracy and efficiency.