【作者】 梁栋；

【导师】 龙熙华；

【作者基本信息】 西安科技大学 ， 计算机应用技术， 2003， 硕士

【摘要】 基于图象序列的三维重建技术是近年来计算机视觉技术与计算机图形学技术相结合而产生的一门新的技术。它利用摄像机拍摄场景或物体不同角度的图象，根据不同图象之间的几何关系估计摄像机的外部参数(即摄像机的位置和方向)恢复场景在射影空间的几何模型，再利用自定标技术估计摄像机的内部参数并进而完成场景在欧氏空间的重建。 本文主要讨论了在射影空间以及欧氏空间通过图象序列进行三维重建的方法。射影重建是基于图象的三维重建技术的基础，它利用图象间的匹配点估计摄像机的位置和方向，为进一步的重建打下基础。本文在研究已有射影重建算法的基础上，利用L-M算法对基于基础矩阵的射影重建算法得到的线性结果进行优化，提高了算法的估计精度和稳定性，并在求得所有图象对应的投影矩阵后利用Bundle adjustment方法对空间结构及投影矩阵进行全局优化。 为了恢复场景或物体在欧氏空间的形状，需要估计摄像机的内部参数，完成从射影空间到欧氏空间的转变，这个过程称为摄像机定标。传统的定标方法需要一个已知空间位置的定标参照物，而且每次使用时都需要重新进行定标工作，使用上多有不便，因此，目前多采用自定标的方法。本文讨论了在摄像机不存在畸变、主点已知而其它参数可以改变时的自定标算法。该方法允许摄像机内部参数特别是焦距任意进行改变，给重建系统的使用带来很大的灵活性。 我国是一个文物大国，各类文物的保护是一个非常艰巨的任务。本文对基于图象的三维重建系统在文物保护中的应用进行了探讨与展望，并开发了一个简单的重建系统，为该技术在文物保护中的进一步应用建立了一个框架。 最后对全文的工作进行了总结，并对系统的进一步完成提出了自己的看法。更多还原

【Abstract】 3-D reconstruction based on image sequence is a new technology in recent that come from the combine of the technology of computer vision and computer graphics. According this technology, first we shot the scene from different angles use digital camera, then utilize the relation of epipolar geometry to estimate the exterior parameters (the position and direction) of cameras and to recover the scene in projective space, after this we use the technology of self-calibration to estimate the interior parameters of cameras and to recover the scene in Euclidean geometry.In this paper, the subjects of we discussed is focus on the technology of projective reconstructing and self-calibration that included in 3-D reconstruction based on image sequence. Projective reconstructing is the foundation of 3-D reconstruction based on image sequence. The aim of projective reconstructing is to estimate the position and direction of cameras through matching points in different images so lay the foundation for further reconstructing. On the basis of current methods of projective reconstructing, we used the Levenberg-Marquardt method to optimize the result of linear method so the precision is be improved, and we use the Bundle adjustment method to entirely optimize the structure of scene and projective matrixes.For recover the real structure of scene we should estimate the interior parameters of camera in further, and transform the structure of scene from projective space to Euclidean space, this process named camera calibrating. Often a reference object that position in space had been known is be needed in the traditional calibrating methods. This brings inconvenience in practice. Hence, many researchers have studied the methods of self-calibration. In this paper, we discussed the self-calibration method that no skew is present and the principal point is close to the center of the image but the focal length can be changed freely.China is an ancient civilized country with a long history. Chinese people have created abundant science and art civilization during its prolonged history process. It’s an arduous task to protect all kinds of culture relics on the ground and underground. In this paper, we discussed the application of the technology of 3-D reconstruction based on image sequence in culture protecting. And we developed a simple system that reconstruct the model in Euclidean space. This laid the foundation for the technology of Image-Based Modeling to be used in culture protecting in further.At the end of this paper, the viewpoint and prospect to the future of this system are mentioned.更多还原