基于双目单视面的三维重建
发布时间：2018-10-23 06:59
【摘要】：以双目立体视觉理论为基础,采用基于模仿人眼的双目单视面所给出的规律,设计出一套简单的三维重建算法。为了利用双目单视面的原理,采用两个光轴交叉放置的相机来拍摄物体。对所得的两幅图片进行坐标转换,将图像坐标转换为角度坐标,然后在球坐标域进行特征匹配,对得到的匹配点进行曲线拟合,将拟合得到的点转换到空间域,进而得到目标物体的三维模型。实验结果表明,此匹配算法可以准确确定目标物体的位置,而且此算法计算复杂度比较低,后期的数据处理采用分段匹配的方式,使得坐标转换以后得到的目标物体的三维模型相对准确且稠密。
[Abstract]:Based on the binocular stereo vision theory, a simple 3D reconstruction algorithm is designed by using the law given by the binocular monocular visual plane based on imitating human eyes. In order to make use of the principle of binocular monoplane, a camera with two optical axes intersecting is used to photograph objects. The two images are transformed into angular coordinates, then the feature matching is carried out in the sphere coordinate domain, and the matching points are fitted by curve fitting, and the fitted points are converted to the spatial domain. Then the three-dimensional model of the target object is obtained. The experimental results show that the matching algorithm can accurately determine the location of the target object, and the computational complexity of the algorithm is relatively low. The 3D model of the target object obtained by coordinate transformation is relatively accurate and dense.
【作者单位】： 华南理工大学物理与光电学院;
【分类号】：TP391.41
本文编号：2288408
[Abstract]:Based on the binocular stereo vision theory, a simple 3D reconstruction algorithm is designed by using the law given by the binocular monocular visual plane based on imitating human eyes. In order to make use of the principle of binocular monoplane, a camera with two optical axes intersecting is used to photograph objects. The two images are transformed into angular coordinates, then the feature matching is carried out in the sphere coordinate domain, and the matching points are fitted by curve fitting, and the fitted points are converted to the spatial domain. Then the three-dimensional model of the target object is obtained. The experimental results show that the matching algorithm can accurately determine the location of the target object, and the computational complexity of the algorithm is relatively low. The 3D model of the target object obtained by coordinate transformation is relatively accurate and dense.
【作者单位】： 华南理工大学物理与光电学院;
【分类号】：TP391.41
【相似文献】
相关期刊论文 前1条
1 杨忠根;;基于最小冗余子集的单视图鲁棒精确复原目标三维位姿[J];模式识别与人工智能;2001年01期
相关硕士学位论文 前2条
1 王超;基于单视图人脸重建的移动平台发型置换引擎的研究和应用[D];浙江大学;2015年
2 张学芳;基于单视图的汽车造型数字化建模研究[D];大连理工大学;2015年
，本文编号：2288408
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2288408.html
