欢迎访问《图学学报》 分享到:

图学学报

• 视觉与图像 • 上一篇    下一篇

消失点计算方法研究

  

  • 出版日期:2016-10-31 发布日期:2016-10-20
  • 基金资助:
    陕西省教育厅专项科研计划项目(15JK1163);陕西理工大学院士工作站建设项目(fckt201501)

Study of the Computational Methods of Vanishing Points

  • Online:2016-10-31 Published:2016-10-20

摘要: 利用消失点(线)进行场景理解与三维重建是计算机视觉研究中的一个重要问题。以往
主要是利用图像点的齐次坐标来计算,由于同一图像点的齐次坐标之间相差一个比例因子,当图像
坐标过大(或过小)时,会引起计算精度的下降。为此,从实际计算角度出发,利用图像点的坐标矢
量和调和共轭特性,研究了消失点的计算方法。经过举例分析和算法验证,证明该方法实用、可靠。

关键词: 计算机视觉, 消失点, 单位坐标矢量, 算法

Abstract: The comprehensive and recognition of 3D scene by using vanishing points (lines) has been one
of the most important research areas in computer vision fields. The conventional computational methods
based on the image homogeneous coordinates will cause deterioration of accuracy when they become too
large or too small, because the homogeneous coordinates of the same image points have different
representations by the scaling factor. The computation methods by use of the characteristics of coordinates
vector and harmonic range are proposed from the perspective of computational processes in this paper. The
results of examples show that this algorithm is more utility and reliable than traditional algorithms.

Key words: computer vision, vanishing points, unit coordinates vector, algorithm