点云曲面的复合几何图像表示及其应用

Compounded geometry image representation of point cloud surfaces and its application

为了得到点云曲面的几何图像,提出一种新的点云曲面的表示方法.将点云曲面的笛卡尔坐标转换为球面极坐标,采用极坐标中的二个角度参数作为二维像素位置坐标,将这个位置上对应的极半径的最小值作为该像素上的灰度值,得到第一幅r(θ,φ)灰度图像.利用同样的方法,在剩下的未被转换的点云中构造第二幅θ(r,φ)图像,由前二次转换剩余的点云转换成第三幅φ(r,θ)图像.将三幅灰度几何图像分别对应彩色图像的R、G、B分量,合成为一幅复合彩色几何图像.结果表明,该方法实现了用彩色几何图像对点云曲面的精确表示,从而可以利用图像处理的方法实现点云曲面的高效压缩和简化.

To create geometry images for a point cloud surface, a novel approach for representing point cloud surfaces was presented. The Cartesian coordinates of 3D points were converted to spherical polar coordinates. By taking the two angular values as the 2D pixel position coordinates and the minimum positive radius value in the pixel as the gray value, a r(ψ,θ) image was created. The left point cloud surface was converted into a θ(r,ψ) image. In the same manner, a ψ(r,θ) image was created. The three geometry images were then combined into a colored geometry image by mapping their gray values into R, G and B components respectively. Results show that the colored compound image can represent the original point cloud surface efficiently and accurately, so high ratio compression and simplification of the original point cloud surface can be achieved with image processing.