高级检索

利用分层深度全景图实现基于图像的绘制

Layered Depth Panorama

  • 摘要: 提出一个快速绘制的IBR新方法,它用分层深度全景图(LDP)作为三维模型的中间表示,通过LDP的重投影变换快速生成新的视图.LDP可以看作物体的一个全景视图,但在每个视线方向有多个不同深度的象素.当用n个平面投影图像建立LDP时,它的数据量不是取决于采样图像的数量,而是取决于物体表面的复杂性.由于LDP可以看作一个独立的图像坐标框架,McMillan的顺次扫描重投影算法能够用于LDP中,避免了Z-缓冲或深度比较过程.该方法在PC机上可达到实时绘制.

     

    Abstract: An efficient image-based rendering method is presented that synthesizes new views in interactive rate.It uses an intermediate data structure called layered depth panorama (LDP) as a compact representation of the object’s surface.LDP can be recognized as a panorama-like representation,but with multiple pixels along each line of sight.When n sample images are preprocessed to construct a LDP,its size is determined by the complexity of the surface,and is independent of n.As LDP is a single view,McMillan’s occlusion-compatible ordering algorithm is adapted.It can also be extended to sphere panorama.

     

/

返回文章
返回