 |
New ideas for digital imaging
|
| The Pointstream 3d Image display method uses sampled measurement data as the drawing primitive. The system differs significantly from conventional 3d textured triangle display systems based upon a geometric representation using quadrilateral texture image topology. |
|
| Points vs Polygons
Methods of display and delivery for 3d image data evolved from mechanical design and entertainment applications. The primary function of these programs is to create a polygonal surface approximating the digitized point cloud. The Pointstream 3d color image can be thought of as a collection of 3d xyz locations with red-green-blue (rgb) color components just as a conventional 2d color image can be thought of as a set of 2d xy locations (pixel centers) with rgb color components. Pointsteam 3d Images are highly optimized for the smaller displays of desktops and web pages. Because only pixels are used, the Pointstream 3d Image representation is not dependent upon dedicated graphics hardware for accelerated polygon rendering.
|
||
The Pointstream advantage
|
||
|
The physical memory limitations, technical skill, and cost of creating useable content from 3D measurement data has limited the use of digital imaging to a few, well published archival projects. Unlike traditional systems that convert scanned 3D data into a polygonal mesh before rasterization, the Pointstream method takes scanned 3D data directly into rasterization.
|
|
|
| Scalable 3d Images
The Pointstream 3d Image representation provides a natural multiple resolution model design. The ability to scale 3d images in the same way that 2d images are scaled allows a single archive quality data set to be easily re-purposed for a variety of applications and devices. Large archival data sets are sampled and compressed to the desired quality level with minimum user involvement. |
|
|
|
Reliable, effective data compression Pointstream lossy compression methods provide low compressive ratios with high frequency 3D models. The greater the surface and color detail, the higher the compressive ratio and reliability can be achieved compared to polygonal models. Although lossy compression has been performed on polygonal models, geometrical topology information must be kept at the desired reliability level, which conflicts with lower compressive ratios. As the geometric complexity increases the compression ratio and reliability decreases.
|
||
 |
 |
| The 3D raster image approach is ideal for unstructured point cloud data collected with long range sensors |
Aligment of range image data is performed using point cloud data |
