PTM's (Polynomial Texture Maps)

Overview | Light Space | Image Enhancement | Viewing a PTM | Building a PTM


Building a PTM

The multiple digital photos used to build PTM's can be captured in a number of ways. Elaborate machines are sometimes used to shoot PTM's. PTM's may also be photographed with only a stationary digital camera, a moveable light source with the ability to evenly illuminate the PTM subject, and a knowledge of the light's spatial positions, relative to the subject.



   Two experimental PTM machines have been constructed by Hewlett Packard Labs. The first machine is suitable for documenting small objects of 6" (15.24 cm) or less. It is a 36" (91.44 cm) diameter black plastic hemisphere with fifty evenly distributed strobe lights mounted such that they illuminate the hemispheric dome's interior. The cartesian coordinates of each light direction vector is known through previous measurement. The digital camera is positioned at the top of the hemisphere and photographs the PTM subject through a view port cut in the dome. The PTM subject is placed slightly below the dome's base and centered in the area enclosed by the dome. The light and camera action is driven by computer commands passed through a controller unit. Computer software triggers the simultaneous operation of the camera shutter and each light in turn.

The second machine is designed to capture images of larger objects. It is made from several metal components. A 90 degree arc 5 feet (1.52m) in diameter is mounted with 12 strobe lights facing towards the center of the arc. One end of the arc is connected to a circular bearing race in the shape of a doughnut. This allows the arc to spin in a 360 degree circle around the bearing race. A computer controlled motor drives the movement of the arc. This assembly sits atop a base that raises it high enough to permit the arc to spin freely. The camera is positioned in the center of the doughnut shaped bearing race with its view passing through the doughnut's hole. As with the smaller PTM machine, the light and camera action is driven by computer commands passed through a controller unit and the software triggers the simultaneous operation of the camera shutter and each light in turn.  

In all PTM capture systems, the photographic images are stored in a numbered sequence corresponding to the order in which they were shot. The captured images are then listed in an LP (Light Position) text file. The first line in an LP file describes the number of input images used. The following lines list each photograph's absolute path name (the location on the computer hard disk), in combination with the corresponding X, Y, and Z coordinates of the light direction vector. The light direction vector's values are expressed in terms of a 3D Cartesian Coordinate System. The three axes of the system are aligned such that the X and Z axes define the camera's image plane, where the subject is photographed, and the 'Y' axis points towards the camera.


The LP file is then fed to Hewlett Packard's PTM-Fitter software which inputs the constituent photographs and synthesizes the output PTM image. The supported input image file formats are ppm, tga, and jpg.

The PTM-Fitter runs from the command line window and the output has two user determined fitting format options, -1 for LRGB and -0 for RGB. LRGB PTMs store each pixel's luminance information separately from its red, green, and blue color information. RGB PTMs store color and luminance together and have the advantage of being half the file size of LRGB PTMs.

Search:


Related CHI Content:
PTM Examples
Reflection Transformation Imaging and Virtual Representations of Coins from the Hospice of the Grand St. Bernard (424 KB pdf paper download)

Related Web Resources:
PTM's
   
HP Labs Resources:
"Making PTM's": highlight PTM info, PTMfitter app download, .lp file creation, (more...)
   
 
Contact | About Us | © Cultural Heritage Imaging 2002-2008