This software specializes in discovering the location and intrinsic parameters of an imaging camera when such information was not recorded in the metadata of an image. It provides a simple command line interface for specifying the input and output for the two camera perspectives.
Using the Get_Platform_Perspective.m function, for a given image, it discovers the position and intrinsic parameters of the camera that took the image. Points in the scene must be correlated to points in real space, typically represented in latitude, longitude, and altitude.
Using the Apply_Platform_Perspec tive.m function, the software transforms an image captured by a known camera at a known perspective to simulate the image captured from an alternate perspective and/or camera.
A straightforward interface with the algorithm allows this to be integrated into a variety of other programs, or used as a standalone analysis tool. Obtaining the platform perspective is a key step in image chain modeling, which aids in the design of new sensor systems.
The toolkit also is useful for simulating alternate views of a scene. This is useful for Earth-observing satellites that are in constant motion and, therefore, are constantly changing their perspective.
This work was done by Steven J. Lewis, David M. Palacios, and Michael C. Burl of Caltech for NASA’s Jet Propulsion Laboratory.