Super-Resolved Image Processing Software (S-RIPS)

The Super-Resolved Image Processing Software (S-RIPS) was developed at NASA Ames Research Center as a way to improve the quality of images from space probes. It creates a super-resolved image, with both spatial and gray scale resolution improved, by using a number of source images of the same object from slightly different perspectives. As the number of digitized source images increase, the software produces even finer resolution images. The software currently implements powerful registration and compositing routines to refine the images. It also includes a Bayesian probability function that exploits the neighboring correlation in real images. The effect of adding the spatial correlation component to the reconstructed image is to smooth out high contrast local variation, which is caused by fitting the noise in the data. The technique produces a model at higher resolution than the original data. As more views are incorporated into the model, the degree of super-resolution increases. This technique is closely related to computer tomography as used in X-ray interpretation, astronomy, and seismology.

Potential Commercial Uses

• Improved resolution of fax transmissions by integrating the output of multiple feeds using S-RIPS.
  
• Much improved composite image from multiple TV images.
  
• Integrating information from different satellites, or the same satellite at different times, to give a higher resolution image.
  
• As a tool in image editing workstations by allowing merging of information from multiple images.
  
• Going beyond the resolution limit of a given scanner by merging the information from multiple images.
  
• Achieving good resolution on output from low bandwidth image transmissions, such as in teleconferencing or videophones.
  
• Extracting better stills from multiple frames of a security surveillance camera.

Benefits

• Improved imaging: Provides higher resolution than might be obtained by standard means.
  
• Numerous applications: Any set of digitized images of the same area can be processed, provided they are all taken from roughly the same direction and under the same conditions.
  
• No extensive equipment needed: Only the digitized images and the software are required. However, additional information such as knowledge of camera characteristics can speed the processing.
  
• Reduced expense: Compared to more expensive cameras used in high resolution imaging systems, the S-RIPS software solution can be much less costly.

Options for Commercialization

NASA Ames is currently seeking partnerships to develop commercial software code. The mechanism for transferring this technology will likely involve close collaboration between a company or companies and the NASA development group at Ames. The research code is currently written in LISP and runs on several platforms. An adopter of this technology would implement these approaches in a more efficient code for a commercial environment.

Contact

NASA Ames Research Center
Commercial Technology Office
Mail Stop 202A-3
Moffett Field, CA 94035-1000
Phone: (415)-604-5761
Fax: (415)-604-1592

http://ctoserver.arc.nasa.gov:80/TechOpps/s-rips.html

Key Words

Image Processing, Image Enhancement.