Innovative Post-processing Tools for Command and Control Applications

The objective of this proposed project is to develop methods and systems for the design of nonlinear filters, based on the so-called trilateral filter recently introduced by the P.I., for removal of a variety of noise processes, including white and impulsenoises, from noisy image data. The trilateral filter is much less complex than current nonlinear PDE and variational solutions, and yet produces at least comparable results in removing white noise while preserving image quality (edges and textures).Furthermore, while the nonlinear PDE and variational approaches are not effective for removing impulse noise, the trilateral filter can even be used to remove mixed impulse and white noises. In addition to optimizing the scale parameters of the trilateralfilter and redesigning its spatial component to achieve even lower complexity, our effort will focus on introducing the adaptivity feature to the new filters in this project. In particular, the new filters will adapt to image content with the goal ofhaving the capability of processing both spatial and multispectral images. The ultimate goal is to integrate the filter systems to UAV missions as a post-processing tool for Command and Control to clean up noisy optical, infrared, and radar images andvideo clips. While the state-of-the-art methods and systems for removing noise from noisy image data without losing image quality are very costly, the innovative system to be developed in this proposed project can do at least the same job with only afraction of the computational and implementational costs. In addition, this new system will be more robust and readily integrated as a postprocessing tool in a wide variety of environments for cleaning up noisy optical, infrared, and radar images andvideo clips. It has the potential of DUAL USE applications. For government applications, it could be used as a postprocessing tool for Command and Control of UAV missions; and for commercial applications, it could be delivered as a JAVA Applet to allinternet users for image de-noising.