Multiresolution Raster Chart Image Digitization and Compression

The proposed research focuses on achieving significant gains in the ability to compress the information contained within raster images of nautical charts in such a manner that aids the characterization and acquisition of metadata pertaining to individual chart features. Present NOAA electronic nautical charts, which consist of raster images that are inherently difficult to compress without loss of data, provide little opportunity for enhancement beyond the mere display of their information. Multiple resolution analysis and neural networks are used to recognize and extract text, symbol, contour, and color data while preserving the original chart features. Data stored as objects in an electronic database enable their future use directly for navigation and information, and lossless compression of the raster chart image is performed to achieve greater compression.