NOAA Raster Chart Conversion Into Digital Charts Using Multiresolution Analysis for Object Recognition and Compression

David Pendleton and R. Glenn Wright

Abstract:	The National Oceanic and Atmospheric Administration’s (NOAA) Office of Coast Survey (OCS), produces the Nation’s marine navigation charts to enable safe efficient passage through its rivers and ports. The OCS has transitioned production operations from manual drafting and engraving processes to routine use of advanced computer technologies using raster chart image files. The resulting raster chart is a georeferenced digital image and is an accurate representation of the familiar paper chart. All 1,000 NOAA paper charts are currently available in raster form. This paper summarizes the results of a highly successful Coast Survey Development Laboratory prototype system developed under contract by GMA Industries, Inc. The current raster chart is limited to simple display of whole chart images, preventing the user from interacting with, interrogating, or selectively displaying the contents. The project focused on overcoming these limits by developing an entirely new lossless image compression technique. The approach used has the beneficial side effect of digitizing individual chart features, thereby enabling the utility of a vector nautical chart. A neural network algorithm recognized and extracted identified chart features at multiple resolution levels, deleted those from the background image, and stored them as feature objects in a database. The simplified background image was then compressed using lossless techniques, which resulted in a four-time reduction in file size, including all 6 color planes. The database was then used for decompression, which resulted in an exact reproduction of the original chart. The database query operation demonstrated an entirely new search, selective display, and highlight capability for the text, soundings, and cartographic features of the original raster chart. These preliminary results could represent the birth of a new generation of raster charts with true digital data query and selective display capabilities not possible with existing image-based raster products. The use of these new techniques could reduce storage and bandwidth requirements for nautical charts and many other types of data rich images. Such automated extraction of data directly from images may contribute to automating raster to vector data conversion and maintenance processes for related electronic chart products and for similar imaging and geographic information systems (GIS) applications.
Published in:	Proceedings of the 58th Annual Meeting of The Institute of Navigation and CIGTF 21st Guidance Test Symposium (2002) June 24 - 26, 2002 Hyatt Regency Hotel Albuquerque, NM
Pages:	295 - 303
Cite this article:	Pendleton, David, Wright, R. Glenn, "NOAA Raster Chart Conversion Into Digital Charts Using Multiresolution Analysis for Object Recognition and Compression," Proceedings of the 58th Annual Meeting of The Institute of Navigation and CIGTF 21st Guidance Test Symposium (2002), Albuquerque, NM, June 2002, pp. 295-303.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In