Application of Computer Image Processing to Underwater Surveys,
Abstract
Computer image processing has often come to be equated with automated pattern analysis, particularly within the context of spacecraft remote-sensing. Automated pattern analysis in the form of multispectral classifications has been widely used for crop and lithologic mapping from Landsat images. Conversely, bottom photographs and side-scan sonar images collected during underwater surveys have rarely been subjected to automated pattern analysis. Among the reasons for this are limitations of the imaging techniques and complexity in terms of scene and pattern descriptions. It is therefore worthwhile to consider alternatives to pattern analysis for helping interpreters to work with images from underwater surveys. Computer dc notch filtering and geometric rectification by resampling are two key image processing techniques which have proved to be successful in achieving this goal. In addition, since image interpreters often compare bottom photographs and side-scan sonar records to other map based data, namely depth soundings or bathymetry, it is advantageous to also display these data as an image. Finally, while the alternative computer image processing techniques are, by themselves, powerful tools for improving image display, they also show promise for alleviating some of the problems facing automated pattern analysis as applied to underwater survey images. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1983
- Accession Number
- ADP003122
Entities
People
- P. R. Paluzzi