SAR Image Processing for Tracking, Display and Prediction of Sea Ice Dynamics.
Abstract
Sea ice is of fundamental importance in weather, climate and other geophysical processes. It is also an important factor for naval operations in the polar regions, in particular regarding transport of personnel and material in regions where sea ice is likely to be found and assessment and prediction of acoustic environments in polar regions. Because sea ice has a large geographic extent and short time scale for variability synthetic aperture radar (SAR) is a valuable technique in studying sea ice, particularly since images can be collected through clouds and at night. SAR information on sea ice will be available from several satellites in the 1990's (ERS-1, JERS-1, Almaz, Radarsat and possibly an EOS SAR). Automated interpretation techniques are required because of the large number and high information content of the SAR images becoming available. Here we report research on automated-computer-based techniques for such interpretation. The general problem that we face is to extract geophysical information from one or more SAR images. The work reported here concerns estimation of sea ice motion. Hence, we collect two images of a given geographical area some several days apart and note the change in sea ice location which gives us the average sea ice velocity vector. While one can do this manually by establishing tie points on the same piece of ice in the two scenes, an automated algorithm is preferable for both cost and consistency reasons. The research reported here covers the image pyramid area correlation (IPAC) and feature tracking algorithms. These two algorithms, when used together, can provide robust ice velocity estimates for ice that translates and rotates.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1993
- Accession Number
- ADA298941
Entities
People
- John F. Vesecky
Organizations
- Stanford University