Craig's Satellite Images Clearinghouse

Overview

This is a short on-line report to assist students in digital image processing who are working with satellite images. Specific pointers to references, and general information are contained herein for completing lab projects for the Spring 1995 Computer Engineering Graduate course in digital image processing. The second lab project worked with Silicon Graphics native file format, using the SGI ImageVision Libraries to process the images for histogram equalization and geometric registration. The third lab focuses on correctly registering the satellite images, and two groups of students are working on two sets of satellite data, one from the GOES West satellite and the other from the Polar Orbiter (AVHRR).


Lab2: Histogram equalization and geometric warping

The primary image that I have given you to work with, in case you don't find your own is /projects/spray/craig/CE261/goes_lab2.rgb This image is in the native SGI image format. You can display the image by doing the following:

cd /projects/spray/craig/CE261

ipaste goes_lab2.rgb

Hitting << ESC >> while the cursor is in the window of the ipaste tool will exit that application. You should be able to see a large storm off of the Baja of Mexico in the image. I would expect that contrast enhancement could highlight the coastline very effectively.

If you want to use the image in other formats, there are various routines that will assist you in converting the image (on the SGI machines).

togif yourfile.rgb resultfile.gif # creates a GIF file from the .rgb

tops yourfile.rgb > resultfile.ps # creates postscript

You can also convert images into the native SGI format using

fromgif yourfile.gif result.rgb fromsun, fromppm, frommac, frombin, etc.

The image that I am giving you was taken by the GOES (Geostationary Operational Environmental Satellite) series. The satellites generate visible light images and infrared images from space. I believe this file is from November 18, 1993 in the visible spectrum. It apparently is one of the quadrants coming from the GOES-W (west) satellite.


Lab3: Proper satellite camera to map projection registration

The registration of satellite imagery involves undoing the projection of the satellite camera device, and then doing the projection of the desired map projection. A good introductory text for these transformations is: K. Castleman. Digital Image Processing. Prentice Hall, 1979.

The back projection for the satellite images is nearly the same for each satellite, but there are differences due to the stabilization used for the satellites. The GOES satellites do not rotate about the earth, but rotate with the earth to maintain their position over a certain point. The polar orbiter crosses over a different part of the earth on each pass of its orbit. Registration information is provided in the downlinked image information that the satellite communicates. In further ground processing, this information can be augmented and additional processing is done. By the time we get the images, they have been contrast enhanced (reversed), quantized to fewer bits, and decimated to smaller resolution. But, there is still plenty of work required to incorporate the satellite images into any application.

General Information about Satellite Imagery

Images Courtesy of NASA Goddard Space Flight Center

GOES (Geostationary Operational Environmental Satellite) series, have coverage of the world. There are two primary satellites, GOES-East, and GOES-West, which are positioned over the equator at 75 degrees West and 135 degrees West longitudes. Someone who has assembled a lot of information on the GOES satellites is Dennis Chesters . NOAA has also publications, manuals, and technical reports discussing the satellites. I looked through some of the NOAA manuals on GOES information that Dr. Chesters had assembled. There are numerous GOES satellites, and on the West coast, we are most likely interested in the GOES-7 which operates as GOES-West for the time being. Higher resolution sensors have been launched such as the GOES-8 (GOES-I), but the East Coast has more interesting weather so that's where it will be stationed.

We have also been using the Univ. of Madison Wisconsin's wiscnav code, which Professor Wendell Nuss of the Naval Postgraduate School obtained. See /projects/spray/spray/wiscnav for the Fortran source files and for the library libwiscnav.a currenlty compiled for SGI IRIX 5.2 (ELF Code). The wiscnav code can be used to navigate the GOES images in the UCSC Spray Rendering application. Spray is an application developed under the REINAS project.

AVHRR (Advanced Very High Resolution Radiometer) is a sensor which flies on the Polar orbiting satellites: TIROS-N, NOAA-6-8-10, and NOAA-7-9-11-12-I-J. Information about the AVHRR and NOAA Polar Orbiter Data can be found in the NOAA Polar Orbiter [TIROS-N, NOAA-6, NOAA-7, NOAA-8, NOAA-9, NOAA-10, NOAA-11 & NOAA-12] Users Guide, by the U.S. Department of Commerce, NOAA, Washington, D.C. (301) 763-8400 (301) 763-8443 FAX. Dec. 1991. And on some Web links and the AVHRR Pathfinder Home Page. Check out this sample AVHRR image CAUTION BIG GIF FILE: .

I currently have a copy of this manual that Eric Rosen obtained from NOAA free of charge.

The GOES manual(s) Have Arrived! Included were the GOES User's Manual , the GOES data User's Manual , and the GOES Archival retrieval system documentation file . The glue documentation required for the files that UCSC are working with are included in the Unidata data feed. This includes information about how the SSEC Univ. of Wisconsin has processed the images, and the particular header information. Professor Wendell Nuss has made this information available to us as he also gets the Unidata feed.

Another source of data on satellite images is the book: Weather Satellite Handbook, 5th Edition, by Ralph E. Taggart, Publication No. 179 of The American Relay League, Inc 1994. Dan Fernandez purchased this book for Satellite related image processing and loading work on the REINAS project.

There appear to be a vast number of issues in working with the satellite images, and there can be no means for solving all of these in course projects. If there are particular questions that pop up, one effective way to get them answered is to contact the people who set up the web pages with the Satellite information. Also, don't hesitate to contact authors of papers when you have a question that directly relates to understanding and/or applying their work to yours.

You might want to look at these weather sites for more examples of collected satellite images:


craig@cse.ucsc.edu