Geospatial Compression and Viewing Software options




A comparative review of low-cost and free software for GIS imagery
by David B. Davis




Although hard drive and other forms of data storage have increased dramatically over the past few years the amount of available data has grown at an even greater rate. Geospatial databases as well as web pages, graphic arts, CAD, and medical technologies are using more and more raster data. With aerial imagery from both photographic and digital sources as well as the various government and commercial satellite imagery providers a city or county project can easily amass tens or hundreds of gigabytes of data.

With ever more powerful computers along with data storage and retrieval systems capable of managing several thousand terabytes of data it is possible to keep up with almost any amount of imagery. That is if your equipment budget is equally as large. Most organizations however, whether public or private, must operate within a limited budget. The ability to easily use and manage an always growing image database with limited equipment and funding can be a very real problem.

The goal was to decrease the file size of the imagery while at the same time maintain the overall quality. Fortunately where there is a need there is usually someone ready to supply the desired product or service. A number of solutions to this dilemma have arisen from the intertwining worlds of academia, defense, government, and the private sector.

There are image compression algorithms available from various sources, ranging from well know commercial software to obscure student projects. The programs may cost thousands of dollars or be available for download at no cost. The basics of the most common compression options will be briefly covered after which the major emphasis here will be on the market leaders. Special attention will be given to the no cost compression programs that have become available in the last couple of years.

Fractal Image Compression

Fractal image compression is a method that appears to be more popular for research among university students than in the commercial world. This type of image compression and creation uses geometric formulas to mathematically represent imagery. It can reduce the size of an image file but is also very useful for creating fascinating designs as well as backgrounds for films or computer games. A few companies offer software that uses fractal methodology, such as MediaBin however it seems that most of the work involving fractal image creation and compression is occurring in academia. Some examples of this can be seen at: Waterloo Fractal Compression Project or Fractal World, a Lock Haven High School Algebra 2 honors class project.

The most widely known and used compression methods are JPEG and GIF. Most people are familiar with these names because of their extensive use as Internet graphics formats. GIF is an acronym for Graphics Interchange Format. This compression methodology uses the LZW algorithm from Unisys. Because of royalties that Unisys requires as well as other limitations described below this compression methodology has progressed very little. JPEG comes from the Joint Photographic Experts Group. The current JPEG standard has not changed for over seven years. While the original version uses Fourier transforms the soon to be released JPEG 2000 will use wavelet transformation to compress imagery up to 200:1.

The GIF format uses substitution to compress imagery. The software looks for patterns and assigns the pattern or sequence a single number. A table is then created that contains the sequence represented by each of the individual numbers. In this way when the image is accessed later on the software can decipher the code and thus display the image as it originally existed. Imagery with large areas of similar tone or pattern, such as a smooth lake surface or desert plain can be compressed much more than regions of complex patterns. The GIF compression method does not lose any of the image data, in other words it is a lossless method. It does however have several limitations, especially regarding GIS usage. First of all it is limited to a 256 color palette. This may work fine for a simple clipart but would not come close to the needs of displaying a complex map or CIR aerial photograph. Secondly, few if any GIS programs offer the option of displaying GIF imagery. Finally, because it is a lossless method the level of compression can not be changed. The software will find all of the feasible patterns within the image, make all of the possible substitutions, and because no data is lost that is as much as it will do.

The JPEG compression format is a lossy method. When compressing the imagery some of the data is lost. Unlike with the GIF method, with JPEG you can decide how little or how much to compress the imagery. The more the image is compressed the more data is lost. Obviously the more the image is compressed the worse the image will look. An image that has been compressed at a high ratio will look much like an image that was scanned at a very low resolution. Depending on the intended use for the imagery often times a happy medium between image file size and image quality can be reached. For use in a GIS the only real choice between using GIF and JPEG for image compression is JPEG. This format is not limited to a 256 color palette and most important of all can be displayed in most any GIS program. Even no cost viewers such as MapSheets Express (ERDAS) , ERViewer (ER Mapper) and the Imaging program in Windows will all work with JPEG compressed images.

Aerial Photograph - Central Uruguay, city of Sarandi del Yi, Flown 1194. Scanned at 256 Grayscale 400 dpi. The images show the difference in high, medium, and low quality compression levels. (Highest quality equals lowest amount of compression.