Visualize, analyze and mine satellite imagery using GLIDER software tool

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Wednesday, 20 January 2010: 8:45 AM
B204 (GWCC)
Rahul Ramachandran, Univ. of Alabama, Huntsville, AL; and S. Graves, T. Berendes, M. Maskey, C. Chidambaram, S. A. Christopher, P. Hogan, T. Gaskins, and M. Smith

Satellite imagery can be analyzed to extract thematic information, which has increasingly been used as a source of information for making policy decisions. The uses of such thematic information can vary from military applications such as detecting assets of interest to science applications such as characterizing land-use/land cover change at local, regional and global scales. However, extracting thematic information using satellite imagery is a non-trivial task. It requires a user to preprocess the data by applying operations for radiometric and geometric corrections. The user also needs to be able to visualize the data and apply different image enhancement operations to digitally improve the images to identify subtle information that might be otherwise missed. Finally, the user needs to apply different information extraction algorithms to the imagery to obtain the thematic information. At present, there are limited tools that provide users with the capability to easily extract and exploit the information contained within the satellite imagery. The ones that do provide these capabilities, such as ENVI and ERDAS IMAGINE, are available are expensive commercial products. As part of a current NASA funded project, we are building a software tool named GLIDER to address this void. GLIDER provides users with a freely available and easy to use tool to visualize, analyze and mine satellite imagery. GLIDER allows users to visualize and analyze satellite in its native sensor view, an important capability because any transformation to either a geographic coordinate system or any projected coordinate system entails spatial and intensity interpolation; and hence, loss of information. Thus, GLIDER allows users to perform their analysis in the native sensor view without any loss of information. GLIDER provides users with a full suite of image processing algorithms that can be used to enhance the satellite imagery. It also provides pattern recognition and data mining algorithms for both parametric and non parametric information extraction. GLIDER allows its users to project satellite data and the analysis/mining results onto to a globe and overlay additional data layers. Traditional analysis tools generally do not provide a good interface between visualization and analysis, especially a 3D view, and GLIDER fills this gap. This feature gives the users extremely useful spatial context to their data and analysis/mining results. This paper will describe the features available in current GLIDER release, a short description of GLIDER architecture and the planned features that will be available in future releases.