Kolam v2.0: Interactive Visualization of Extremely Large Geospatial Datasets


Kolam is a software package for interactive visualization of massive geospatial datasets and 2D imagery on standard desktop and mobile computers. Kolam is a platform and operating system independent software developed at the University of Missouri, that also takes advantage of commodity high performance graphics processors or GPUs. Kolam supports embedded datasets at multiple resolutions ranging from one km global to 30cm urban mosaics that may be hundred of gigabytes to terabytes in size. Such large datasets are too large to fit entirely in the main processor memory of inexpensive computing platforms. A new multiresolution tiled pyramid file layout structure combined with a quad-bundle memory data structure is used to enable interactive display. In addition to extremely rapid roam, zoom and hyper-jump spatial operations, Kolam supports an arbitrary number of simultaneously visible embedded layers, on-the-fly colormap lookup and histogram enhancement, projection of images onto a spherical surface and elevation maps or terrain rendering. Kolam was originally developed in 2001 and either predates or was cotemporaneous with the currently popular satellite imagery plus map web services from Google Earth (i.e. Keyhole), MSN VirtualEarth, Microsoft Terraserver, Terrafly, etc.

AVI Movie (~11 Mb)

Zipped AVI Movie (~3.7 Mb)


Ian Roth's Thesis Document (PDF format, ~11.5 Mb)

Thesis Presentation (MS PowerPoint format, ~4.0 Mb)

The ever-increasing size of spatial datasets in a variety of applications such as remote sensing and scientific visualization requires the development of new data structures and software tools for efficiently managing and manipulating very large datasets. The new generation of sensors, instruments and numerical models combined with enhanced computational capabilities, produce datasets that are many times larger than the physical computer memory available to view that data. Individual simulations in computational fluid dynamics are now 300 GB in size for which highly interactive browsing combined with automatic feature extraction is desired for understanding complex phenomena (Bryson et al. 1999). How does one ensure interactivity when loading a 100 GB dataset into memory can take up to an hour even with high-performance hardware. Another example from remote sensing is to interactively browse a Landsat mosaic of the Unites States (US). A multispectral (6 band) Landsat TM mosaic of the conterminous US at 30 m resolution, requires 428 scenes (single coverage) and results in an image with dimensions of 218,000 x 95,000 pixels that uses 160 gigabytes (GB) of storage space at multiple resolutions (Plesea and Curkendall 2000). This mosaic was created at the NASA Jet Propulsion Laboratory, and was one of the largest seamless single images at that time; JPL has now produced a global (pan-sharpened) 15m resolution Landsat 7 ETM+ that is about 4TB in size and can be accessed at WMS Global Mosaic. Strategies for dealing with large datasets for scientific visualization include sparse traversal and compression (Bryson et al. 1999). Methods for viewing large digital imagery using tiling and discrete wavelet transform (DWT)-based compression are described by Bradley (1998), Hovanes et al (1999) and Diego et al (2000). Furthermore, extremely large datasets often reside on a remote computer system and hence require integrated support for network access to view and manipulate this data. The objective of Kolam is to design a system for interactively viewing large datasets that not only exceed available memory resources but potentially exist only in the secondary storage of a remote system such as a digital library. We will refer to this software as kolam (K-tiles for Optimized muLtiresolution Access with coMpression).


Kolam was written by Joshua Fraser, Ian Roth and Prof. K. Palaniappan, University of Missouri-Columbia. The current version is maintained by Joshua Fraser. References:

Palaniappan K, Fraser JB (2001) Multiresolution tiling for interactive viewing of large datasets, Seventeenth Int. Conf. on Interactive Information and Processing Systems (IIPS) for Meteorology, Oceanography and Hydrology, Jan. 14-19, 2001, Albuquerque, NM, AMS, 2001, pp.338-342.

Kolam is under continuous development. Please refer to our web page (http://meru.cecs.missouri.edu/mvl/kolam) for the latest version, additional information, and resources.

Joshua Fraser - fraser@meru.cecs.missouri.edu
Ian Roth - ijr3d2@meru.rnet.missouri.edu
K. Palaniappan - palani@cecs.missouri.edu


Executable Binaries:

SAMPLE DATASETS (pyramid file format)