Rapid Image Access Technology Without Compression

July 23, 2007

Please note that contributed articles, blog entries, and comments posted on GIScafe.com are the views and opinion of the author and do not necessarily represent the views and opinions of the management and staff of Internet Business Systems and its subsidiary web-sites.

by Susan Smith - Managing Editor
Each GIS Weekly Review delivers to its readers news concerning the latest developments in the GIS industry, GIS product and company news, featured downloads, customer wins, and coming events, along with a selection of other articles that we feel you might find interesting. Brought to you by GISCafe.com. If we miss a story or subject that you feel deserves to be included, or you just want to suggest a future topic, please contact us! Questions? Feedback? Click here. Thank you!

Welcome to GISWeekly!

GISWeekly examines select top news each week, picks out worthwhile reading from around the web, and special interest items you might not find elsewhere. This issue will feature Industry News, Top News of the Week, Acquisitions/Agreements/Alliances, Announcements, People, New Products, Around the Web, and Events Calendar.

GISWeekly welcomes letters and feedback from readers, so let us know what you think. Send your comments to me at

Email Contact

Best wishes,

Susan Smith, Managing Editor

Industry News

Rapid Image Access Technology Without Compression

by Susan Smith

[

Click to Enlarge ]

Ingesting imagery into servers quickly has been a challenge for government and the public sector who need to move large volumes of imagery and data.

Pixia Corp. announced at ESRI that they had been selected by the Department of Homeland Security (DHS) to provide a baseline capability for serving geospatial imagery via the Department's Integrated Common Analytical Viewer (iCAV) system. According to the press release, “iCAV is a web-based geospatial analytical and situational awareness system that helps the DHS mission partners to better prepare, prevent, respond and recover from natural and man-made disasters.” Pixia is a provider of rapid image access technology that allows the rapid ingestion and dissemination of large amounts of satellite and aerial imagery, which is what DHS wanted.

“Pixia's software allows for high performance image access via Open Geospatial Consortium(R) (OGC) standard web services, such as a Web Map Service (WMS) and a Web Coverage Service (WCS). This provides a fully interoperable framework that easily integrates into an Enterprise Service Oriented Architecture (SOA).”

Our technology is a raster file structure,” said Patrick Ernst, director, Business Development. “We’re all about increasing performance and scalability when it comes to storing and retrieving data, particularly image data and pixels although it does apply to other data. It’s not a compression format, it’s more of a file system within a file.”

[

Click to Enlarge ]

Ernst explained that there are two branches of Pixia technology, one is the file system within a file where they’re dealing with raw sensor data. You might have a file system container for standard files like JPEG 2000 files. The other branch, a more traditional capability, allows for the creation of large mosaics in the file structure, which is called NUI, for New Universal Image format. Pixia licenses this technology. “Our flagship product is the nuiGeoEncoder much like the Adobe PDF model that allows you to create very large PDF files but they are large mosaics, a bit like MrSID or ECW but different in that there’s no compression being added in there,” Ernst

noted. The nuiGeoEncoder reorganizes and optimizes the way in which the pixels are stored to provide improved access performance and scalability. “We do enable JPEG 2000 compression to be added if customers want it, but mostly we’re just laying the data out in an optimized fashion to ensure very rapid access capability, very rapid writing and reading of that data on the disk drives or any other type of storage media, from solid state right on down to CDs.”

A lot of Pixia’s customers are in the defense, intelligence and homeland security markets, where compression is not always the ideal way to handle data. Organizations often compress at very high compression ratios because that’s the only way it becomes easily manageable, said Ernst. Pixia lets you store that uncompressed data and as in the example of DHS, you could take the entire continental U.S. and make one file out of it.

[

Click to Enlarge ]

“There’s a tradeoff between disk space and performance and access – we do enable JPEG 2000 compression, so if we have customers who are concerned about disk space, then they can just apply compression,” Ernst pointed out. “But if we have customers very interested in maintaining the integrity of the data and potentially they want to utilize servers for local access as well as serving it through a SOA through a web service, we have OGC connectors, such as Web Map Service (WMS) and Web Coverage Service (WCS) (no web feature service, since we only deal in raster domain, not vector). In the WMS environment, compression usually happens on the fly, so it’s

advantageous to have data already uncompressed, so you don’t have to decompress it and recompress based on a get map request. If you have clients who want a JPEG file vs. a PNG file, you can create that compression right on the fly and then send it down the pike.”

Compared with compression, Ernst said Pixia is “extremely fast, we’re talking roughly 100 milliseconds for accessing anywhere randomly within a gigantic image. You could be talking hundreds of thousands of pixels by hundreds of thousands of pixels and you’ll access random. We have pyramid levels in there, and you can reduce resolution data sets to whatever level you want. So you’re not constrained to base powers of 2. If for some reason you want a 6-inch pyramid level if you then want a 8-inch or 9-inch or a 1.2 meter you can set it at whatever you want.”

If you have a data set of IKONOS imagery, that’s one meter per pixel ground sample distance, and see that its native resolution is the level zero, Pixia as well as Mr. SID, ECW, IMG, even National Imagery Transmission Format (NITF), allow you to create pure mid levels of the imagery. This is so that when a user wants to look at a larger area but they don’t want to zoom in all the way, that scaling has already been done for you. The way the other formats allow you to do this is by taking every other pixel and dropping out the middle pixel. Pixia lets you set those pyramid levels at whatever you want which works well for sensor fusion capabilities. This is good when you have

different sensors at different resolutions and you want to actually find common ground there, as in fusing together a DigitalGlobe QuickBird image with a 1 meter IKONOS image to create a 1 meter dataset.

“What we have in some ways a geo raster database within a flat file. What that allows us to do is to stack these different images on top of each other. You can update them quite quickly. So you’re looking at over 40 terabytes of imagery that’s being served up by this one server,” explained Ernst.

The resolution of the data used is really a function of the sensor and the data constraint, Ernst stated. “We can super sample it and zoom in past the resolution of the data. We can handle any resolution data you have, no limitation. Because we’re actually overlaying different resolutions of data, it’s not like we’re saying here’s the base resolution for the whole world and let’s plop that in there and now you’re constrained to that. You can just pull in data at different resolutions.”

For the DHS iCAV program, Pixia will be doing one meter of the whole country with 133 cities sprinkled inside the one meter. There will be 15 meter of the whole world, at different resolutions. The 133 cities, based on critical infrastructure protection, will be much higher resolution varying from 6-inch to 1-foot. The 1-meter will be much coarser resolution yet it gives you a broad, synoptic coverage of the entire continent of the U.S.

“The reason we’re able to ingest much more rapidly than your average spatial database, or spatial and relational database stack, which are the more traditional approach for ingesting big amounts of imagery,” Ernst explained, “is you have a limitation of 2-5 gigabytes per hour to bring that data in. That’s a hardcore limitation because you’re actually building up that database. When you take a flat file approach, we’re seeing an excess of 50 gigabytes an hour per processor. So if you’re looking at serving imagery across an enterprise and you have quite a few processors

available to you, you’re suddenly looking at exponential increases on ingest performance.”

1 | 2 | 3 Next Page »

You can find the full GISCafe event calendar here.

To read more news, click here.

-- Susan Smith, GISCafe.com Managing Editor.

Rating: