Future trends in imagery SSI ASIBA WALIS Imagery Workshop 28 th November 2005 Stuart Nixon, Founder.

Future trends in imagery

SSI ASIBA WALIS Imagery Workshop28th November 2005

Stuart Nixon, Founder

Scope

Please refer to my SSC 2005 keynote for a discussion on imagery value

This talk looks at requirements that are driving future trends

Working back from needs Most people don’t know what imagery

can do for them because they can’t get access to imagery in the first place

This makes it difficult to look at future trends and their impact

So instead, look at future needs Working back from these needs will

show us what future trends and changes are required

Needs – public use of imagery Public use and need is often overlooked Examples:

Classified gravity maps of little use to unless you are planning to launch ICBM nuclear weapons

Classified high resolution imagery are of little use to disaster victims

Imagery locked up in government departments is of little public use

Public use and availability is the key!

What we need for Public Use

5cm resolution over cities 25cm resolution everywhere else Repeated at least once per month Free use Easy distribution Easy online access These are minimum requirements

What will meet Public Use?

Most demand will be met by airphotos

First address the 80% need (high resolution and high repeat imagery)

Imagery public use requires a value chain of many stages…

Acquisition Processing Storage Use enablement Web delivery Web serving Integration

Thanks to AEROmetrex for this 3.5cm resolution imagery

Future trends on each stage

Where are we today? What is missing? What future trends will help us?

Acquisition

This is what we have today

Solutions exist today for 5cm imagery that can be flown monthly

Digital acquisition solves many problems

Thanks to AEROmetrex for use of this 3.5cm resolution imagery

Processing A big challenge Today: 1TB/week We need: 1TB/hour Future trend will be towards grid

computing solutions for geolocation, mosaicking, colour balancing and compression

An area that deserves active R&D

12,000 airphotos (1TB): - Geocode - Reproject - Colourbalance - Mosaic - Compress

Processing trends We are choking on data volumes Grid computing the way to go (and a

nice technical fit, too) Multi-core CPUs will help a lot A lot of software development

needed A Government+Private “imagery

grid computing” initiative would be interesting and very valuable

Storage 3TB per year is required to image

an entire city 12 times a year at 5cm using 10:1 compression

This is not a lot of storage This year a press release proudly

announced Telstra’s NAS storage costs dropped to $100,000/TB from $1million/TB

This is a joke

Storage trends 500GB drives today; 2TB drives soon Government agencies need to take a

hard look at their IT storage approach Today $10K buys a 5TB RAID NAS

capable box with grid computing ability Future approach to backup imagery is:

Cheap delivery solutions (multi-location) Off-site hard disk replication

Use enablement

Can I use my imageryin common applications?

Use enablement prettymuch solved (Google Earth, ECW, etc)

Use trends Mainstream (Google, Microsoft et al) going

to drive new imagery use – a nasty shock to GIS vendors used to holding the whip

GIS vendors will have to start paying more than lip service to open standards

Would be nice if companies stopped inventing new image formats for each new satellite or imagery solution

We must encourage initiatives like GDAL

Web delivery

Peer to peer delivery proven to share TB’s of imagery to lots of users

35% of all Internet traffic today is peer to peer

GeoTorrent.org demonstrates value Keep it simple!

Web serving

Pretty much solved today Big challenges:

Increasing integration Adding value (location based ads, etc) Availability of imagery

Integration trends

Going to be driven by main stream Advertising revenue probably the

main lead requirement New uses will open up we never

thought of Our industry’s internal politics and

squabbles largely irrelevant

Speculative musings Negative refractive optics – if possible – will

allow sub-centimeter imagery Holographic storage in 2006 will offer

300GB DVD-style disks. Petabyte storage a real long term possibility

Commercial cameras from Canon etc will offer 32+Mega pixel multi-visible spectral solutions at < $15,000/camera soon

High resolution micro-satellite imagery in the future once digital airphotos show the way?

Conclusions

Perhaps surprisingly, most of what we need is here today

Other than for processing, few technical hurdles remain

Big challenges are in other areas – licensing of imagery remains a problem

Airphotos are going to be the driver