"Remote Sensing Comes of Age" was the title of the keynote presented by Kass Green, immediate past president of ASPRS, and President, Kass Green & Associates. Green is uniquely qualified to present this speech to this crowd at this particular time, as her 20 years of research and application development focuses on the integration of GIS and remote sensing for environmental and policy analysis. Her textbook, Assessing the Accuracy of Remotely Sensed Data (coauthored with Russ Congalton), was recently released in its second edition.
Some fundamental changes in remote sensing and GIS over the past decade include automating the integration of GIS and remote sensing. This involves data exploration (mining) to establish relationships between the imagery, GIS ancillary information, and features on the ground. Kass stated that these methods can be more cost effective and consistent than manual photo interpretation, allowing the computer to label the easy-to-classify features, thereby preserving the skilled remote sensing analysts' efforts for the difficult and complex features.
Kass reflected, "I can remember being at an ESRI user conference 20 years ago and stating, ‘I can't imagine using remote sensing software without GIS software, or using GIS software without remote sensing software.'"
Green is president of Kass Green & Associates and consults on geospatial strategy, technology, and policy issues to private, educational, and public organizations. Green serves on the boards of several for-profit and nonprofit organizations and is currently a member of the U.S. Department of Interior's National Geospatial Advisory Committee. Green also serves on the University of California, Berkeley, Foundation Board of Trustees as well as the university's College of Natural Resources Advisory Committee and Geospatial Innovation Facility Advisory Committee.
The data provider panel was moderated by Mark Baker, Professional Services Business Development Manager, ESRI. Panelists were:
Q. Let's talk about the commitment of our profession to national programs. What are your views for a properly funded organization like "Imagery for the Nation"?
One of the challenges is continued commitment. We are learning the need for more current imagery and not only every 3rd year. We need continued commitment and not only for the initial program, but also for the future.
Q. We talked about greater synergy between remote sensing, imagery and GIS, and bringing them together in a more meaningful way. What are the things to be aware of from a data perspective?
Gerry Kinn: We are only now beginning to realize the benefits of bringing GIS and imagery together. Traditional imagery users had a more scientific basis and specific questions. There are not many different users. The answer lies in extracting or unlocking the appropriate information for each of their communities. We are working with our partners to enable the different information to be extracted. GIS will help. This information extraction will continue to evolve to the point when required information is extracted for specific communities. We will then further see the remote sensing and GIS industries grow. There is still more potential to grow. Imagery is helping GIS data be real time, and GIS is helping give context to imagery.
Rich Cowart: We are still in the Stone Age regarding ability to consume the information in the imagery. With much more information becoming required, it will come down to smart agents to distill this out of imagery. We are a long way from finding all the information, but solutions may not be such a long way off.
Roger Mitchell: To create GIS we need to make observations. Most observations are coming from imagery. That part is very expensive. We need to become more efficient in working it out. It needs to be the case where we can extract this automatically in near real-time. This will bring a new paradigm. Users will gain more confidence in the datasets.
"Google exposing street data results
in less heat on satellite providers
(from a privacy standpoint)." – Joel Campbell, GeoEye
Q. In the post-911 era, regarding the privacy of individuals in society, what data can we share publicly in web applications? How do you respond to these issues in your own businesses, with products such as StreetView?
Joel Campbell: We face this every day. Google exposing street data results in less heat on satellite providers. The information that we share is nothing that cannot be observed from aircraft or driving. We saw the same thing with property boundaries. They are now available. We need to see this on a more global scale. In the next years we will see many different electro-optical satellites over countries. Satellites do not see manmade boundaries, or political boundaries.
There are features we want to protect for various reasons. There are policies in place to protect critical infrastructure. We need to realize that these policies will be different in other countries. Interesting data of other areas will become available.
John Auble: We may be forced into self-regulation. These are observations you can make by different means. We don't know what technology will be coming next, so we need to keep asking the question. We need to be careful of things like blurring out hospitals. By blurring things, we only help point out such features.
Roger Mitchell: Self regulation! From a personal standpoint, for most data paid by the public, that should be available. Some should not be, such as our medical records (though we all contribute to Medicare). Property records are available. There were concerns about this in the past. We have been looking at other people's ‘backyards.' By exposing things, in making data publicly available, we are making this accessible, which can be good. For example, Amnesty International exposed Darfur, where we could show that the world was watching. Much greater transparency will make a safer world. We need to lead the way and show the good about being open.
Russ Cowart: The case can be made that making data available will help poorly funded people (who want to do us harm) could use this data to do us harm. We do have to realize that these tools do make it easier for them.
Gerry Kinn: Openness has always won out over closed. We can see this from history. From a software perspective, we want to make things possible. We need to provide solutions that enable users to make data available or private, and we need to make these solutions robust. Society has to work out what is best for society. It will probably end up much more open. Concepts of privacy are often localized. For example, in small towns, there is less privacy; everyone seems to know what everyone else is doing. In a large town, there appears to be more privacy. We are going to try and make sure that tools required to control the level of data availability are available.
Q. Let's talk about new data products being delivered to our profession.
Gerry Kinn: There are data appliances – several flavors – and more people are stepping up. We have different bundles with different scopes of data; each has different datasets maintained by different people: situational awareness bundles, one for intel community. We want to make sure that these are updated appropriately. We also have ArcGIS online, which enables people to share what they want to make available.
Russ Cowart: i-cubed is one of the providers for data to ESRI's data appliance, and we are updating frequently.
Roger Mitchell: Data appliance is a big box with pre-loaded data that's ready to go, that is in place and will need to be supplemented with more current data to make it more usable. Data on such an appliance is generally out of date. The model should be less about data appliance and more about service-oriented architecture that is updated in near real-time. Data appliances are useful at certain times, but more should be via the internet, so when I do go get it, it's ready.
John Auble: By the time we get our 3rd satellite, we will have nearly daily updates. Therefore what is required are appliances that can also be kept up-to-date with the new data, in case the required communication infrastructure fails.
Joel Campbell: The heart of the issue is about information, not data - trying to provide information authoritatively that's easily accessible. It's about increasing processing capability and bandwidth. GeoEye was fortunate enough to have a satellite passing over President Obama's inauguration. We used the opportunity to determine if we could capture the scene and get it out quickly. We took images 20 minutes before he was sworn in, and about 2 hours after that, they were on CNN and around the world. We need near real-time capture and dissemination of useful information and images.
Russ Cowart: Appliance is an interim step. It is about getting imagery available in an extremely fast manner without being limited by bandwidth. Imagery should be pre-cached until we have ubiquitous bandwidth. Meanwhile, appliances are a very cost-effective way to get imagery into workflows.
Gerry Kinn: What is limiting us is not technology, but our own imaginations. The data appliance was created because there was a business need; it provided value to those who needed it. So we must use our imaginations to find new value beyond what we have been doing for 30 years, and this concept of real-time GIS could actually become a reality.
Q. We have seen fantastic increases for spatial resolution. Are there ceilings from a technology standpoint?
Joel Campbell: This question hits home with us. GeoEye-1 is 41-cm at nadir in pan and 1.65 in multi-spectral. NOAA restricts imagery to 50-cm resolution or courser. So we are already disadvantaged by 20%. We have in planning a second satellite with 25-cm resolution. There is still a pretty big disconnect between what is available commercially and technically… Technology is not the problem; the problem is policy and all good-intentioned people are looking at it. When the policy was created, no one ever imagined that this resolution would be available so soon. If we want to continue to lead, we must have the ability to create better and better technology.
John Auble: This has to be a discussion among many nations. What could we do to guarantee that we would only use a 20-cm sat over the U.S. and 50-cm res everywhere else? Who knows what will happen, in terms of someone coming up with a filter or something else?
Q. What about co-resolution challenges and how to address those?
Roger Mitchell: Whatever the best resolution is what you co-register to. High-res aerial photo data (1-foot) is standard, so why would you co-register to anything else? We will always want to gravitate to the best res at that time. Globally, this is currently 15 meters.
Gerry Kinn: This is becoming a larger challenge. As resolution increases, so does height distortion, for example, from building lean. You have to know where edges of buildings are. You may have LiDAR data, but you still don't know where the edge of the building is. We will not run out of challenges or enthusiasm to address them.