By Jay Tilley, executive vice president, Sanborn (www.sanborn.com), Colorado Springs, Colo.

Application-specific knowledge is what most nontechnical end users need to make informed decisions. Yet, for nearly two decades, developing a true decision-support system (DSS) has remained an elusive goal in the geospatial industry. Although imagery providers and value-added resellers can skillfully extract useful information from raw spatial data, the industry has yet to effectively bridge the gap between information and knowledge.
 

Geospatial firms typically enhance, annotate, merge and classify satellite images or aerial photos and label them as DSSs. Although these products may play a role in decision making, the delivered information still requires an end user to apply knowledge and analysis to reach an intelligent decision. Most end users don’t have the remote sensing expertise or GIS background to accomplish this on their own.
 

Therefore, the need for true decision support is greater today than ever, as the future of the geospatial industry depends largely on the ability to extend spatial data use to the mass market. Creating DSSs that seamlessly exploit geographic information and feed answers to nontechnical end users will help achieve this objective.
 

“People want end-to-end solutions,” explains Vic Leonard, vice president of Vision at DigitalGlobe. “The further we move away from the power users to the novice users, the more we have to simplify the process so the user makes minimal inputs and the system delivers choices from which decisions can be made.” 
 

A Fresh Approach
Sanborn, a geospatial information provider based in Colorado Springs, Colo., is taking a fresh approach to decision-support system development by attempting to bridge the information-knowledge gap. The basic premise of the company’s solution is to integrate spatial information with an application-specific database in an environment in which automated analysis and modeling algorithms can be applied. Outputs must be quantifiable results that place knowledge in end users’ hands, enabling them to make black-or-white decisions.
 

A hypothetical example from the insurance industry illustrates this idea. To determine the risk of fire, an aerial image can isolate a single house, reveal what combustible materials surround it and possibly determine its composition. This information is certainly valuable, but the insurance company has to factor in the cost of the house and initiate a risk-assessment model to calculate whether the policy is worth writing and how much the fire insurance premium should be.
 

Sanborn is working to combine these disparate elements and technologies, most of which already exist, into a true DSS—an integrated, automated system fueled by spatial and customized data. Once this vision becomes reality, the potential benefit to the geospatial industry is enormous because users at the federal, state and commercial levels will require steady streams of spatial information to feed their systems.
 

True Decision Support
To be of value to most decision makers, a DSS must answer financial- or risk-based questions and integrate key decision indicators from the decision-maker’s operations. Most imagery or photogrammetric systems perform a measurement Sanborn calls a “remote sensing indicator” (RSI). An RSI must be mapped for a Decision Indicator (DI) to have DSS value. Typically RSIs are fused with other operational and ancillary data in a decision model to produce DIs. Although DIs often are displayed in a spatial context, such as an annotated 3-D visualization of an emergency situation, DIs can be as simple as text reports.
 

To deploy a true DSS, Sanborn is focusing on two important developmental issues. The first is improving the extraction of information from raw spatial data sets. Extracted information must be rich in content, accurate in location and tailored to meet specific user needs. Just as the insurance company wants to know the composition of building rooftops and volume of combustible materials to assess risk, emergency-response personnel need to ascertain the location of every fire hydrant and back alley to respond effectively to emergency calls. These features and many others can be derived from high-quality imagery.
 

The next important step is to arrange this information in data tables that can be merged with customized data supplied by end users. In the emergency-response scenario, identifying nearby utility junction boxes and tenants in threatened buildings helps responders determine a course of action. Obtaining this information requires geospatial organizations to form partnerships with end-user companies and agencies to examine the layout of existing databases and determine how spatial data sets can be integrated with other data.
 

In many cases, the geospatial data tables also must be designed to feed data seamlessly into complex computer models and analysis routines as well as displayed and reported to the end user. In the emergency-response example, a self-coordinated evacuation plan that is tailored to the specific event location and takes into account personnel safety needs, routes, risks and responder interface can be quickly generated and distributed to emergency personnel.
 

Sanborn has made great strides in accomplishing these first two objectives, which are common to nearly all DSS deployments. Beyond these steps, however, the architecture and delivery mechanisms must be tailored to meet specific user needs. Prior to each deployment, a DSS developer must determine who within each user organization should receive the intelligence and where it will be delivered.
 

Desktop PCs, communication devices and/or personal digital assistants are the most likely delivery points, but the location of the system itself also must be determined, taking into account security and proprietary data concerns. In nearly all cases, powerful computers and high bandwidth telecommunications will have to be established either within the end user’s location or between it and the DSS service provider.

DSS Market Demand
The advantage of DSS development is that the technology will have universal appeal to all levels of end users. The U.S. federal government, for example, already has announced plans to create “situation rooms” within various agencies, including the Office of Homeland Security and the Environmental Protection Agency. A variety of data will pour into these rooms from multiple sensors, nodes and sources to support critical strategic and tactical decisions.
 

Many of these applications will employ models already developed by the Department of Defense to forecast the impact of terrain on troop mobilization, the spread of airborne contaminants from a biological or chemical weapons attack, or the damage caused by a conventional explosion. In these and many other cases, the key variables are geographic in nature and can be derived from spatial data. Moreover, these types of models may require constant updating of geographic conditions so answers can be delivered at a moment’s notice—the ideal situation for DSS deployment.
 

Federal situation rooms will rely on information fed from similar DSSs at the state and county government levels. These systems will analyze emergency preparedness and response data pertaining to situations that must be handled by first responders who could benefit from using a routing-based DSS to put resources into place quickly at the scene of an incident and efficiently evacuate civilians from the area.
 

But Homeland Security applications aren’t the only reasons local governments are eyeing DSS technology. With cities and counties pinched for tax revenues, government agencies need a more current way to keep abreast of changing property information. By merging parcel ownership and tax databases with feature information derived from aerial photos, taxing districts will immediately assess the values of building additions and accurately determine wastewater rates based on impervious surfaces on a land parcel. A DSS can serve these needs automatically.
 

Similarly, there are numerous potential DSS applications in the commercial market. For example, Sanborn teamed with Risk Management Solutions (RMS), Newark, Calif., to develop catastrophic risk management systems. Insurers use such systems to better understand the risks posed by earthquakes, hurricanes, floods and terrorism on individual-insured locations or entire portfolios. RMS offers an Accumulation Management DSS developed in the wake of the 9/11 attacks.

“The terrorist attacks on the World Trade Center in New York alerted insurance companies to the potential over-exposure they face by insuring multiple businesses in the same building structure,” explains Paul VanderMarck, RMS executive vice president.
 

According to VanderMarck, many insurance companies don’t even realize they have multiple clients in a single building, which could be destroyed by one disaster. To assess the situation, Sanborn has created CitySets for 20 major urban areas. These digital databases are extracted from aerial photos and contain the locations of every building structure as well as critical attributes, such as construction type, composition, height, age and number of stories. Some of this information is gleaned from imagery, while the rest comes from other sources along with topographic, soil and geologic data.
 

RMS used complex models comprising these attributes as inputs to determine the risk for an individual building by multiple types of natural disasters. The models were built with long-standing insurance industry matrices so costs can be calculated to quantify each risk factor. When an insurance company wants to use the DSS, it sends a standard database file of its portfolio holdings to RMS, which simply loads the data into the system. No special technology or expertise is required on the part of the insurance company. In fact, many companies have purchased the RMS system and run it themselves internally.
 

“The system can generate a detailed map showing the insurance company what its exposure is in single buildings or in particular areas of a city,” says VanderMarck. “Or the system can analyze an individual structure and give the company a simple red- or green-light response to the question of whether it should write a new policy at that location.”
 

The Accumulation Management DSS has been received enthusiastically by the insurance industry, and Sanborn and RMS are continuing to derive new spatial data sets and models that will benefit this important commercial market.
 

In Search of the Holy Grail
To view DSS development as the Holy Grail of the geospatial industry is only a slight exaggeration. As vendors continue to provide seamless, end-to-end solutions to new markets, geospatial organizations will expand the demand for spatial products. More importantly, DSSs have enormous potential to attach a quantifiable value to geographic products and services recognized in the government and commercial markets.

  See  Featured Images
 
  Subscribe to Earth Imaging Journal

  Go to Home Page