Digital Mapping and Large Format Scanners

By Jeremy D. Wilson

• Introduction
• What is Digital Mapping?
• Scanning as Data Capture Solution
• Cost-effective Data Conversion
• Managing System Implementation
• Maximizing the Hybrid Environment
• The Maturing of Scanning Technology
• Software Advances
• Data Capture Strategies
• Raster-based Analysis
• Conclusion

INTRODUCTION

Digital mapping encompasses some of the most diverse and complex of today’s enterprise information technologies. It also presents some of the most acute data management challenges in the industry today.

In particular, data-capture, conversion and verification account for 60 to 80 percent of the costs associated with most digital mapping projects. While the relative costs of hardware and software for these systems have declined in the last five years, data conversion costs have changed little. Why does data conversion persist as such a costly and difficult endeavor? There are several reasons:

• The size and scope of the data is so profound that it becomes a mammoth job to convert an entire collection of maps, land and infrastructure information.
• The variety of data types, from hand-drawn maps, to remote sensing and aerial photography provides a cornucopia of information. But that information must be consolidated from a host of inconsistent and incompatible formats into a singular data structure.
• The paper documents themselves may be old, inconsistent and frequently outdated.
• Data conversion is a labor-intensive pursuit. A person has to sit down at a computer terminal and physically input the information from a paper map to create a CAD or other graphics-based document.

Many digital mapping users have looked to large-format scanners as a solution to some of these challenges with varying degrees of success. The complexities of large-scale data conversion, the need for ongoing data input and the integration of high-quality remote sensing images make it vital to include the means for data entry capabilities provided by scanners. But only recently has scanning delivered the functionality necessary to fully meet the complexities of digital mapping.

Scanning systems have matured in the last two years. With a synthesis of new software, refined hardware and increasing processing power, scanning systems have improved by quantum leaps, particularly when it comes to data capture and management solutions for digital mapping applications.

WHAT IS DIGITAL MAPPING?

More than a single application, the term digital mapping defines a broad range of systems applied in a multitude of variations. The scope and breadth of companies using digital mapping applications spans a surprisingly large number of industries. The market for digital mapping systems may be classified in three main categories:

• Natural resource management, which includes activities like forestry, oil and gas exploration, fisheries and agricultural mapping.
• Environmental management, including environmental protection and cleanup, oceanography, meteorology and environmental research.
• Infrastructure management, such as gas and electric utilities, telephony and cable, transportation, municipalities and regional agencies.

Digital mapping represents one of the fastest growing segments of the information technology market, with annual growth rates currently estimated at 14 to 20 percent, depending on whose numbers you believe. It is a complex computer environment, built to address the management challenges of its unique data types. They may be identified as one of several technologies, including Geographic Information Systems(GIS), Land Information Systems (LIS) and Automated Mapping and Facilities Management (AM/FM) systems.

Digital Mapping: Maps are an increasingly important part of the enterprise data environment. Acquiring this data and making it available in a useable form across the enterprise remains the greatest challenge in implementing digital mapping.

In all these forms, digital mapping is an extended variation of Engineering Drawing Management (EDM). Whatever the individual moniker (and no one has found an adequate acronym to define the whole market), they share certain common critical missions:

• They manage data across a large geographic space.
• They provide a graphic representation of the land and infrastructure resources to be managed.
• They link that graphic representation of each element to a corresponding database (or multiple data-bases) containing relevant attribute information.
• They support large-scale maintenance, planning, engineering and development functions related to the management of land and infrastructure assets.
• They must interface with other organizational information systems, such as document management, customer information, executive decision management and a host of industry-specific operations support systems.

Clearly, these are enterprise challenges. How a particular enterprise confronts these challenges varies based on its needs, objectives and information technology philosophy. Consequently, there is a wide range of differences in what these systems look like and what elements they use. A firm managing private land for forestry or oil and gas exploration, for example, will have far different objectives and needs than a utility managing its electric distribution lines, or a phone company creating a support system for its new broadband communications network. Even within an industry, no two organizations manage their land and infrastructure information in quite the same way. Consequently, they implement their digital mapping systems in radically different ways.

One common thread exists in each implementation, however, that binds together all digital mapping implementations: no segment of the information technologies market presents more complex or more intractable data capture and management challenges.

SCANNING AS A DATA CAPTURE SOLUTION

Digital mapping encompasses a series of interrelated processes and technology issues that make it difficult and expensive to acquire good digital data in a form that can add value to an organization’s long-term operations. At the heart of these challenges is the creation of useable digital data from paper source documents. These challenges are magnified by the size and scope of data sets, the complexity of the data structures, and the nature and quality of source documents. This environment is ideally suited to the use of large format scanners for several reasons.

1. Scanners provide a cost-effective means for the initial conversion of historical paper documents. By eliminating the tedious redrawing of existing architectural and engineering drawings, CAD resources may be optimized for new designs and updates. Information contained on old, deteriorating documents may be preserved more effectively.
2. System development and implementation cycles may be reduced significantly, achieving productive usability in a shorter time and increasing the return on investment.
3. Scanners also enable users to maximize use of the hybrid raster-vector environment of today’s advanced digital mapping software. Using scanned documents in a hybrid raster/vector environment can enhance productivity and increase efficiency.

Cost-Effective Data Conversion

Using a large format scanner to create raster digital files from existing paper maps is one of the few technologies to demonstrate any reasonable impact on the time and cost of converting historical records.

Particularly organizations involved in infrastructure management have an overwhelming repository of historical map and engineering data. A local telephone company or electric utility, for example, probably started building its infrastructure near the turn of the century. Many cities and municipalities may have land and property records dating back two hundred years - even further for many European cities and towns. This historical data is stored on countless thousands of paper maps drawn at different times, with different media and different mapping standards, symbologies, based on different field data collection procedures. The result is a complex and inconsistent set of records that may number in the thousands or tens of thousands.

Scanning provides a direct, efficient means for creating digital files. Users can establish a starting point for electronic records management and move the digital mapping implementation process along more quickly in order to reach a production mode and realize at least some return on investment.

Consequently, scanning can serve as an integral part of most digital mapping environments, enlisted to forge a link between the new electronic system and all the paper maps and historical records.

Link from paper to digital maps: Scanning can serve as an integral part of most digital mapping environments, enabling users to input data quickly and efficiently as a starting place for their electronic mapping programs.

Scanning is immediately useful for creating a digital representation of source documents. It also provides a mode by which deteriorating documents may be captured and archived, insulated from the deterioration of time and use. Once the documents are in the system. They can be used in a number of ways:

• As templates for converting information to vector form. Many data conversion strategies incorporate scanning to one degree or another. In fact, the fast-track strategies most typical in today’s results-oriented implementations use scanning to get drawings into the system fast. Scanned drawings may be used only temporarily to help convert information to vector form. The images are displayed and then copied, either manually or with auto-conversion software, to develop a new vector-based data set.
• As archives for periodic reference. Some data may be less vital to the day to day operation and analysis that it is not cost-effective to vectorize it, and yet it may be necessary to keep that data for occasional reference or reporting. In these cases, whole documents may be scanned, indexed and stored for use when needed.
• As backdrops for vector drawings. In other cases, images may be incorporated as part of the ongoing database. USGS quads, for example, may be loaded as a backdrop and only key information, such as street centerlines, utility right-of-ways, addresses, and other network-centric information will be digitized. If a higher resolution or more current detail is required, aerial photos may be input as the basemap image.
• As a permanent reference document attached to the vector file. Many users will incorporate drawings as an integral part of their digital mapping systems. For example, an electric utility may scan and store individual service drawings into its digital mapping systems, linked to street addresses or some other key vector feature. If the item is not critical to vector-based network analysis, or if conversion can be placed on a lower priority and delayed to a future phase, it may be beneficial to store the record in this manner.

Managing System Implementation

Most current users are exploring fast track implementation approaches for their GIS projects. Historically, implementing a GIS could take five to ten years, with little or no payback until the project neared completion. Unfortunately, by the time a system is implemented it also may be obsolete. This cycle of protracted implementation followed by immediate obsolescence is unacceptable in today’s business environments. Consequently, GIS developers and implementers have taken a more practical, incremental approach, with the intent of producing measurable benefits within months instead of years.

At the core of these strategies is the use of scanners to input paper map files quickly so the electronic system can be used immediately. That data which needs to be in vector format, to support analysis and reporting applications, then can be prioritized and digitized over time. That data that offers the most benefits from the vector-based analysis capabilities of the system are digitized first.

A utility, for example, may digitize its distribution network first and allow streets and building records to follow later. A county, on the other hand, may first digitize its property boundaries in order to modernize its land ownership and property tax systems. Both will likely focus on areas of highest activity first, so the system will support their most pressing day-to-day design and reporting needs.

Return on Investment: In the traditional data conversion model, the return on investment is delayed until most of the conversion is completed - perhaps for several years. An incremental approach, on the other hand, puts the system to work much sooner, even though the entire system may be incomplete. This approach accelerates the ROI and provides measurable gains in productivity earlier in the project.

A fast-track strategy is essential for these systems to be successful. To survive they must succeed in accomplishing four objectives in the corporate environment:

1. Deliver a productive environment in minimum time. Instead of waiting for the entire system to be complete, an incremental approach enables organizations to get certain core functions operational quickly and demonstrate gains in productivity. Time to productivity has become a new and important measure of success for digital mapping projects.
2. Achieve and maintain management support. Support is increasingly based on performance. Management must be able to recognize a system’s benefits to the overall goals and information needs of the organization.
3. Win user acceptance. User acceptance is based in a large part on the ease of transition from the old paper-based system to the new. Scanned images can help support the transition by providing the familiarity of the old documents within the digital environment.
4. Demonstrate a viable return on investment. In this survival-of-the-fittest environment, information management systems are no exception. Those projects that do not effectively and measurably benefit the bottom line are at risk of termination. Only a productive system can show a positive impact on the bottom line.

Maximizing the Hybrid Environment

Utilizing a hybrid digital mapping system in which raster and vector data can be displayed and used simultaneously, opens new options for data management and storage. Not all data need be converted to vector form. Only that information which is required for vector-based analysis need be converted. Other information, which may be useful to the operator for reference, but not critical to network analysis, for example, may be left in raster form and displayed behind the vector network information.

Grayscale and color images, such as aerial photos and satellite images, may also be incorporated as a backdrop to vector data. The introduction of images to the digital mapping environment has proven unexpectedly beneficial to these systems.

The interpretive quality of photo-realistic images, held as raster files in the digital mapping system, adds powerful benefits, particularly for organizations that must gain approvals from non-technical administrators or that must meet demanding regulatory reporting requirements. People understand a photo better than they comprehend the lines of a map. In digital mapping, a picture is worth a thousand vectors.

The availability of inexpensive land data is probably the most compelling change in the industry today. With the availability of new remote sensing products, including one-meter resolution satellite images, low-level aerial photography and radar imaging, the need for a detailed vector landbase could be significantly reduced in just a few years. Users will store data in a hybrid environment, and convert only that data which must be in vector form to support analysis. Everything else will be stored in the most economical form, as raster images.

Moreover, the ability to integrate scanned drawings and photographs into the digital mapping environment is now being followed by the development of raster-based analysis tools. Using spectral analysis and other tools on grayscale and color images, these tools to can automate processes like identifying tree growths, locating buildings and calculating footprints, determining the effects of the tree growth on overhead power lines and other analyses which were formally done by first converting data to vectors.

While these new advancements provide growing benefits and functional capabilities for digital mapping operations, scanning will remain an important element in the overall system. As the primary - and most cost-efficient - conduit between hard copy and digital documents, scanning will continue to evolve and improve, along with the other system components, to satisfy the demanding needs of digital mapping which continue to push the boundaries of document management beyond its limits.

Footnote: While raster files are much larger and require more processing power than vector files, the high cost of creating vector records makes raster images far more economical. With today's high-powered low cost workstations and personal computers, hardware is significantly cheaper than manpower.

MATURING TECHNOLOGY

As a technology, scanning has matured significantly in recent years. It now provides a level of functionality and ease of use that enhances its value to the enterprise. With sophisticated design and manufacturing, scanners are among the most reliable pieces of equipment on the market. Support software for the scanning process and for scanned image management is now catching up.

Software Advances

In the last twelve months a new generation of scanning support software has reached the market making scanners exponentially more useful in the digital mapping arena. The culmination of years of development, these new systems provide advanced functionality in three key areas:

Indexing.

In large-scale operations, users may scan thousands of images. They must then determine their relative order, add geo-positioning and index the files to build a contiguous data display. The latest generation software can provide a clear scan, intelligently positioned and set in an index schema that accounts for both scale and location. Such automated indexing capability makes it possible to establish a basemap in much less time than by manual input methods.

Hybrid Reporting.

A major element in digital mapping is the ability to parse data relative to geographic location and report that data in a geographic representation. Once indexed, software can now provide fundamental view and query capabilities, which may be applied to geospatial reporting structures.

Data Conversion.

Software solutions to labor-intensive data conversion continue to advance at a rapid pace. Whether it includes optical character recognition, automated line following or interpretation from photographic images, the ability to automate more and more of the data conversion process helps gradually reduce the associated costs.

Data Capture Strategies

The specific strategy for incremental deployment may vary significantly from one user-organization to the next. And, while the ultimate goal is still to convert key data to a vector format, new software is making it possible to extract more and more information directly from raster images. Recent advances in software for raster images have raised scanning to a higher level of effectiveness and value for mapping users. With this increasing functionality, scanning systems have evolved equally refined strategies for using scanning to convert data to vector form from original paper drawings. Some of these strategies are introduced below.

Screen Digitizing.

Screen, or heads-up, digitizing is the most basic use of scanned images. A paper map is scanned and displayed on the computer screen. An operator simply traces the information creating a vector copy of the raster image. Once completed, the scanned image may be archived or discarded. As a data conversion strategy, this approach has definite advantages over copying a paper map sitting on the table next to you. Speed increases, the error rate decreases. However, the process is still manual, labor intensive and costly.

Interactive Vectorization.

At the next level, conversion is partially automated. Software will convert certain elements and prompt the operator for help, based on a table of rules. If a question occurs, such as a break in a line or uninterpretable text, the software will prompt the operator for help. At its basic level, this may be a simple line-following function. More advanced systems may read additional graphic types, like polygons. They may also open data entry screens to let the operator add additional information or link a particular feature to a related attribute database record.

With this semi-automated approach, both graphics and attribute data can be captured and linked in a streamlined, rules-based environment that reduces the time and increases the accuracy. However, operators remain central to the conversion process. They determine what elements should be converted and when. They continually monitor the process and intervene whenever the system encounters a question or discrepancy it cannot handle based on its limited rules base.

Batch Conversion.

Most recently, batch conversion software has emerged that enables users to pre-load a series of drawings, capture the graphics and populate the associated database records unattended. In some cases, even randomly rotated and hand-drawn text and sentence strings can be read and entered into the appropriate data fields.

Depending upon the nature of the documents, these tools may improve conversion productivity by as much as 40 to 90 percent. It is important to note, however, that interactive vectorization or batch conversion systems are not off-the-shelf solutions. Both will require substantial customization in order meet the specific needs of a given organization. A complete rules base must be developed based on the user requirements and the nature and condition of the original drawings and the types of scans performed.

Raster Analysis

Driven by the availability of low cost aerial and satellite imagery. R&D teams are working on raster-based analysis tools that will eliminate the need to vectorize much of the data now being converted. While most of these tools are still in development, it is already possible for software to "read" an aerial photo and identify roads, railways and building outlines. Spectral analysis can interpret types of vegetation, determine irrigation inconsistencies and identify diseases or fungus in crops long before they are visible to the human eye.

These uses are just the beginning. Expect to see sophisticated raster-based analysis tools in the next several years to support virtually any potential use for digital imaging. The technology is driven by simple economic logic. As images become more detailed and refined, it requires better tools to make use of it. As more users accept the product, the more its price declines and user interest increases.

CONCLUSION

The value of scanners for digital mapping has increased substantially in the last two years, as the technology has finally begun to mature. Scanning systems now provide all the critical components to deliver on their promises of fast, accurate input of usable data - with the emphasis on usable.

Remarkable advances in the support software continue to push the envelope of what can be done with raster data. Traditionally raster information was considered "dumb" data, compared to "intelligent" vector data, which can be analyzed and manipulated to create new information. But new software tools are making it possible to add an interpretive element to these dumb pictures, which can, in itself, be used in analysis. In fact, every serious CAD or GIS software package available today provides a means to use both raster and vector data in a common environment.

On the hardware side, the migration of high-end graphic functionality from the workstation to the desktop has put the ability to use high-accuracy, complex images, like aerial photography, in the hands of more and more users.

At the end of the day, large format scanning makes its biggest contribution to digital mapping by enabling systems to become productive faster. Time to productivity has become the measure of success in today's rapid-deployment computer environments.

Organizations implement digital mapping system to improve productivity, reduce operating costs and increase access to geospatial information. The faster information can be loaded into the digital environment and made available in a usable format, the sooner these organizations can accomplish their objectives. To date, the best input device for achieving productivity remains the large format scanner.

About the Author: Jeremy D. Wilson is freelance writer and analyst in Aurora, Colorado. He specializes in the business-case and marketing aspects of geotechnologies and other high-tech industries, particularly related to the operations of infrastructure-intensive organizations.

This white paper is provided courtesy of VIDAR Systems Corporation, Herndon, Virginia.

Copyright Open Archive Systems, Inc. All Rights Reserved.