| GIS: Keeping Abreast of Data Conversion Technologies Impacts the Bottom Line
A Look at Current and Future Technologies
By Nirav Shah Data conversion is the dirty work of GIS development. It can be time consuming, labor intensive and expensive. In fact, data conversion is often the most expensive component of a GIS implementation project, accounting for 40 to 80 percent of total development costs.
By their nature, some GIS projects require more conversion work than others. Natural resource, cadastral and planimetric GIS databases, for instance, usually require very specialized and accurate data that can only be created from some form of existing maps or legal descriptions. These must be converted into digital formats for use in the GIS, often at considerable cost.
Insurance, real estate, retail and other business geographics disciplines, on the other hand, primarily use TIGER files, census demographics and other commercial data.
Thankfully, data conversion technology is being improved all the time. Better hardware and software are resulting in more efficient conversion procedures. This translates into less time and less money, and often a better end product.
If you are planning to implement a GIS project in the near term or in the distant future, it is important for you to be familiar with the state of existing data conversion technology and of advancements that soon will be made. Taking Stock of Current Technology
There are currently four basic data conversion technologies available. Choosing the conversion method that best fits your situation is crucial because of the long-lasting effects it may have on the project.
Computer-Aided Drafting - It is rare, but sometimes organizations find themselves with no maps and very poor legal records defining street centerlines and property boundaries. In this case, they have little choice but to start from scratch and use CAD technology to create their digital maps, a time consuming, expensive method.
Heads-up Digitizing - This is another tried-and-true technology that may be attractive because it can be performed in house with a digitizing table and commercial software. But in comparison to other newer methods, it requires a lot of manual input which results in a high degree of human-induced error. The quality can be very good if the operators are skilled, but in the end it usually proves inefficient and costly.
Semi-Automated Raster-to-Vector Conversion - In this technique, paper maps are scanned to create a raster image file. Semi-automated software is then used to convert the rasters into vector files.
This method provides very high quality digital products, which can be expensive, but will be worth it for projects requiring a high degree of accuracy. Generally, this process should be considered if you have very good paper maps to work from.
Completely Automated Raster-to-Vector Conversion - Similar to the above mentioned method, maps are first scanned, but then the digital files are run through batch processing which converts every raster to vector in a fully automated manner. Operator interaction is minimal.
This is a down-and-dirty method because it can be done inexpensively with results adequate for simple GIS projects. If you want to assemble complete GIS coverage including attribute data, however, you will need to conduct considerable manual "cleaning up" of the vectors to include attributes such as street names and parcel numbers. Coming Attractions
Completely automated conversion could be voted "Most Likely To Succeed" because upgrades are constantly being made to commercial software. In the next couple of years, these packages will drastically improve in their ability to convert text and numbers such as street names and property identifications.
When this capability is enhanced, completely automated conversion will offer a high-quality, complete GIS coverage product at a reasonable price.
One new technique now under development is called Rapid Digital Conversion of Analog Maps. Although it is limited in applications, this process may prove one of the more accurate and inexpensive methods because it employs currently existing technology.
In this method, a color thematic map is photographed with a digital camera to create a raster image. Then standard image processing routines are applied to derive classification layers related to the map colors. In experimental projects, this technique has scored accuracies in the range of 93 percent. Its use, however, is limited to those disciplines that routinely display their data on thematic maps.
Another promising new technology is Video Tracing. It utilizes a video camera mounted on a scanner stand and a motor that moves the camera around a map. The map data is displayed in real time on a screen where a CAD operator assigns the various lines to appropriate GIS layers.
Also in its experimental phase, video tracing appears to be a very fast and efficient technology. Its quality is roughly equal to that of the semi-automated conversion process. The only downside to this process is that it requires the purchase of specialized hardware and software. It's a Dirty Job
As I said at the beginning, data conversion is a dirty job, but it has to be done. If you are going to the trouble of designing and implementing a GIS, it will pay to keep abreast of new advancements in this technology - or at least develop a relationship with a GIS vendor who does - because choosing the right technique can positively impact the quality and economics of your GIS. Free offer from the author - Nirav Shah is president of EI Technologies in Aurora, Colo., a full-service GIS consulting firm. Contact him to receive a free special guide, "What Everyone Needs to Know About GIS." He may be reached by phone at 303-750-8084, fax at 303-750-8488, and e-mail at gis-eitek.com. You may also visit the EI Technologies Web Site at http://www.eitek.com Back |
