Data Acquisition Techniques in GIS

Spatial Data Acquisition Techniques in GIS

A potential GIS user quickly realizes that having needed spatial information in digital form is critical. Without supporting spatial data, it is nearly impossible to conduct any meaningful analysis. Public data repositories (or government agencies), private vendors, collecting and organizing it yourself, or paying (or hiring) someone to collect and organize it are all options for obtaining the spatial information required for any project or study.

Existing GIS Data Sources

There are numerous existing digital spatial information sources that can be used in a GIS. Some are free or easily accessible to the general public, whereas others must be purchased from a private vendor. Census data  is perhaps the most common source in the United States, summarizing population characteristics of the nation, states, counties, cities, and towns at varying scales. Government agencies also produce data such as 100-year floodplains, vegetation classifications, and transportation. GIS vendors frequently include some basic data with their system, such as political boundaries of nations and states. There are ongoing efforts, such as Project Alexandria  and Geospatial One-Stop, to bring together freely available spatial information sources in digital libraries.

There are also numerous commercial spatial data providers. Some specialize in transportation data, such as Tele Atlas  and NAVTEQ , while others, such as Claritas, focus on geodemographic and market research (www.claritas.com). Given the importance of digital spatial information in the use of GIS, it is not surprising to see the data provider and service industry emerge and thrive, with annual sales currently exceeding $5 billion and expected to grow significantly in the coming years. As a result, commercial data providers profit from the sale and distribution of digital spatial information to users. The point is that while spatial data exists, it may be expensive to obtain.

Semi Existing GIS Data Sources

By semi existing sources, we mean that the spatial information is not always in a digital format that is GIS-compatible. For example, one might have a map from the mid-1800s that shows where gold has been discovered in California. Technically, the information about where gold was discovered exists, but it is contained on a paper map rather than as digital data. A spreadsheet of addresses indicating the residential locations of customers who have purchased a specific product is another example. Again, the information exists, and in this case, it is in digital format, but it is not suitable for use in GIS for geographical evaluation. However, in both cases, this information can be processed to create a digital form that can be used in a GIS. We will now look at three fundamental approaches to processing semi-existing data sources: scanning, digitizing, and geocoding.

Scanning is the process of converting a hardcopy map into a digital image. Most people are probably familiar with scanners, which are used to convert text on a page into digital form. A similar process is used for maps, in which the scanner is used to detect the presence of information on the map pixel by pixel (or raster cell by raster cell). The resulting scanned image is a digital version of the map that can be accessed in some way by most GIS.

Digitizing; the process of capturing or creating vector objects from hardcopy maps or other geographic information sources is known as digitizing (e.g., photographs and images). Manual digitizing relies on a piece of equipment known as a digitizing table, to which a cursor or puck is attached to trace points, lines, or polygons of interest on the map. When the geographic information source is in digital form, such as an image or photograph, another approach is heads-up (or on screen) digitizing. As an example, given a scanned map image, one could import the image into GIS and then manually digitize vector objects in the image.

Geocoding; the process of converting a street address to a latitude and longitude on the earth's surface is known as geocoding. A database with records of street segments, the geometry of each street segment, and the address ranges on each side of the street segment is required. Because the centerline of streets represents the geometry of street segments, this is commonly referred to as a street centerline database. If a street and address are not found in the database, the associated latitude and longitude of the address cannot be determined. When a street is found in the database and the address location is estimated, the latitude and longitude of that address are successfully geocoded. A point on the earth's surface can be found using latitude and longitude, and it corresponds to the street address. Most commercial GIS packages include geocoding functionality, and there are commercial vendors who specialize in geocoding services. However, because successful geocoding is dependent on the street centerline database used, problems can arise if this database is out of date, inaccurate, or of poor quality.

Surveying and Airborne Approaches

Surveying and airborne approaches are the two final data acquisition approaches to be discussed. These are grouped together because they are becoming more interconnected and/or interdependent. The following fundamental approaches to generating spatial information are now discussed: surveying, GPS, aerial photography, and remote sensing.

Surveying is a method of generating vector-based spatial data (points, lines, and polygons) by measuring angles and distances from known positional locations. The importance of known positional locations or reference points here cannot be overstated. Traditional surveying methods use transits and theodolites to measure angles and measuring tapes to determine distance. This requires the cooperation of two people. Total stations are increasingly being used to measure angles and distances in surveying due to advances in technology. Surveying, in general, ensures high positional accuracy—even down to the millimeter level in some cases. It is, however, a time-consuming approach.

GPS (global positioning system) is a satellite navigation system run by the United States Department of Defense that was originally intended for military use. It is a satellite constellation that orbits the Earth at a distance of about 20,000 kilometers. The satellites contain atomic clocks that transmit highly accurate radio signals that handheld or mounted receivers can read. This allows for the determination of position on the earth's surface, as well as velocity and time, assuming that the receiver is in view of a sufficient number of satellites. Vector data can thus be generated. A receiver, for example, could be used to locate a bus stop (point), record the route of the vehicle (line), or demarcate the catchment area of a watershed (polygon). Signal errors can be corrected with differential GPS, and positional accuracy to the centimeter level is possible. This is accomplished by using ground reference stations to adjust GPS readings.

Aerial photography is done from above the earth's surface, possibly in a hot air balloon, plane, or helicopter. This results in a digital image (or possibly a photograph that is subsequently scanned into a digital image). A georeferenced digital image can be used to derive features or attributes on the earth's surface. Heads-up digitizing, for example, could be used to create vector objects such as roads, lakes, rivers, buildings, fields, or forests. Positional accuracy can often be achieved to the fraction of a meter.

Remote sensing is commonly used to generate raster-based spatial data. Sensors mounted on satellites specifically measure solar energy (electromagnetic radiation), though sensors can also be mounted on planes or helicopters. This allows for the deriving of physical, chemical, and biological properties on or near the earth's surface, but it necessitates the processing and interpretation of sensor readings. Spatial and temporal resolution can vary significantly, with some platforms producing measurements for a raster cell of a few meters or less in size and others producing measurements for a raster cell of up to 10 km or more in size for an individual cell.