1 FUNDAMENTALS OF GPS

The explosion in interest in GIS as a management toolhas been accompanied by the development of anumber of enabling technologies, one of the moreimportant of which is the Global Positioning System(GPS). While GIS technology offers tremendouscapabilities for more informed management decision-making, rendering competent decisions still dependson having reliable data. In order to realise the benefitsof GPS, and not to mis-apply the technology, it isimportant to understand its limitations. This chapterdescribes how GPS works and how to obtain reliabledata using it. Accuracy (in the surveying sense) ofGPS data is an important consideration. The first partof this chapter covers how the GPS accuracy isobtained and the second part discusses how to useGPS for GIS data capture. The term ‘GPS’ is usedinterchangeably here to refer to satellite-basednavigation systems in general and the ground-basedgeographical data collection instruments in particular.

1.1 Overview of GPS

The Navigation Satellite Timing And RangingGlobal Positioning System, or NAVSTAR GPS, is a

satellite-based radio-navigation system that iscapable of providing extremely accurate worldwide,24-hour, 3-dimensional (latitude, longitude, andelevation) location data (Wells 1987). It is one of twosuch satellite systems, although its Russiancounterpart, GLONASS, will not be discussed here.The system was designed and is maintained by theUS Department of Defense (DoD) as an accurate,all-weather navigation system. Though designed as amilitary system, it is available with certainrestrictions to civilians. The system has reached itsfull operational capability, with a complete set of atleast 24 satellites orbiting the Earth in a carefullydesigned pattern.

1.1.1 The fundamental components of GPSThe NAVSTAR GPS has three basic segments:space, control, and user. The space segment consistsof the orbiting satellites making up the constellation.This constellation is comprised of 24 satellites, eachorbiting at an altitude of approximately 20 000kilometres, in one of six orbital planes inclined55 degrees relative to the Earth’s equator. Eachsatellite broadcasts a unique coded signal, known aspseudo random noise (PRN) code, that enables GPSreceivers to identify the satellites from which the

467

33Using GPS for GIS data capture

A F LANGE AND C GILBERT

Depending on the particular equipment utilised and the techniques used, Global PositioningSystems (GPS) are capable of recording position to a high level of accuracy. Differential GPSis required to obtain the accuracy required by many GIS data capture applications.Differential GPS requires the use of a base station at a known location to remove systematicerrors from the GPS signal. Modern GPS/GIS data collection devices have the capability tocollect a wide range of GIS attribute data in addition to position of the feature of interest;attributes include point, line, and area features. Since GPS can be used in many diverse GISdata capture applications (including applications as diverse as road centreline mapping,utility pole mapping, and wetland boundary mapping), it is important to understand thelimitations of GPS in order to derive full benefits from the technology.

signals are coming (Hurn 1989). The satellitesbroadcast the PRN codes as modulation on twocarrier frequencies, L1 and L2. The L1 frequency is1575.5 MHz and the L2 frequency is 1227.6 MHz.

With 24 satellites in the constellation, and thedesign of the orbits and the spacing of the satellitesin the orbital planes, most users will have six or moresatellites available at all times. There are a number ofdifferent programs used to plot the satelliteavailability for any geographical location. Theuniversal reference locator (URL) for a World WideWeb (WWW) site with a GPS satellite visibilityprogram on-line is http://www.trimble.com/satviz.

The control segment, under the United StatesDepartment of Defense’s direction, oversees thebuilding, launching, orbital positioning, andmonitoring of the system. It provides two classes ofGPS service: Standard Positioning Service (SPS) andPrecise Positioning Service (PPS). Monitoring andground control stations, located around the globenear the equator, constantly monitor theperformance of each satellite and the constellationas a whole. A master control station updates theinformation component of the GPS signal withsatellite ephemeris data and other messages to theusers. This information is then decoded by thereceiver and used in the positioning process. Of thetwo classes of service, the PPS is only available toauthorised government users while the SPS isavailable for civilian use.

The user segment is comprised of all of the usersmaking observations with GPS receivers. Thecivilian GPS user community has increased

dramatically in recent years, because of theemergence of low-cost portable GPS receivers(Figure 1 and Plate 23) and the ever-expanding areasof applications in which GPS has been found to beuseful. Such applications include surveying,mapping, agriculture, navigation, and vehicletracking (Figure 2; Plate 24). The civilian users ofGPS greatly outnumber the military users.

A F Lange and C Gilbert

468

Fig 1. A hand-held 12-channel GPS receiver.

Fig 2. Off-road navigation using GPS.

1.1.2 The limitations of GPSAlthough in theory GPS can provide worldwide, 3-dimensional positions, 24 hours a day, in any type ofweather, the system does have some limitations. First,there must be a (relatively) clear ‘line of sight’ betweenthe receiver’s antenna and several orbiting satellites.Anything shielding the antenna from a satellite canpotentially weaken the satellite’s signal to such adegree that it becomes too difficult to make reliablepositioning. As a rule of thumb, an obstruction thatcan block sunlight can effectively block GPS signals.Buildings, trees, overpasses, and other obstructionsthat block the line of sight between the satellite andthe observer (GPS antenna) all make it impossible towork with GPS. Urban areas are especially affected bythese types of difficulties. Bouncing of the signal offnearby objects or the ground may create anotherproblem called multi-path interference. Multi-pathinterference is caused by the inability of the receiver todistinguish between the signal coming directly fromthe satellite and the ‘echo’ signal that reaches thereceiver indirectly. In areas that possess these type ofcharacteristics, as in the urban core of a large city,inertial navigation techniques must be used tocomplement GPS positioning.

The receiver must receive signals from at least foursatellites in order to be able to make reliable positionmeasurements. In addition, this number of satellitesmust be in a favourable geometrical arrangement,relatively spread out. In areas with an open view ofthe sky, this will almost always be the case because ofthe way the satellite orbits were designed. Thedilution of precision, or DOP, is a measure of thegeometry of the GPS satellites. When the satellitesare spaced well apart in the sky, the GPS positiondata will be more accurate than when the satellitesare all in a straight line or grouped closely together.

An additional limitation of GPS is the DoDpolicy of selective availability (SA). This policylimits the full autonomous accuracy only to officialgovernment users. This policy, and methods used toameliorate its effects, are described in section 1.1.6.A Presidential Directive of May 1996 has declaredthat the policy of SA will likely be discontinuedwithin a period of more than four years and lessthan ten years. However, even if SA is eliminated,GPS will not be accurate enough for many mappingand GIS data capture applications.

1.1.3 How a GPS receiver calculates positionThe position of a point is determined by measuringthe distance from the GPS antenna to three satellites

for a 2-dimensional reading (latitude and longitude),and at least four satellites for a 3-dimensionalposition (latitude, longitude, and altitude). The GPSreceiver ‘knows’ where each of the satellites is at theinstant in which the distance was measured. Thesedistances will intersect at only one point, the positionof the GPS antenna. How does the receiver ‘know’the position of the satellites? Well, this informationcomes from the broadcast orbit data that are receivedwhen the GPS receiver is turned on. The GPS receiverperforms the necessary mathematical calculations,then displays and/or stores the position, along withany other descriptive information entered by theoperator from the keyboard.

The way in which a GPS receiver determinesdistances to the satellites depends on the type ofGPS receiver. There are two broad classes: codebased and carrier phase based.

1.1.4 Code-based receiversAlthough less accurate than their carrier phasecounterparts, code-based receivers have gainedwidespread appeal for applications such as GIS datacapture. Their popularity stems mainly from theirrelatively low cost, portability, and ease of use.

Code-based receivers use the speed of light andthe time interval that it takes for a signal to travelfrom the satellite to the receiver to compute thedistance to the satellites. The time interval isdetermined by comparing the time in which aspecific part of the coded signal left the satellite withthe time it arrived at the antenna. The time intervalis translated to a distance by multiplying the intervalby the speed of light. Position fixes are made by thereceiver roughly every second, and the GPS receiversdesigned for use in GIS data capture applicationsenable the user to store the positions in a file thatcan be downloaded to a computer for post-processing or analysis.

Under normal circumstances, autonomousposition fixes made by code-based receivers wouldbe accurate to within 25 metres. This is the specifiedaccuracy for the SPS. The DoD, however, beganimposing its selective availability policy in July 1992,and this limits position fix accuracy to within100 metres. The purpose of selective availability is todeny potential hostile forces accurate positioningcapabilities. Military P(Y)-code receivers are notaffected by SA, and are not available to the generalpublic. In order to ameliorate these limits in

Using GPS for GIS data capture

469

positioning accuracy, differential GPS (DGPS)techniques have been developed. DGPS enables theuser to improve on the SPS, and to remove theeffects of SA and some other sources of error. Thesedifferential correction techniques can producepositions generally accurate to within a few metres.If and when SA is eliminated the accuracy of GPSwill be more accurate than SPS specifications, toabout 10 to 15 metres. This is not good enough formost GIS data capture applications and differentialGPS will still be required.

There are also now code-based receivers capableof sub-metre differential accuracy. Some sub-metrereceivers require longer data collection times (up to10 minutes). They perform best under veryfavourable satellite geometry, and with anunobstructed view of the sky. Some newer receiverscan provide a sub-metre accurate position eachsecond. These receivers cost more at the outset, butprovide a good return on the added investment byfacilitating substantially higher productivity.

It is very important that users of code-basedreceivers understand the positional accuracylimitations of the receiver. Because of SA, eachcoordinate viewed on a non-differential GPSreceiver’s display is only accurate to within 100metres. This accuracy may be improved on bytaking an average of 200 or so repeated positionobservations of the same point. The resultingaccuracy would nevertheless still be below whatmany users would consider acceptable quality.In order to produce acceptable results, GPSdata collected in the field must be differentiallycorrected either in real-time, or by post-processingof the data.

1.1.5 Carrier phase receiversCarrier phase receivers, used extensively in geodeticcontrol and precise survey applications, are capableof sub-centimetre differential accuracy. Thesereceivers calculate distances to visible satellites bydetermining the number of whole wavelengths andthe partial wavelength. Once the number ofwavelengths is known, the range may be calculatedby multiplying by the wavelength of the carriersignal. It is then a straightforward (but cumbersome)task to compute a baseline distance and azimuthbetween any pair of receivers operatingsimultaneously. With one receiver placed on a point

with precisely known latitude, longitude, andelevation, the calculated baseline can be used todetermine the coordinates of the other point(Leick 1990).

The relative cost of the carrier phase receivers ishigh, but technological advances have made thedual frequency (using both L1 and L2) carrierphase receivers much more efficient than the singlefrequency (using L1 only) receivers that were state-of-the-art only a few years ago. Dual frequencyreceivers allow for correction of ionosphere delays,and are therefore inherently more accurate atlonger baseline lengths over 50 km. With some ofthe most recent dual frequency receivers veryprecise measurements (± 1 cm) can be made in real-time. These receivers are used in machine controlapplications requiring a high degree of accuracy.

1.1.6 DGPSDifferential GPS can be employed to eliminate theerror introduced by SA and other systematic errors.DGPS requires the existence of a base station, whichis a GPS receiver collecting measurements at aknown latitude, longitude, and elevation (Hurn1993). The base station’s antenna location must beknown precisely. The base station may storemeasurements (for post-processed DGPS) orbroadcast corrections over a radio frequency (forreal-time DGPS), or both.

The assumption made with the base stationconcept is that errors affecting the measurements ofa particular GPS receiver will equally affect otherGPS receivers within a radius of several hundredkilometres. If the differences between the basestation’s known location and the base station’slocations as calculated by GPS can be determined,those differences can be applied to data collectedsimultaneously by receivers in the field (Figure 3).These differences can be applied in real-time(especially applicable for accurate navigation) if theGPS receiver is linked to a radio receiver designedto receive the broadcast corrections. In some GISmapping applications, these differences are appliedin a post-processing step after the collected fielddata has been downloaded to a computer running aGPS processing software package. GPS processingsoftware is typically integrated with GPS hardwareand thus is provided by the receiver manufacturer.As a rule, post-processed DGPS is consideredslightly more accurate than real-time DGPS.

A F Lange and C Gilbert

470

1.1.7 Base stations – the source for reference data for DGPSThere are many permanent GPS base stationscurrently up and running in the USA that can providethe data necessary for post-processing differentiallycorrecting positions to the users of code-basedreceivers over an electronic bulletin board.

In many parts of the world a real-time DGPSbeacon system is operational. These stations are partof a large network of coastal and river valleystations. These beacons broadcast in the frequencyrange of 285 to 325 kHz. The range of many ofthese stations is 300 to 500 kilometres. These stationscan provide differential accuracy in the 1-metrerange, depending on distance from the station andthe quality of the roving receiver.

Carrier phase base differential processing, forcentimetre accuracy, does not have the range ofcode-based differential GPS. To obtain centimetreaccuracy, it is now necessary to have the carrier basestation within approximately 10 kilometres for real-time carrier differential (sometimes called real-timekinematic) operation. Extending the range of carrierdifferential beyond 10 kilometres is an active GPSresearch topic.

1.2 GIS data capture considerations

Several key issues need to be explored whenconsidering whether GPS is an appropriate tool forcapturing coordinate data for a GIS database. Firstand foremost is the need to determine the positionaccuracy requirements. If the data will be used for

site-specific analysis that requires position accuracyto be within a metre, high-quality code-baseddifferential GPS receivers will be necessary. If stillhigher accuracy is required, of the order of10 centimetres or better – as for property boundarymapping – then carrier phase differential GPStechniques will be required.

Every GIS database must be referenced to a basemap or base data layer, and the reference datum ofthe various data layers must be the same. Ideally, thedatabase should be referenced to a very accuratebase map. If, however, the base map is 1:125 000 orsmaller, there could be problems when attempting toview the true spatial relationships between featuresdigitised from such a map and features whosecoordinates were captured with GPS. This can be areal problem if the GIS analyst decides to use aparticular GIS data layer that was originallygenerated using small-scale base maps as a base towhich all new generated data are referenced (seeWeibel and Dutton, Chapter 10). The best way toavoid such incompatibility is to develop an accuratebase data layer, based on geodetic control andphotogrammetric mapping.

Map datumUnderstanding the concept of a map datum isimportant if useful results are to be obtained fromany GPS mapping exercise. A datum is amathematical model of the Earth over some area.GPS, being a worldwide system, has a datumapplicable over the whole earth. This is called the

Using GPS for GIS data capture

471

Fig 3. Use of base stations for precise global positioning.

Base station antennapositioned over known point

Post Processing

300 km

GPS PC

Base station

Remotestations

Basefile

Remotefile

Correctedremote

file

PFINDERdifferentialcorrection

module

World Geodetic System, 1984 (WGS-84). There aremany local datums like European Datum 1950 (ED-50), North America Datum 1927 (NAD-27), andNorth America Datum 1983 (NAD-83) that havebeen used to make local maps. A particular point onthe surface of the Earth will have different latitudeand longitude coordinates depending on thereference datum. For GIS applications, it isrecommended that all data be collected anddisplayed in the most up-to-date datum available. Inthe US this datum is NAD-83. Fortunately for usersin most areas of the world, region-specific datumsare quite similar to the GPS natural datum, WGS-84. It is important to use the same datum for all datalayers if the data are going to be overlaid in a GIS.

2 USING GPS WITH GIS

There is a range of ways in which GPS is usedalongside GIS. First, GPS may be used to identify orrefine the geographical coordinates associated withsatellite imagery. GPS is used to reduce distortionsand to improve the positional accuracy of theseimages. When three or more distinctive points (themore the better) can be located both on a satelliteimage and on the ground, GPS receivers can be usedto collect accurate geographical coordinates at theselocations. The rest of the image can then be adjustedso that it provides a better match to the real-worldcoordinates. Second, GPS can be used in the groundtruthing of satellite images. When a particularsatellite image has a region of unusual orunrecognised reflectivity or backscatter, thecoordinates of that region can be loaded into a GPSreceiver. The GPS receiver will then enable the userto navigate directly to the area of interest so that thenature of the vegetation or the Earth’s surface can beexamined. With real-time differential GPS, users caneven navigate directly to an individual 1m pixel.Third, GPS has developed into a cost effective toolfor updating GIS or computer-aided design (CAD)systems. Using GPS to collect data is analogous todigitising a map by moving a mouse or digitisingpuck over a map. The users of GPS equipmentsimply move along the surface of the Earth and thegeographical coordinates of where they went arecomputed and stored as a continuous series.

However, GPS technology is better suited to somedata collection applications than others. Forexample, GPS is not particularly well suited torecording the corner stones of skyscrapers because

the building itself blocks a large percentage of thesky, and hence the GPS signal. Traditional surveytechniques would probably be more efficient. GPS isnot necessarily the most efficient tool for evaluatingsubtle gradational changes in vegetation across alarge region. Satellite imagery might be better suitedto this task. However, GPS is an excellent tool fordata collection in many environments where the usercan generally see the sky and is able to get close tothe objects to be mapped.

2.1 How is GPS used for GIS attribute collection?

The typical GPS-based data capture tool is a GPSreceiver combined with a hand-held computer. Thesetwo components may be connected by a cable, orthey may be combined as a single, integrated, hand-held unit. The hand-held computer is used forattribute entry and display of data to the user. Thereare several variations on this theme, including theuse of pen computers, laptop computers, and small,hand-held units. The design goal is that the user canfocus on observing and entering attribute data, sincethe position data ‘just happens’ automatically.Figure 4 illustrates how a bar-coded system may beused to record the locations of utility conduits.

Mapping roads is an important application. Bydriving the road once, users can record the locationof the centreline and both kerbs by noting the offsetbetween the line traversed by their GPS antenna andthe linear feature of interest, such as a kerb (seeWorboys, Chapter 59). Additionally, they can recorda number of attributes such as name, width,condition, materials. The amount of attribute datathat can be collected is limited only by how fast theoperator (or accompanying colleague) can enterthem while driving past. The concepts employed inroad mapping can be extended to any type of linearor polygonal feature, such as shorelines, paths, forestfire boundaries, land usage zones, vegetationboundaries, powerlines, and utility right-of-ways: seePlate 24. Presently there are many consulting firmsthat use GPS to record the attributes and location ofthousands of kilometres of road every month.

Recording the location and attributes of point-likefeatures is another very popular application of GPS.To record a point-like feature with GPS, the user willtypically walk or drive up to the object to be mapped(such as a pole, storm drain, or soil sample site) andthen place the GPS antenna on that location ofinterest. While the user is busy for a few moments

A F Lange and C Gilbert

472

observing and entering attribute data, the GPSreceiver is automatically recording position data.

Unlike the hand digitising of a paper map, thereis no ‘scale’ directly associated with this form of‘GPS digitising’. If the user is doing detailed, highaccuracy work, it might make sense to record astorm drain as a small polygon (remember thebest GPS receivers are accurate to better than1-centimetre). On the other hand, if the data will beused to populate a crude map, the user might chooseto define an entire airport as a point feature. Usersare free to define their features as required to best fittheir application. GPS is readily used to record theattributes and location of street furniture (such aslamp posts, telephone poles, storm drains, drainageaperture covers, valves, etc.). It is not uncommonthat some users see their productivity increase froma typical 100 to 200 points/day before GPS to over1000 points per day after GPS. The increasedproductivity allowed by GPS can be stunning.

The use of GPS can be extended to mappingobjects which are difficult or dangerous to occupy.By combining the use of a hand-held laserrangefinder with a GPS, users are able to record thelocation of an object even if they cannot physicallyoccupy that location: see Plate 25. The user canremain in a location that is either convenient ormore conducive to good GPS signals, then measurethe distance and direction (an offset) to the object ofinterest with a laser rangefinder. In some GPS, thelaser can be connected electronically so that the laser

measurements are transmitted directly to the GPSdata collection system; alternatively, the user couldenter the offset information by hand. The location ofthe GPS antenna and the offset measured by thelaser are automatically combined to compute anaccurate location for the object of interest.

There are many other types of electronic sensingdevices that can be used in conjunction with a GPSreceiver. Typically when sensor data are coupledwith GPS positions, the user has the intention oflater presenting these data in the form of a thematicmap. The contours of these maps may representinformation as diverse as radiation levels (isoradmaps: Runyon et al 1994), temperature (isothermmaps), water depth (bathometry: Chisholm andLenthall 1991), signal strength for cellulartelephones, or even biological density.

Some features of modern GPS data collectionsystems make them extremely effective tools forattribute collection. One common (and quiteeffective) way that GPS users prepare for datacollection is to predefine a customised list of thefeatures and attributes that are to be collected in thefield. This list is then transferred into the GPS datacollection system so that, in the field, the user will beprompted by the GPS for the appropriate attributes ateach location. For example, if the user approaches atelephone pole in the field, the first step is to select theitem ‘telephone pole’ from the screen of the GPS.After having selected the feature name, the GPSwould then automatically prompt the user for

Using GPS for GIS data capture

473

Fig 4. Use of GPS in utilities mapping.

whichever attribute details are appropriate totelephone poles. These attributes (such as material,condition, identification number, etc.) are taken froma list that was created by the user, are displayed at theappropriate time and in the appropriate sequence, andmatch the structure of the target database.

When attribute data are collected, they shouldalways be stored in the appropriate form.Alphanumeric and numeric data are stored in theappropriate field types, and numeric values arevalidated so that they are within predefinedmaximum and minimum values. Character stringsare automatically checked so that they do not exceedthe maximum allowable field length of the targetdatabase. Whenever appropriate, precise date andtime attributes are generated by the GPS, andmanually entered date and time values are checkedso that they fall within valid ranges. When theattribute is an image, perhaps from an attacheddigital camera, the data collector is able to store thefilename of the image as an attribute value. Theattribute values that are collected may be edited atany time after they have been collected – either in thefield or during subsequent processing. Collectively,features like the ones described above will greatlyreduce field data collection errors.

GPS also feature pause and resume functionswhich, although apparently simple, are neverthelessinvaluable in the field (Gilbert 1994a). Imagine whatthe path you digitised would look like if you couldnot momentarily pause the data collection when yourhat blew off! When recording roads, many users findthat they wish to interrupt data collection every timethey stop at a traffic signal or a gate. A problemoccurs when a GPS receiver is stationary andcontinues to record positions once per second, sincethe resulting road features contain periodic, puzzlingclumps of tangled vertices in the roadway.

The ability to repeat previous attribute values isanother function which is very simple, yet is oftencrucial to efficient data collection. After entering 200attributes for the previous utility pole you will surelyappreciate the ability to press one button and havemost or all of your attributes transferred to the nextpole. The repeat function can also be very usefulwhen recording linear features that containdynamically changing attribute values. For example,on approaching the end of a block of buildings, auser may wish to end one segment of the road andimmediately begin another segment that shares acommon node. A well-designed GPS data collector

will allow the user to terminate the first segment andinitiate the second while retaining most, or all, of thesame attribute values for both nodes – with as littleas one keystroke.

Many GPS receivers are unable to collect accuratecoordinates effectively while they are moving. This isnot a limitation of the satellite GPS themselves, as thebest GPS receivers can provide centimetre level resultswhile moving at high speeds. This limitation is simplya by-product of some of the receiver designs andsoftware products on the market. It is sometimesdesirable to record point features while moving. Onevery powerful data collection function allows users tostore point features accurately while driving or flyingover the point of interest. If it is not possible to traveldirectly over the point of interest, some GPS will evenallow the user to define an offset distance from theGPS antenna to the object of interest. Typically, thedirection of the offset is defined as being either to theleft or the right, and orthogonal to the direction oftravel. Depending on the variability of the attributedata, such a data collection system can allow a skilleduser to record thousands of point features per day. Itis important to recognise, however, that such rapidrecording is often less accurate than even a onesecond static occupation – most commonly becauseof user errors in pressing the button at the precisemoment the feature is being passed.

2.2 Exporting GPS data to a GIS, CAD system,or database

The final stage of the process of GPS-based datacollection for GIS is to transfer the data from thefield device to the target database. In the interests ofpreserving valuable memory, most GPS store datainternally in their own proprietary formats. Thisdata transfer is most often accomplished by runninga translation program that will convert the data fromthe compact internal storage format to the databaseinterchange format of the user’s choice.

At the most basic level, this task can beaccomplished in a very simplistic manner – that is,simply dumping the coordinates and attributes in acomma delimited file. However, most users prefer toemploy a translation program that will allow them tomodify slightly the output data, so that it matchesthe characteristics of the target database as closelyas possible (Gilbert 1994b). Some of the moreimportant considerations of a translation programare as follows.

A F Lange and C Gilbert

474

2.2.1 Creation of metadataA useful feature of a data translation program is theability to generate metadata (Guptill, Chapter 49;Goodchild and Longley, Chapter 40). The resultingoutput file(s) will contain not only the attributevalues that were entered in the field by the user, butalso a variety of additional attribute data that aregenerated by the translation program (Gilbert 1996).Some of these generated attributes are common tothe entire dataset, such as the receiver type used tocollect the data or the name of the original data file.However, the most useful of the generated attributesare specific to the individual features. Severalexamples of GPS metadata are described in thefollowing paragraphs.

Quality information about the accuracy of theGPS coordinates is very valuable (see Fisher,Chapter 13). This may take the form of a singlequantitative accuracy estimate, or it may bepresented as a variety of accuracy indicators thatcan be used individually or together to infer dataquality (Veregin, Chapter 12). Although a singleaccuracy estimate is simple to use, it may be lessreliable. The user must develop an understanding ofwhether any particular manufacturer’s algorithm isreliable enough. Additionally, the user often has noway to tell how the GPS accuracy value wascomputed and whether it represents a probabilityof 99 per cent, 95 per cent, 68 per cent, or even aslittle as 50 per cent. There is also the uncertainty asto whether the reported value refers to horizontalaccuracy or 3-dimensional accuracy. It is notuncommon for the horizontal and vertical accuracyof a GPS coordinate to differ by a factor of two.There is no industry standard on to how the GPSreport spatial accuracy (see Smith and Rhind,Chapter 47, for a discussion of industry spatialdata standards).

Alternatively, the translation software mayreport a variety of accuracy indicators that can beused individually or together to infer data quality.Such indicators may include occupation time,number of satellites used, maximum dilution ofprecision, quantity of GPS data collected, or thetype of differential processing that was performed,if any. Although this is more complex, it doesprovide the user with greater understanding andflexibility. When these indicators are availableindividually as metadata, the users are free toestablish whatever data quality rules they desire fortheir applications. Some agencies will define

multiple quality assurance guidelines that varydepending upon the intended use of the data.Storing multiple accuracy indicators has the addedbenefit that they can be used to verify that thefield operator adhered to the appropriate datacollection procedures.

2.2.2 Creation of macro filesThe transformation of GPS data to anappropriate format can be greatly enhanced by theautomatic creation of macro or batch files. Whilethe data are being translated to the desired format,some translation programs will read and computethe information required to create a set ofcommand scripts. These scripts are unique to thetranslated data and are used to complete theimport process quickly and easily. They canautomatically create new layers and tables (orappend existing ones), automatically associateattribute values with the coordinates, or evencreate additional data such as labels positioned atthe centroid of a polygon feature.

2.2.3 Geodetic datum transformationAs stated in section 1.2, the native geodetic datumof GPS data is WGS-84. Programs that translateGPS data into GIS or CAD compatible formatsare usually able to transform the coordinates tothe projection, coordinate system, and datum ofthe user’s choice. Most of these translationutilities include a long list of datums andcoordinate systems. The more sophisticatedprograms will also allow the user to create acustomised coordinate system.

2.2.4 Configurable ASCII data translationThere are many GIS and CAD compatible importformats. Despite the abundance of formats, GPSmanufacturers have discovered that some of theGIS community utilise home-grown, customisedimport formats. Therefore, even a translationprogram that supports the dozen most popularformats will seem incomplete to some users. Themainstream GPS data collection systems are allequipped with powerful and flexible configurableASCII export programs. Such programs give usersthe ability to translate their GPS data intocustomised formats that the GPS manufacturerscould never have anticipated.

Using GPS for GIS data capture

475

SUMMARY

Depending on the particular equipment utilised andthe techniques used, GPS is capable of a wide range of accuracy, from tens of metres to centimetres.Differential GPS techniques are required to obtainthe accuracy required by many GIS data captureapplications. Differential GPS requires the use of abase station at a known location to removesystematic errors from the GPS signal. ModernGPS/GIS data collection devices have the capabilityto collect a wide range of GIS attribute data inaddition to the position of the feature of interest,including point, line, and area features. GPS can beused in many GIS data capture applications,including applications as diverse as road centrelinemapping, utility pole mapping, and wetlandboundary mapping. It is nevertheless important tounderstand the limitations of GPS to derive fullbenefits from the technology.

References

Chisholm G, Lenthall P 1991 Combining GPS and echosounders to monitor river siltation. GISWorld 5(2): 46–8

Gilbert C 1994a Attribute collection with GPS part 1: Varietyis the spice of attributes. Mapping Awareness 8(9): 28–30

Gilbert C 1994b Attribute collection with GPS part 2: Assessingease of use in the field. Mapping Awareness 8(10): 30–2

Gilbert C 1996 The software with your GPS Part 1 – dataexport. Mapping Awareness 10(8): 32–4

Hurn J 1989 GPS: A guide to the next utility. Sunnyvale,Trimble Navigation Ltd

Hurn J 1993 Differential GPS explained. Sunnyvale, TrimbleNavigation Ltd

Leick A 1990 GPS satellite surveying. New York, John Wiley& Sons Inc.

Runyon T, Hammitt R, Lindquist R 1994 Buried danger:integrating GIS and GPS to identify radiologicallycontaminated sites. Geo Info Systems 8(4):28–36

Wells D 1987 Guide to GPS positioning. Fredericton, CanadianGPS Associates

A F Lange and C Gilbert

476