Author: Susanne Hansen, Denmark
ABSTRACT
In 1995 The National Survey and Cadastre of Denmark launched a Global Positioning System (GPS) projekt with objective to make GPS operational for cadastral surveys.
A network of 87 ground control points has been established in a 4000 km² test area. The distance between neighbouring control points is less than 10 km. The control points - coordinated in x, y, z - are used as GPS reference stations. Comprehensive tests made it clear, that the density of the network (one control point pr. each 45 km²) is sufficient to ensure a total Real Time Kinematic coverage which is stable enough for cadastral surveys.
On this background it was decided to implement the results of the GPS project. The time scale of the implementation phase is, that GPS should be in operation at regional cadastral offices in Denmark by the end of 1997.
This paper describes the methods developed during the project. The impact of GPS on the data quality is evaluated. And the efficiency of GPS applied in cadastral surveys is briefly discussed.
ZUSAMMENFASSUNG
1995 setzte das Bundesvermessungsamt in Dänemark das Projekt Global Positioning System (GPS) in Gang mit der Zielsetzung, GPS für Kataster-Vermessungen nutzbar zu machen.
In einem 4000 km2 großen Testareal wurde ein Netz aus 87 trigonometrischen Festpunkten festgelegt. Der Abstand zwischen zwei benachbarten Festpunkten beträgt weniger als 10 km. Die Festpunkte haben x-, y- und z-Koordinaten und dienen als Referenzstation bei der GPS-Messung. Umfassende Tests haben ergeben, daß die Dichte des Festpunktfeldes (ein Festpunkt pro 45 km2) ausreichend ist, um eine deckende totale RTK zu sichern, die verläßlich für Kataster-Vermessungen ist.
Auf diesem Hintergrund wurde entschieden, die Ergebnisse des GPS-Projektes umzusetzen. Der zeitliche Rahmen der Implementierungsphase sieht vor, daß bis Ende 1997 GPS in den regionalen dänischen Katasterämtern zur Anwendung kommt.
Dieses Referat beschreibt die Methoden, die während des Projekts entwickelt worden sind. Der Einfluß von GPS auf die Datenqualität ist untersucht worden. Die Effektivität der GPS-Anwendung bei Kataster-Vermessungen wird kurz umrissen.
GPS APPLIED IN CADASTRAL SURVEYS.
In Denmark cadastral surveys are performed by the licensed surveyors in private practise, except in the southern part of Jutland, where local offices of the National Survey and Cadastre perform all cadastral surveys. For historical reasons all parcelling out has been surveyed, prepared and approved by these local offices. This project is developed by the South Jutland Branch of the National Survey and Cadastre.
The southern part of Jutland has been under the Prussia from 1864 to 1920. That’s the reason why the cadastral maps and the cadastral coodinate system are Prussian. The coordinate system is named the system Ostenfeld. In connexion with digitizing the cadastral maps it was decided to use the national grid: the system 34. Before the maps were digitized 1,200 system Ostenfeld control points were cooordinated in the system 34. For each of the 23 municipalities a transformation was made to convert the maps from the system Ostenfeld to the system 34. The accuracies of the transformations are between 25 to 35 cm, and this accuracy is sufficient for the digital cadastral maps. At the end of 1997 all the cadastral maps in Denmark have become digital maps (Hansen 1995).
The number of control points was not sufficient to support the cadastral work for the future in the system 34. So we needed to establish new control points in the system 34. With the same density of the control points as we have in the other part of Denmark we should have established 35,000 control points.
In the light of the technological development in the field of Global Positioning System (GPS), we found it out of date to do this establishment, because it is possible with the use of GPS to make measurements at large distances. The original projekt was changed in 1994. The purpose of the new project became to examine the possibilities of doing GPS operational for cadastral surveys (Ryttersgaard 1995).
THE USE OF GPS AT THE SOUTH JUTLAND BRANCH
At the South Jutland Branch the GPS methods Fast Static and Static have been used to coordinate control points since 1991. In 1994 we got the possibility to use the Real Time Kinematic method (RTK method). This method has especially been used for staking out/retrieving existing system Ostenfeld control points. The experience with RTK was successful - we found the control points (undergrounded vertically placed drainpipes), and the method was 5-6 times faster than traditional methods.
The figure shows how the Real Time Kinematic method works. The one GPS receiver is placed in a control point (named the reference station) with known coordinates. By means of a radioconnection between the reference station and the rover it is possible to transmit data from the reference station to the rover. The rover is placed in the point with unknown coordinates. The received data from the reference station are used for processing the coordinates for the new point, imidiately.
To avoid problems the RTK method is only used in distances up to 5 km. The radiolink sets the limitation, not the obtainable accuracy. With this limitation in mind a network of 87 ground control points (named basic points) has been established. The distance between neighbouring ground control points is less than 10 km. This density is sufficient to ensure a total RTK coverage for cadastral surveys. By placing the basic points it has been taken into consideration that the basic points shall be suitable for GPS reference stations for RTK measurements. It means that obstructions must not disturb the receiving of data from the satellites, and the location shall ensure a stable radio communication between the reference station and the rover (no buildings, no tall dense trees).
For each of the 87 basic points a check control point has been established in a distance less than 100 m from the basic point. The function of theese check points is to ensure that the correct coordinates and ellipsoidal height have been used at the reference station at the start up, and that the radio connection operates. Before leaving the area where the reference station is placed a measurement is made at the check point. When we have finished the survey and before the reference station has been broken down the check point is measured again.
The basic points as well as the check points are coordinated in x, y, and z in the system 34 and the system UTM.
The network consisting of the 87 basic points is a strong network, in proportion to cadastral measurements, with a high relative accuracy (2 cm), which means that it won’t give any accuracy problems in areas, where there is overlap between two neighbouring basic points.
The development in the field of GPS runs fast, so it will be possible to make measurements at larger and larger distances. In a few years we can handle with less than 87 basic points. That’s the reason why we haven’t got any policy on maintenance of the 87 basic points.
SELECTING METHOD OF MEASURING - TOTAL STATION OR/AND GPS
Unlike using total stations for surveying the use of RTK means that it is not possible to measure all detailed points directly (e.g. corners of houses), because GPS requires free sight to the satellites. It depends on the number and the kind of obstructions how many detailed points it is necessary to measure by using off set functions. Normallly all the detailed points are measured by use of RTK. If there are a lot of detailed points, where the use of off set functions is unavoidable, 3 - 4 points are established by using RTK in the measure area. Theese points can be identical to some of the detailed points. From theese 3 - 4 points the rest of the detailed points are measured by use of an total station.
The time required for either using total station or RTK is almost the same, when it comes to measuring detailed points, because it takes time to set up the reference station. But the instrument can be operated by one person, and that gives more flexibility in planning the work.
The time spent on the survey is minimal compared with the time spent on driving to the survey area, so it is possible to save time, if the measurements are gathered in a way so the same basic point can be used throughout the day.
CONSIDERATIONS ABOUT ACCURACY
Generally there are some conditions which you have to become aware of to obtain a satisfying accuracy by using RTK.
Horizontal accuracy.
To test the accuracy of the RTK method a temporary instrument testing site has been established. 25 points have been placed in a distance from 0 km to 6 km from a basic point, and they have been measured 3 times. The test results include the centrering of the GPS antennas. Theese are centered with the accuracy normally used by cadastral surveys.
This test resulted in a horizontal standard error of 2.0 cm, which is smaller than the expected value. The expected value lies between 2.1 - 2.8 cm, corresponding to distances from 0 km to 6 km between the reference station and the rover. The expected standard error includes also errors from centrering the GPS antennas.
The RTK method affords high absolute accuracy, but the requirement of cadastral surveys in Denmark is high neighbouring accuracy. To ensure that the RTK-measurement affords the desired accuracy and to ensure that no blunders have been made, it is necessary to measure all the detailed points twice. It is also necessary to evaluate the deviation between the two measurements. When the RTK method is used, the results are evaluated in the field, because it is possible to see the deviations imidiately and to decide, whether a divergent point has to be measured a 3. time.
The large distance between the reference station and the rover implies, that the deviations sometimes are not as small as expected. To understand this, we have to look at the way to calculate the expected accuracy. This consists of 2 parts. The basic error and a error which depends on the distance between the reference station and the rover. Normally, when total stations are used, we work at short distances, where the influence of the distance-depending error is minimal. By RTK surveys this error has an important influence of the expected accuracy for long distances.You have to decide, if this expected accuracy is sufficient for your task. Sometimes there will be a difference between two measurements at the same point, which is larger than we are used to by using conventional means and which is not acceptable, because the task requires high neighbouring accuracy. But you can’t be sure that the deviations indicate that there is something wrong with the instrument!
The figure shows the horizontal deviations for 75 measurements, including short and large distances.
For staking out boundaries the RTK-method can afford the required accuracy, but you have to be aware of the neighbouring accuracy. In Denmark it means that existing neighbouring detailed points have to be measured (controlled), too, as documentation for the neighbouring accuracy. The existing boundaries are often coordinated in local coordinate systems or the system Ostenfeld. For staking out we need to transformate from theese local systems, first to the system 34 and then to the system WGS84. Accuracy can be lost in the transformation from the local system to the system 34!
If the task consists in retrieving boundaries, which already are marked (it means that the required accuracy is about 30 - 50 cm) , our experience is that the use of RTK is successful.
Taken together the evaluation is that the RTK method - considering the accuracy - can replace other methods of measuring in the cadastral surveys - and in tasks which require the same accuracy.
Vertical accuracy.
GPS and with that the RTK method, too, is based on a 3 dimensional coordinate system. That means that both x, y and z are required for the basic points. The required accuracy of the height at the basic points depends on the task.
To apply the RTK method in cadastral surveys it is necessary that the basic point has a high horizontal accuracy, whereas it is sufficient for the determination of the z-coordinate to use a height from a topographical map (in Denmark the accuracy of the contour lines is about 1 m in topograpfical maps in scale 1:25.000). The 87 basic points are also used for technical tasks, and theese tasks often require a high accuracy of the z-coordinate, too. Therefore the heights of the 87 basic points have been determined by spirit levelling.
The vertical accuracy, which the RTK-method gives, has also been tested.
The figure shows the deviations for 75 measurements. Again the deviations are interesting, when the requirement is high neighbouring accuracy.
The test resulted in a standard error of 1.7 cm. The test result is better than expected. The expected accuracy in the interval 0 km to 6 km lies between 2.1 cm and 3.3 cm. The largest deviation was 7.5 cm, and it must be characterized as a blunder. The next largest deviation was 4.4 cm. Theese two detailed points are neighbouring points, so the next surveyor can be able to demonstrate the difference of 12 cm!!
A deviation on that scale gives problems in the sort of tasks, where the definition of the points is exact. In this situation it is possible to test the heights by spirit levelling. The typical deviations are +/- 1.3 cm. Testing by spirit levelling can directly reveal such a discrepancy.
To find blunders and the unacceptable deviations between neighbouring points, it is necessary to measure all the points twice, if you want to use the RTK method for tasks, where the definition of the points is exact.
For other types of levelling tasks, where you do not find the same exact definition of the points, the RTK method is a good alternative to conventional means, if there are not too many obstructions.
The conclusion is that the RTK method does not give a vertical accuracy, which can be used for all sorts of levelling tasks, where the spirit levelling normally are used. Particularly where the requirement is at cm-level the RTK method can give problems. For registration of wires-/pipes and contour levelling the RTK method is superior to total stations, provided that there are no obstructions.
The considerations about accuracy are supported by the references (Lejonhufvud and Wiklund, 1996) and (Frederiksen, Jeppesen, Madsen and Poulsen 1995).
EVALUATING THE USEFULNESS OF THE 87 BASIC POINTS
To ensure that it was possible to use the 87 basic points not only for fast static and static measurements, but also for the RTK measurements, where the radiolink is necessary, the 87 basic points have been tested by coordinating 500 single points distributed over the Southern Jutland, and by measuring 10 cadastral measurements. From theese tests we found that we had to make a few alterations. To day the 87 basic points have been used for about 200 cadastral measurements made by the RTK method.
CADASTRAL SURVEYS BY THE RTK AND THE DANISH RULES
The rules in Denmark for cadastral surveys are broad (elastic), when the question is about the accuracy.
The rules for the accuracy of measurements are expressed in the following 3 points:
The measurement has to be so thorough that it is possible to
It is the surveyor’s etimate, whether the measurement and the results are sufficient.
For certain sort of tasks there are rules for establishing control points, but with the 1,200 control points from the map project, we have a total covarage, which means that it will be rather seldom, that we have to establish new control points.
The conclusion is that with observance of the Danish rules and
we don´t require further control points for cadastral surveys.
IMPLEMENTATION.
We have bought 4 sets of equipment for RTK measurements. The instruments are used daily by 6 sections. Each section have on average 3-4 surveyors.
The education is organized by an experienced surveyor, who has been working with GPS for the last four years. He works together with one surveyor at the time for one week with current tasks. During the week the planning software, the instrument, and how to download data to the CAD software are gone through. By using the RTK method it is possible to choose, whether you want the output in vectors or in coordinates. When we work with cadastral surveys we want as less processing as possible. The vector processing requires more education. Therefore we have decided to use only the coordinates.
About the transformations between the system 34 and WGS84 we have chosen to develope a program, so the cadastral surveyors don’t have to think of geodetic parametres, but only of the name of the input file.
Through this education schedule the RTK method has been quick implemented, and we haven’t seen many bad results. Specially it is important to understand that it is necessary to use the planning software, because then it is possible to save time.
The conclusion is that it is easy to use the RTK method for cadastral surveys. You don’t need to spent much time on education. But we still need qualified staff in the field to make decisions about the cadastral problems. When we use the RTK method, we evaluate the results in the field, because it is possible to see the differences imidiately. That also means that the staff has to be well educated.
CONCLUDING REMARKS
REFERENCES
Frederiksen, J. B., Jeppesen, B., Madsen, J.T.B., and Poulsen, J.L., 1995, Anvendelse af Trimble GPS til matrikulär maaling, Aalborg University, Denmark.
Hansen, B. S., 1995, Article in Landinspektören 1-95, Digitale matrikelkort i Sönderjylland, og hvordan de anvendes i det daglige.
Lejonhufvud, C. and Wiklund, P., 1996, Undersökning av seminkinematisk GPS-mätning i realtid, Lantmäteriverket, Sweden, Gävle.
Ryttersgaard, J., 1995, Article in Landinspektören 1-95, Ändringer af referencenettet og brydningstider i opmälingsmetoderne.
Biographical notes.
Susanne Hansen, Denmark
1979: Degree of Cand. Geom. (Graduate in Geodesy, Cadastral
Science and Planning), Aalborg University.
1996-98: Studies at Aalborg University, Master of Technology
Management in Cadastral Administration.
1979- 84: National Survey and Cadastre, Denmark, South Jutland
Branch.
Regional office, cadastral and technical works.
1984- 92: National Survey and Cadastre, Denmark, South Jutland
Branch.
Project manager, cadastral map project.
1992- 98: National Survey and Cadastre, Denmark, South Jutland
Branch.
Technical Department, Head of Section: digital cadastral maps,
control points, GPS, survey-documents projects.
1th March 98: National Survey and Cadastre, Denmark, South
Jutland Branch.
Cadastral Department, Head of Section, Haderslev.