School of Surveying and Spatial Information Systems
The University of New South Wales
RTK-GPS for Cadastral Boundary
Surveying in NSW
by Adam Veersema
Supervised by Dr. Craig Roberts
October 2004
Abstract
Real Time Kinematic GPS has been used for a variety of different surveying applications and the use of this GPS method for cadastral work is becoming commonplace. A case study survey was conducted to investigate the use of the RTK technique for boundary surveys. Both a Total Station and an RTK survey were conducted over six lots in Inverell, and the results showed up to 57mm difference between the coordinates derived from the two survey methods. It was concluded that the RTK technique competes well with traditional surveys methods in terms of accuracy, cost and efficiency, and also complies easily with Class C survey standards as required by current legislation.
Contents
Survey Results and
Comparisons
Comparisons of Reference Mark Joins
Introduction
The use of Global Positioning Systems (GPS) is becoming commonplace in most large surveying practices and more surveyors are taking on this new technology for a variety of different surveying applications and solutions. GPS is being used commonly for large detail and contour surveys as well as a variety of engineering applications. Though many surveyors feel that the high capital outlay for the technology, and lowered productivity for small cadastral surveys when compared to traditional Total Station surveying techniques do not justify investing in GPS.
With continual research and development into GPS, the techniques and systems developed have become more reliable, cheaper and more productive, making GPS more attractive for a range of surveying solutions. The question then arises that if RTK-GPS is now accurate and viable enough to compete with Total Station methods, can this new technology be used for boundary surveys, most notably boundary definitions and redefinitions?
Case Study Fieldwork
In order to determine if the RTK technique is suitable for this type of
cadastral work, a case study survey was conducted. The work was completed in
Inverell in northern New South Wales and the site chosen for the survey
consisted of six lots comprising about 10 hectares located on the town’s
outskirts. Two surveys were completed, one using a Nikon DTM821 Total Station
and the other using a Leica SR530 RTK Field Unit and Base station with a radio
link.
The aim was to compare the resulting coordinates obtained from the RTK
survey to those derived from the Total Station survey. The two surveys would be
compared in terms of accuracy, efficiency, cost as well as practicality to the
surveyor.
Figure 1: Nikon DTM 821 Total
Station
Figure 2: Leica SR530 RTK Unit
operating as a Base Station
Survey Results and Comparisons
Both surveys were competed from the 10th to the
14th of July 2004 and the measurements were reduced, tolerances and
miscloses checked using CivilCAD for the Total Station survey and Leica
Geomatics Office (LGO) for the RTK data and any bad observations were deleted
from the RTK data. In order to compare the two surveys SSM 56420 was held fixed
and MGA azimuth was obtained by adopting the join to SSM 56423. The Total
Station survey was then reduced onto the MGA grid to allow the comparison of
the two methods, as the RTK measurements are all done on the MGA grid.
The results between the two techniques can be found in Table
1 below. Point 106 appears twice as this is SSM 56422 and an extra RTK
occupation was made to improve the survey redundancy of this mark.
|
|
Total Station Coordinates |
Mean RTK Coordinates |
|
||
|
Point No. |
Easting |
Northing |
Easting |
Northing |
Misclose Vector (m) |
|
72 |
316374.942 |
6706371.228 |
316374.963 |
6706371.216 |
0.024 |
|
73 |
316349.891 |
6706452.916 |
316349.897 |
6706452.904 |
0.013 |
|
74 |
316372.080 |
6706550.028 |
316372.090 |
6706550.007 |
0.023 |
|
76 |
316684.692 |
6706502.030 |
316684.702 |
6706502.043 |
0.017 |
|
80 |
316476.586 |
6706352.126 |
316476.617 |
6706352.105 |
0.037 |
|
82 |
316365.684 |
6706219.635 |
316365.735 |
6706219.610 |
0.057 |
|
83 |
316385.448 |
6706365.814 |
316385.464 |
6706365.787 |
0.032 |
|
88 |
316367.572 |
6706543.203 |
316367.570 |
6706543.177 |
0.026 |
|
94 |
316667.439 |
6706374.767 |
316667.460 |
6706374.768 |
0.022 |
|
95 |
316659.626 |
6706317.286 |
316659.660 |
6706317.304 |
0.038 |
|
97 |
316529.314 |
6706334.906 |
316529.348 |
6706334.886 |
0.039 |
|
98 |
316529.418 |
6706335.912 |
316529.460 |
6706335.889 |
0.048 |
|
99 |
316508.840 |
6706396.259 |
316508.857 |
6706396.207 |
0.054 |
|
100 |
316502.057 |
6706338.627 |
316502.099 |
6706338.611 |
0.044 |
|
106 |
316092.590 |
6706602.636 |
316092.600 |
6706602.621 |
0.018 |
|
106 |
316092.590 |
6706602.636 |
316092.599 |
6706602.623 |
0.016 |
Table
1: Differences between Total Station and RTK derived
coordinates.
The misclose distances between the RTK and the Total Station surveys range from 13mm to 57mm, with an average misclose of 32mm and a standard deviation of 14mm. When comparing the Easting and Northing coordinates for each point obtained by each method some interesting trends were discovered. The easting differences were almost all negative values, suggesting that the RTK derived coordinates were further East than the Total Station derived coordinates by an average of 22mm. The differences in the northing coordinates also showed a trend in that the RTK derived coordinates were almost all South of the Total Station derived coordinates as a large majority of the difference values were positive by an average of 16mm.
Figure 3: Misclose vectors between the
RTK and Total Station derived coordinates.
Point numbers are concurrent with Table 1. Figure not to scale.
Comparisons of Reference Mark Joins
To determine the boundary from the survey work completed it would be necessary to compare the measured joins between reference marks (RMs) such as GIPs, CBs, DH&Ws as well as boundary pegs and any other monuments found on DPs that have been surveyed.
To compare the measured joins with the calculated joins, via both distance and bearing it is necessary to have all the measurements on the same azimuth. As the subject and the adjoining DPs were all in the ISG system, and the RTK survey work was completed in MGA, a swing of -7˚33’31” had to be applied to have all the measurements in the ISG system, as on the DPs. This was calculated by adopting the join between the CBs on lots 16 and 20 along Vintage Close in DP716360.
When comparing the calculated joins between the RMs from the relevant DPs showed some varied results. The total station survey came in quite good with a lot of the joins, as should be expected, with differences ranging from 9mm up to 102mm which is quite a large and unexpected distance to find between two pegs.
The joins from points 72 to 73 and 74 to 76 are not on the DPs as they are traverse stations used during the surveys and were included in the table to show the good agreement between the RTK and Total Station surveys.
|
Point Pair |
RM Type |
RTK vs. TS |
RTK vs. DP |
TS vs. DP |
|
72 - 73 |
GIN |
0.005 |
NA |
NA |
|
94 - 95 |
PEG |
-0.018 |
-0.009 |
0.009 |
|
95 - 97 |
PEG |
-0.005 |
-0.107 |
-0.102 |
|
97 - 100 |
PEG |
-0.008 |
0.002 |
0.010 |
|
98 - 99 |
GIP |
-0.019 |
0.012 |
0.031 |
|
82 - 83 |
CB |
-0.007 |
-0.033 |
-0.026 |
|
83 - 88 |
CB |
0.003 |
-0.027 |
-0.030 |
|
74 - 76 |
DPY |
-0.005 |
NA |
NA |
Table 2: Differences between the RTK and
Total Station and DP calculated
Join distances for a number
of different reference marks. All
distances are
in metres.
The comparison of the measured angles between the different techniques and those calculated from the Deposited Plans is also an important consideration when determining the boundary. The table below shows the differences between the measured bearings from the two methods and also those calculated between the RMs on the DPs.
|
|
|
RTK vs. TS |
RTK vs. DP |
TS vs. DP |
||||||
|
Point Pair |
Code |
D |
M |
S |
D |
M |
S |
D |
M |
S |
|
72 - 73 |
GIN |
|
|
-34 |
|
|
|
|
|
|
|
94 - 95 |
PEG |
|
|
-33 |
|
|
28 |
|
1 |
1 |
|
95 - 97 |
PEG |
|
|
-59 |
|
-2 |
17 |
|
-1 |
18 |
|
97 - 100 |
PEG |
|
|
38 |
|
3 |
43 |
|
3 |
5 |
|
98 - 99 |
GIP |
|
-1 |
47 |
|
|
-58 |
|
|
49 |
|
82 - 83 |
CB |
|
|
-48 |
|
|
-48 |
|
|
0 |
|
83 - 88 |
CB |
|
|
-21 |
|
-1 |
11 |
|
|
-50 |
|
74 - 76 |
DPY |
|
|
-22 |
|
|
|
|
|
|
Table 3: Differences between the RTK,
Total Station and DP calculated
Join bearings for a number of different reference marks.
Class of Case Study Surveys
The Surveyor General’s Directions for GPS Surveys, which is available from the Dept. Lands Website (direct link unavailable at the time of web page construction), requires that all surveys in which GPS is being used must conform to Class C standards, which can be achieved by following Standards & Practices for Control Surveys, which is the ICSM publication often called SP1. To assign class to the RTK survey, the guidelines on page B-28 of the SP1 document were followed. Class was determined for the lines or joins between similar points, such as GIP’s, CB’s and Pegs. It was chosen to test these lines, as the joins between such marks would be checked for agreement with the applicable DP’s when determining the boundary. The results of this class test show some interesting trends.
Out of the eight joins tested, 3 failed the class B test. The calculated standard deviation of the misclose between all eight pairs of points was 10mm. This value was obtained again by using the manufacturers specifications of 10mm + 1ppm for the SR530 unit in RTK mode as required by SP1. The results for the class tests can be seen in the tables below.
|
Point Pair |
Distance (m) |
Line SD |
95% CI |
Satisfy |
|
72 - 73 |
85.447 |
0.01046 |
0.01049 |
Y |
|
94 - 95 |
57.991 |
0.01048 |
0.00948 |
N |
|
95 - 97 |
131.493 |
0.01047 |
0.01218 |
Y |
|
97 - 100 |
27.503 |
0.01045 |
0.00836 |
N |
|
98 - 99 |
63.740 |
0.01045 |
0.00969 |
N |
|
82 - 83 |
147.502 |
0.01045 |
0.01277 |
Y |
|
83 - 88 |
178.291 |
0.01048 |
0.01390 |
Y |
|
74 - 76 |
316.270 |
0.01048 |
0.01897 |
Y |
Table 4: Results of the Class B test on pairs of RM’s
|
Point Pair |
Distance (m) |
Line SD |
95% CI |
Satisfy |
|
72 - 73 |
85.447 |
0.0105 |
0.0210 |
Y |
|
94 - 95 |
57.991 |
0.0105 |
0.0190 |
Y |
|
95 - 97 |
131.493 |
0.0105 |
0.0244 |
Y |
|
97 - 100 |
27.503 |
0.0104 |
0.0167 |
Y |
|
98 - 99 |
63.740 |
0.0104 |
0.0194 |
Y |
|
82 - 83 |
147.502 |
0.0104 |
0.0255 |
Y |
|
83 - 88 |
178.291 |
0.0105 |
0.0278 |
Y |
|
74 - 76 |
316.270 |
0.0105 |
0.0379 |
Y |
Table 5: Results of the Class C test on pairs of RM’s
It is interesting to note that all the joins that failed the class B test had join distances of less than 65 metres, which when computing the suggested tolerance for these lines, they all result in tolerance values of less than 10mm at the 95% level. Using the manufacturers specifications in the equation, with the aim to obtain a class B tolerance of 10mm at the 95% level, then the distance between the two points should be no less than 72.109 metres. All the joins between these points easily complied with the tolerances at the 95% level for a Class C survey as can be seen by Table 9 on the previous page.
These results show that RTK can conform easily to Class C standards and could comply with Class B standards if required as long as special detail was paid to a strong network design and the survey had good redundancy.
Cost Comparisons of Surveys
An interesting way of comparing the two surveys, which is of great interest to the practicing surveyor, is which method is more viaible in terms of cost efficiency. The total Station survey took 9.5 hours to complete the surround survey and connect to SSM’s, where as the RTK survey took 4 hours for both occupation sessions. So would it be cheaper for a survey firm, with all necessary Total Station equipment at their disposal, to hire RTK and complete the survey task? The tables below show the cost comparison.
|
Total Station Survey |
|
|
|
|
|
|
|
Search |
|
|
|
|
|
|
|
4 DP's |
4 |
|
@ |
$14 |
ea |
$56 |
|
4 SSM's |
4 |
|
@ |
$3.50 |
ea |
$14 |
|
2 Man Party |
9.5 |
hrs |
@ |
$170 |
hr |
$1,615 |
|
Data Input and Processing |
2 |
hrs |
@ |
$90 |
hr |
$180 |
|
|
|
|
|
|
TOTAL |
$1,865 |
Table 6: Table of expected costs for case study Total Station survey.
|
RTK Survey |
|
|
|
|
|
|
|
Search |
|
|
|
|
|
|
|
4 DP's |
4 |
|
@ |
$14 |
ea |
$56 |
|
4 SSM's |
4 |
|
@ |
$3.50 |
ea |
$14 |
|
Equipment Hire |
|
|
|
|
|
|
|
2 Field Units |
1 |
set |
@ |
$715 |
day |
$715 |
|
2 Man Party |
4 |
hrs |
@ |
$170 |
hr |
$680 |
|
Data Input and Processing |
1 |
hrs |
@ |
$90 |
hr |
$90 |
|
|
|
|
|
|
TOTAL |
$1,555 |
Table 7: Table of expected costs for
case study RTK survey using most
expensive hire price.
The RTK survey is quite a bit cheaper, by $300, even when hiring quite an expensive long range RTK unit. When considering hiring cheaper units, the price difference between the two methods is almost $700. This saving occurs when hiring a short range RTK unit (up to 2.5 kms), which for the case study survey would be more than adequate.
Conclusion
It can be seen from the various comparisons between the two methods presented in this report that the RTK technique of survey agrees quite well with the traditional total station methods of survey in terms of accuracy, efficiency and cost.
The RTK method is definitely more efficient as the jobs become larger and very cost effective when compared to traditional Total Station methods of survey. There are some slight angle differences, which result in up to 50mm difference over some lines, but this can be reduced by improved survey technique, methods of point observation and increasing redundancy on points that might cause concern or points that are vital to the survey fix. Distance measurement and join wise the surveys are quite similar and agree with DP distances.
These results all combine to show that the RTK survey complies easily with Class C survey standards and only fails Class B standards on shorter lines, which is more than adequate enough for most Cadastral survey applications. Thus for larger jobs such as a rural boundary job, as carried out in this case study, the technique of Real Time Kinematic GPS will be able to deliver results and accuracies equivalent to those expected from a traditional Total Station survey, with the benefit of reduced cost and improved productivity for the surveyor.
Further Information
For more information, please contact:
Dr. C. Roberts
Email: c.roberts@unsw.edu.au
Mail:
School of Surveying and Spatial
Information Systems
University of New South Wales
UNSW SYDNEY NSW 2052
Australia
Phone: +61-2-9385-4203
Fax: +61-2-9313-7493
WWW: http://www.gmat.unsw.edu.au/
Adam Veersema
Email: adamveersema@yahoo.com.au