THE USE OF AERODIST FOR FILLING IN BETWEEN TELLUROMETER TRAVERSE LOOPS

by

B.P. Lambert,

Director of National Mapping,
Canberra, Australia

(Lambert, B.P. (1967), The Use of Aerodist for filling in between Tellurometer Traverse Loops, in Proceedings of Conference of Commonwealth Survey Officers, Cambridge, Paper B3, pp.93-106)

 

Introduction

Once it has been established that Aerodist has a sufficient range of operation, is rugged enough for long spells of field operation and will produce horizontal control of the necessary accuracy, its use or otherwise, for filling in the primary network of horizontal control becomes a matter of management decision based on consideration of other relevant factors.

These other factors will, of course, vary from country to country and will be greatly affected by the immediate objectives of survey and mapping programs.

For these reasons the scope of this paper must necessarily be very largely restricted to the tasks with which the author is familiar. It therefore deals with surveys in Australia, makes some mention of operations in New Guinea and reaches certain conclusions that are based on the results obtained from these surveys.

Range of Operation

Using the line crossing technique, occasional lines of 200-250 kms length have been measured with standard equipment but the optimum distances have proved to be 100-150 km while quite satisfactory measurements can be made over lines of 40-50 km length. (1).

Recent tests have been carried out with a master station mounted in a helicopter and so connected that the signals could be switched via a standard flat antenna or via an antenna taken from a remote set of equipment. The results indicated that a ground to air distance of 160 km could be comfortably measured through the remote antenna under conditions in which the normal antenna gave no result at all.

The possibility of regularly using the curved antenna in a fixed-wing aircraft is under investigation as the added range capacity, if developed in time, could greatly affect the operational economics.

Ruggedness and Reliability of Equipment

Two sets of equipment, after preliminary trials and testing, have been operated in Australia and New Guinea since 1963 and they generally gave trouble in the early years.

However, electronic technicians now accompany the field parties and it is the accepted regular procedure for the equipment, on return from the field, to be thoroughly overhauled and then field tested prior to commencement of the next field season.

This practice, together with the greater experience of the survey staffs, has ensured practically trouble-free operation over field periods of about 6-8 months and, looked after in this fashion, the equipment now appears to be adequately rugged and reliable for a number of years of additional field work.

Accuracy attainable by Aerodist Survey

The determination of station elevations has a considerable bearing on the attainable accuracy. The general practice in Australia is to carry third order levelling into each Aerodist ground station and use near vertical Aerodist measurements to calibrate aircraft altimeters.

Early assessments of Aerodist measurements (2) led to the assumption that a standard error of ± 3.5 metres was practicable from a single line crossing for lines of 100-150 km length and Canadian results (3) seem better still. However, continuous local operation over a number of years indicate that about ± 4 metres would be a more appropriate figure. (See Annexure A).

Investigations carried out (Annexure B) indicate that quite large blocks of Aerodist trilateration adjusted to perimeter control (assumed free of error) will result in average co-ordinate position errors of about one-half the standard error of a line measurement and indicate that the average correction to individual measurements is also about half of the standard error of the line measurements.

In practice, these errors and corrections will be directly affected by the residual inaccuracies of the primary network and the possible non elimination of lines with large errors. The results quoted in Annexure A give some indications of the precision obtained in practice when 6-9 good line crossings are made. From these results, it would seem that a co-ordinate precision can be expected of approximately ± 1 to 2 metres relative to surround control.

On this basis it is interesting to speculate that, if minimum quadrilaterals of 2° x 2° are assumed and extrapolation made on the analogy of photogrammetric investigations into block adjustments (4), it would seem that adequate first order horizontal control could have easily been provided over the whole continent by containing Aerodist quadrilaterals of this size within a first order perimeter traverse and by making a greater number of crossing measurements per Aerodist measured line.

However, for mapping control, a close spacing of control would be required and in order to save costs by reducing the number of occupied ground stations, attempts have been made to supplement the ground station control by photo-trilateration techniques.

In this technique, a Wild HC1 horizon camera was geared to a manually levelled Vinten vertical camera, with which consecutive photographs were taken while flying over a particular point of ground detail. Aerodist distances to ground stations were recorded at the instants of exposure.

On the resultant photographs, it has been very difficult to locate the image of the horizon, furthermore, the metric qualities of the positioning camera are poor, so that to date, the results have not been as good as expected.

Consideration is being given to the installation of a normal angle air survey camera and subsequently orienting the photographs in a stereo plotter to fit approximate control derived from the RC9 photographs controlled approximately for scale, by the available 1:250,000 compilations, and for elevation, by the raw airborne profile data.

Accuracy Required

Basically this has been related to the accuracy of the horizontal control necessary for :

-      the present Commonwealth program of topographic mapping (Annexure C) which has the objective of compiling the whole of Australia and New Guinea by the end of 1975 at a scale of 1:100,000 and with a normal 20 metre contour interval.

-      the desirability of permitting a further breakdown of control for larger scale mapping and for other survey purposes.

Horizontal Control for 1:100,000 Scale Mapping

In Australia the climate is usually very favourable for air photography and the existence of 1:250,000 scale planimetric maps prepared from air photography permits systematic and careful planning of flight strips and control patterns. The photography for the 1:100,000 program is taken with RC9 cameras at a scale of 1:80 - 84,000 and the regular provision of an 80% overlap facilitates photogrammetric block adjustment.

The system used for this adjustment directly affects the intensity of horizontal control required.

The Royal Australian Army Survey Corps, for the Australian content of its allocated portion of the 1:100,000 program, is proceeding with a 1° x 1° pattern of control using the Wild A9 (or a stereocomparator) for bridging and adjusting in blocks of 1½° of longitude by 1° of latitude (5).

In New Guinea the photography is of such a heterogeneous nature that it causes very difficult photogrammetric bridging problems. It is quite obvious that numerous horizontal control points of the nature readily provided by Aerodist trilateration and Aerodist controlled photography will be of the utmost value. The use of Aerodist in this country by the Survey Corps is illustrated in Annexure D.

Investigations (Annexure E) were conducted by the Division of National Mapping to ascertain what horizontal control spacing was necessary for the 1:100,000 program, assuming 1:80,000 scale RC9 air photography with airborne profile records along the sidelaps. The results obtained show that control at the corner of 1° x 1° blocks will suffice even for 1:50,000 mapping and that something less than this would be satisfactory at 1:100,000 scale.

The Division has traditionally used slotted templates and the results obtained from the tests with precision nadir templates have been most encouraging.

If these templates can be used for horizontal bridging, airborne profiles used for height control and plotting based on automated stereo-plotting, there will not be the large requirement for skilled stereoplotter operators that might otherwise prejudice timely completion of the 1:100,000 program.

Furthermore, templates can be laid over very large areas that are difficult to handle by measurement and computation, other than by internal sub-grouping with consequent accuracy loss.

In order to determine the accuracy limitations of the template method, the Division has established Aerodist ground stations in a 1° x 1° pattern over a particular operational-area and blocks of varying size will be covered by templates fitted to perimeter control and tested for accuracy by using the internal control points for checks. (It is normal practice to increase accuracy by laying an overspill of templates around the perimeter.)

It is confidently expected that control established either at ½° intervals around the perimeter of a 1½° x 1½° quadrilateral, or at the corners of a 1° x 1° quadrilateral and supplemented by a centre point, fixed by phototrilateration, will suffice for 1:100,000 mapping.

While awaiting the results of these tests, the Division will provide a pattern of ½° x ½° control at Aerodist ground stations in those areas where regular 1:50,000 scale mapping is likely within the foreseeable future and in other areas Aerodist ground stations will be established at the corners and centres of 1° x 1° quadrilaterals and supplemented by phototrilateration control at the other ½° points.

Concurrently, the Division will endeavour to extend the practical operating range of the Aerodist equipment.

The success of these endeavours could materially affect the economics of future operations.

Horizontal Control for Mapping of Larger Scales than 1:100,000 and for Other Survey Purposes

In other countries of the size of Australia the largest basic mapping scale appears to be about 1:25,000 with the occasional use of a 1:10,000 scale in more developed areas.

The results obtained with Aerodist show that a positional standard error of ± 1.5 metres can be obtained for a 30’ x 30’, quadrilateral network and if it is assumed that extension from this by one of the helicopter hovering techniques will give an additional ± 1.5 metres, then a combined positional standard error will result that would readily permit 1:25,000 scale mapping.

If a central control point is similarly fixed from a number of surrounding Aerodist points and additional control obtained by hovering helicopter at the corner of each model, then still larger scale mapping of isolated areas could be undertaken (when allowance is made for the greater error tolerance at these scales).

In so far as other types of survey are concerned, the horizontal control from Aerodist will provide starting co-ordinates with a standard error of the order of ± 1.5 metres relative to the primary network and in the less developed areas of a country such as Australia this capacity could be invaluable, but in the more developed areas, ground control survey methods will be necessary in order to meet the normally more stringent accuracy requirements.

When planning the Aerodist ground station pattern, it is of considerable importance to allow for future intensification of horizontal control by ensuring that the whole terrain can be covered from these stations by (preferably three) radial measurements with E.D.M. equipment.

The earlier mentioned tests with Aerodist equipment would indicate an upper limit of 180 km for these radial distances.

Other Factors to be considered

It is not intended in this paper to deal with the factors common to most survey and mapping operations and which have to be considered in the normal managerial appreciation of the situation. However, one important factor that will be considered is possibility that other techniques and equipment can be applied to the task more effectively than Aerodist.

Of the available equipment the only one offering serious competition under normal Australian conditions is the Tellurometer and experience has shown that Tellurometer traversing if controlled by frequent azimuth observations can readily provide control of second order accuracy.

This technique has the advantage that trigonometric height control can be provided simultaneously and data can be very quickly computed, whereas, Aerodist distance measurements only become available after a considerable amount of chart reading and computation.

The choice between equipment will most likely depend on the size and nature of the area to be surveyed and the relative cost and time comparisons.

The main terrain factors that adversely affect Tellurometer traversing and that conversely favour airborne surveys are the necessity to :

 

(a)

erect towers

 

(b)

clear vegetation

(c)

travel over ground that is unfavourable to ground transport

 

In Annexure F an approximate analysis has been made of cost and time factors for various types of terrain on the assumption that these cover quite extensive areas.

It will be noted that in assessing the cost of Aerodist survey a complementary airborne method of extending height control has also been taken into account in order to maintain balance with height control provided during traversing.

From these analyses there is no doubt that for a limited area of ideal surveying country and a requirement for a fairly close network of control, Tellurometer traversing is the best answer.

On the other hand, in extensive terrain that is heavily timbered, or extremely rugged, airborne survey methods would be a much more practical proposition.

Where limited areas of airborne control are a necessity, it should be possible to develop a trilateration technique using a hovering helicopter, a vertical survey camera located at the unknown point together with E.D.M. equipment capable of measuring to an airborne remote station vertically from a master station located at the ground level of the unmown point and radially from a known station to another helicopter borne remote station.

It would be necessary to expose the camera and simultaneously record E.D.M. measurements on voice command from the helicopter.

With this setup, the photogrammetric measurements on the photograph and the vertical height would permit of appropriate corrections being made to the radial measurement.

Such a technique would have the benefit that the unknown stations could be marked for future use.

If helicopter costs and range of operation ever become competitive with those of fixed-wing aircraft this technique could be applied to extensive trilateration schemes.

There is, of course, the possibility of using Tellurometer traversing along accessible routes and extending control sideways with airborne equipment.

Conclusions

If the region to be mapped contains extensive areas that can be more appropriately controlled by airborne survey techniques, then there is no doubt that Aerodist is rugged enough, accurate enough, and economical enough, to provide the necessary breakdown within the primary traverse network.

However, in easy survey country and probably in reasonably populated areas, Tellurometer traversing will be the best technique while in limited areas of difficult access it may be worthwhile to develop a helicopter or hovering technique using locally available equipment.

There will, of course, be fringe areas where the benefits either way are marginal and in these areas no doubt everything available will be used in order to complete control surveys as early as practicable.

This paper, having established that Aerodist can be effectively used under certain conditions for the breakdown of a primary horizontal control net, now finishes on the note on Which it commenced - the final decisions are of a managerial nature. It is hoped that the material presented herein will help others who may be faced with these decisions.

As a matter of interest the current allocation of areas between Aerodist and Tellurometer in Australia is illustrated in Annexure G.

(Editor’s note : The original Annex was not suitable for publication and has been redrawn showing National Mapping Aerodist operations, completed and proposed, to 31 December 1966 and the area proposed for future Tellurometer and Aerodist operations enclosed by the red line).

 

LIST OF ANNEXURES

A

Report of precision obtained in various Aerodist survey operations.

B

Investigations into error patterns in the block adjustments or trilateration models.

C

Diagram illustrating Australian 10 year 1:100,000 scale topographic mapping program.

D

Diagram illustrating status of the Royal Australian Survey Corps Aerodist surveys in Papua New Guinea.

E

Investigations into horizontal control requirements for 1:100,000 scale mapping.

F

Comparisons of relative cost and time factors between Aerodist and Tellurometer horizontal control surveys.

G

Provisional proposals for areas of Aerodist and Tellurometer survey in Australia.

LIST OF REFERENCES

(1)

Lines J.D., "Aerodist in Australia", Australian Surveyors, Vol. 21, No. 2, 1966,

(1)

Turner L.G., "Aerodist operations in Australia" Document E/CONF. 52/L.44. (*)

(2)

Lambert B.P., and other officers of Division of National Mapping, "Aerodist Surveys-Report on operations carried out by the Australian Division of Mapping since 1983”, presented at the I.A.G. Symposium on Electromagnetic Distance Measurement, Oxford, England, 1985.

(3)

Numerous reports on Aerodist operations published by the Department of Mines and Technical Surveys, Ottawa, Canada.

(3)

Tuttle, A.C., "Aerodist in geodetic surveying in Canada". Document E/CONF. 52/L.95. (*)

(4)

Ackermann F., "On the theoretical accuracy of planimetric block adjustment". Paper presented at the International Symposium on spatial aero triangulation, Illinois, U.S.A. February-March 1966.

(5)

Buckland F.D., "An approach to 1:100,000 and 1:50,000 scale mapping using super wide angle photography". Document E/CONF. 52/L.56. (*)

(*) Presented at 5th United Nations Regional Cartographic Conference for Asia and the Far East, Canberra, Australia, March 1967.

 

ANNEXURE E

INVESTIGATIONS INTO HORIZONTAL CONTROL REQUIREMENTS FOR 1:100,000 SCALE MAPPING

Introduction

In view of the following facts :

(a)

two-thirds of Australia has a slope of less than 5%;

(b)

the greater proportion of the country does not exceed 700 metres in elevation and the highest point has an elevation of 2,200 metres;

(c)

the close verticality of present day photography;

(d)

the possibility of obtaining A.P.R. profiles along the sidelaps of the air photography;

the Division of National Mapping decided to thoroughly investigate the practicability of radial triangulation techniques. Both slotted template and mathematical (Radial Triangulator) techniques were examined.

Test Details

For convenience and economy the test area was located near Canberra, and the range of ground elevation in the test area was approximately 300 metres.

An RC9 camera was used for the photography, 80% longitudinal overlap was provided and the average and maximum tilts were respectively ½° and 1½°. Four flight strips of 8 overlaps were tested, the overlaps being selected so as to line up perpendicular to the flight directions.

Control points were established by ground survey at the corners of each overlap and in such positions that each control point was common to four adjoining overlaps. The marked up control then provided the pass points for the adjustment.

These surveys were undertaken subsequent to the flying of the original air photography. The control was signalised on the ground, photographed by hand held camere from a low flying aircraft and marked on the basic photography with the aid of a differential stereoscope.

In some of the template tests, this marking was done with a needle point and in other cases with the Wild PUG 3 but no significant difference was noted between the two processes.

Tests were carried out with thin film diapositives using slotted templates and the Wild RT-1 Radial Triangulator (with PUG marking only).

Control Patterns

The control patterns generally used for the tests were :

(a)

4 point - at each corner of test area.

(b)

5 point - (a) plus a centre point.

(c)

8 point - (a) plus points mid way along the perimeter sides.

(d)

9 point - (c) plus a mid point.

(e)

perimeter - (a) plus 12 to 14 well spaced points around the perimeters.

In addition, special tests were made with the Radial Triangulator observations fitted to complete perimeter control around a block of 3 flight strips each containing 6 overlaps.

Slotted Template Tests

The template assembly scale was 1:40,000 and sets of templates were prepared with slots radiating from both the photo centre and the near nadir.

As a test of the overall accuracy of both the template assemblies and the control, the centre point and near nadir templates were each laid to fit all 45 available control points. A perfectly flat assembly was easily obtained.

Results obtained with radial slotted tem plates (at photo scale of 1:35,000)

 

Control Pattern

 

Average Vector m.s.e. in mms

Range of Vector

m.s.e

No. of Tests

 

(a)

4 point

± 0.20

0.13

10

(b)

5 point

± 0.16

0.03

3

(c) & (d)

8 or 9 point

± 0.14

0.02

4

(e)

perimeter

± 0.10

0.01

-

Radial Triangulator Tests

These were carried out using a near nadir centre, reading angles, once stereo‑

scopically and once monocularly, and using the variation of co-ordinates (Varycord)

technique for adjustment to control. Special additional tests were made using 3 rows of

6 overlaps fitted to complete perimeter control in an endeavour to establish the limits

of accuracy obtainable by this technique.

 

Results obtained with radial triangulator

 

Control Pattern

 

Average Vector m.s.e. in mms

Range of Vector

m.s.e

No. of Tests

 

(a)

4 point

± 0.12

0.0

2

(b)

5 point

± 0.11

0.03

2

(c) & (d)

9 point

± 0.08

0.01

2

(e)

perimeter

± 0.07

0.02

2

(f)

special test

± 0.06

0.0

2

Assessments of Results

The above results indicate a limiting co-ordinate accuracy of ± 0.4mm in co-ordinate error which would correspond to ± 3 metres from 1:84,000 scale photography. However, where plenty of perimeter control is available, some allowance should perhaps be made for inaccuracies in transfer of ground control points to the photographs but these inaccuracies would have a negligible effect on the 4 point control assemblies.

Extrapolation

The first extrapolation was made by assuming that the results obtained from 1:35,000 photography are directly proportional to those that would be obtained from 1:84,000 scale photography.

The second extrapolation is based on material in Ackerman's paper (4) On the theoretical accuracy of planimetric block adjustment where it will be noted that, in respect of a block of 4 x 8 photographs, the ratio of 2 to 1 here obtained in practice between 4 corner control points and dense perimeter control agrees with Ackerman's theoretical ratio.

Experience in the Division has always supported the contention that use of additional photogrammetric data around the perimeter of a block will improve the work when control is limited. In fact it has been the practice for many years now to lay templates to the next line of control outside the perimeter of any block.

Ackerman's work indicates a 2 to 3 improvement ratio from the inclusion in the adjustment of models beyond the perimeter of the block.

This ratio has therefore been applied to the results previously quoted and thereby obtain the following-figures :

 

Control Pattern

 

Slotted Template m.s.e. (mm)

Radial Triangulation m.s.e. (mm)

 

 

4 point

± 0.13

± 0.08

 

8 or 9 point

± 0.09

± 0.05

The third extrapolation has been from the ½° x  ½° quadrilateral to a 1° x 1°
quadrilateral using a ratio of 5 to 3, deduced from Ackerman's investigations, and
the following figures have been obtained:‑

 

Control Pattern

 

Slotted Template m.s.e. (mm)

Radial Triangulation m.s.e. (mm)

 

 

4 point

± 0.22

± 0.13

 

8 or 9 point

± 0.15

± 0.09

Validity of Extrapolations

In order to assess the validity of the extrapolations, comparison was made with reports of similar investigations that had been conducted elsewhere on block areas (of similar model content to 1° x 1°).

Template Methods

No comparisons are known in respect of nadir slotted templates but the following figures for comparable blocks are available in respect of stereo-templates :

(a)

4 point control

 

The I.G.N. report to I.S.P. Study Group into Block Adjustment for 1:100,000 Scale Mapping (Lisbon 1984) gave a vector m.s.e. of ± 0.24mm at photo scale.

(b)

9 point control

The A.S.P. Manual of Photogrammetry (Third Edition, Vol.1, pp. 445) quotes a vector m.s.e. of ± 0.13mm.

Analytical Methods

The I.G.N., at Lisbon, reported the results of a number of tests of a block that was very similar to the Australian proposed 1° x 1° block and using 9 point control. The average vector m.s.e. was ± 0.10mm at photo scale.

The I.T.C. also reported to the same Study Group that with 10 point perimeter control they obtained a vector m.s.e. of ± 0.09mm at photo scale for the same block.

In the light of these results it is considered that the figures deduced from the Division's tests are reasonably valid.

Conclusion

The allowable m.s.e. for the plotted position of well defined detail is ± 0.31mm, at the 1:100,000 map scale which corresponds to 0.35mm at the photo scale of 1:84,000, therefore any of the considered extrapolations of the near nadir template technique would suffice but having regard to the normal increase of error between controlled investigation and practical production it would not be wise to normally rely on 4 control points at the corner of the 1° x 1° quadrilateral.

The 8 point perimeter control at corners and ½° spacing combined with an overspill of templates to the next line of control appears admirably suited for an area of this size.

The overriding observation in respect of the Division's tests, and those submitted at Lisbon by many organisations, was the incidence of human error which, even if only present in a small percentage of cases, will adversely affect the results.

Operational procedures designed to eliminate this source of error are essential.

 

ANNEXURE F

EXPLANATORY NOTES

(Editors note : The Annexure has been truncated on the left but the top four items relate to using Tellurometer traversing and the bottom four items to the use of Aerodist)

Basic Assumptions

Types of Terrain

The following types of terrain have been considered :

(a)

Type 1

Easy survey country with plenty of roads and presenting simple logistic problems.

(b)

Type 2

Remote desert type areas involving cross country ground transportation and presenting difficult logistic problems.

(c)

Type 3

Flat country covered with timber of medium density with road transportation along a limited number of roads.

(d)

Type 4

Equatorial rugged jungle covered country.

Air Photo Coverage

It is assumed that the Terrain Types 1, 2 and 3 would be covered by regularly flown super-wide angle photography and that the Type 4 Terrain would be covered by miscellaneous wide angle photography taken through breaks in the cloud cover.

Assumed Costs

These are based on operational costs at Divisional level and do not include any component for Departmental overhead or for Divisional overhead that is likely to be common to both techniques.

The assumed costs are approximate only and are of necessity based on many arbitrary assumptions but are reasonably proportional in respect of the techniques and procedures that have been considered.

In assessing the relative costs only Terrain Types 1 and 2 have been considered for Tellurometer traversing and Types 1, 2 and 3 for Aerodist trilateration.

It is not likely that Tellurometer traversing would normally be undertaken for Types 3 and 4 although it might be used along favourable routes and the control pattern extended laterally by airborne survey techniques.

It was found impossible to produce any worthwhile estimate for operations in Terrain Type 4.

In assessing the Aerodist trilateration costs, provision was made for additional A.P.R. flights along each degree of meridian connecting between the Aerodist stations, the elevations of which are usually determined by spirit levelling.

In allocating costs it has been assumed that an extensive area is to be covered and costs have therefore been shared between adjoining configurations.

Manpower Involved

This has been assumed at about 24 operative personnel in each case.