100kMapping

Control Surveys for 1:100,000 Mapping

J.D. Lines
Division of National Mapping
Department of National Development

Background

The compilation of the R502 map series was completed for the entire area of Australia at the end of 1966. This series, comprising 541 map sheets, will be published at a scale of 1:250,000 and includes a small percentage of maps contoured at 250 feet intervals, the balance being planimetric maps with relief shown by hill-shading techniques.

The Commonwealth Government is now actively embarking on a programme of re-mapping continental Australia over a period of 10 years at a scale of 1:100,000. This encompasses some 3200 map sheets of which about 45% will be published as fully contoured 1:100,000 maps in accordance with a new publishing specification now being prepared, and the remainder published as 1:250,000 topographic maps.

Re-photography of Australia is being undertaken with Wild RC9 super-wide angle photography at a rate commensurate with the new programme, on the premise that all 1:100,000 mapping will be undertaken only with this photography.

The Geodetic Survey of Australia will provide the basis for horizontal control extension for mapping purposes and the national programme of 3rd order levelling will provide the basis for vertical control. The geodetic survey is completed and the levelling survey has the year 1970 as the target for completion and a national adjustment.

Introduction

Australians for many years have become familiar with medium scale maps at 1:253,440 (four miles on the ground to one inch on the map) or 1:250,000 (practically the same), on the one hand, and the larger scale maps of 1:63,360 (one inch to 1 mile), 1:50,000 and far more commonly, 1:31,680, on the other hand.

This latter scale has, until quite recently, been widely used by the various States for basic topographic mapping, although it has never achieved any Commonwealth acceptance. Mapping at this scale has virtually been confined to methods embracing wide-angle photography, high quality and sometimes quite extensive ground control surveys, and photogrammetric block adjustments using observational data from either analogue or analytic equipment.

Between the two common post-war map scales of 1:250,000 and 1:31,680 there is an area where a re-thinking of methods could well yield quite valuable dividends in savings of money and time, combined with making the best use of the limited skilled manpower available in this country.

The control surveys made specially for 1:250,000 scale mapping were largely astronomical fixations, where observations were confined to one night only with one second theodolites, and time determinations that included the personal error of the observer. Although these "fixes" additionally contain varying and unknown components of the deflection of the vertical in each case, they offered, in general, an acceptable solution to the planimetric control because they were :-

(a) fairly consistent in observational error, ±3 seconds of arc,

(b) a unique determination of position not dependent on any other computations or adjustments of previous survey work,

(d) geographically placed specifically for mapping purposes.

As the Geodetic Survey of Australia progressed and it became possible to compare positions of points based on astronomical fixations with their corresponding geodetic values based on both the old Sydney origin and Clarke 1858 spheroid and the new Australian Geodetic Datum (1), it confirmed the premise that the astronomical control provided a better overall result for geographical position for 1:250,000 mapping than the existing nets of triangulation of various orders of accuracy.

Also, as the older triangulation results were progressively shown to have shortcomings of one sort or another by the geodetic surveys conducted with modern equipment, and calculated and adjusted with very powerful electronic computers, there emerged a period in recent years when surveyors made new closed circuit surveys around the perimeter of areas which were to be mapped at larger scales, e.g. 1;31,680.

These surveys were often made on the basis that, provided a single map sheet or a small group of map sheets were rigorously controlled for scale, azimuth and height, the fitting of the topography to a national framework would be a residual job for some other time in the future. Under this system, of course, there could be a large number of self-contained surveys based on different origins, the origin being usually one or two major triangulation stations, not at that time incorporated into the Australian Geodetic Datum.

This sort of thing is a natural corollary to any mapping programme which is not soundly based on a national homogeneous geodetic framework, properly classified in order, properly marked on the ground and properly recorded and readily available through a data retrieval system.

Or a corollary to having no truly national system at all! We are not in this position any more.

Apart from the saving grace that we are, in Australia, now commencing a transition from yards and feet to metres in our topographic mapping presentation, the amount of duplication of effort necessary to convert maps based on other than a national datum, to the national datum, is a luxury which developing countries cannot afford.

With modern equipment and techniques supported by forms of transportation unknown, or unavailable, to surveyors of even 20 years ago, there are now less and less reasons and excuses why mapping programmes and surveys should not be connected to the national framework of survey control.

In Australia, there is reason to take some comfort from the despatch with which the geodetic survey of the continent has been completed and results made available. Reference to the amount of work achieved in the last 15 years with the moderate resources used, is evidence of what can be done.

One important lesson which did come out of the National Geodetic Survey is that first order traversing and triangulation and geodetic astronomy has shed a lot of its occult reputation, and demonstrated to the profession at large that it is possible, these days, to set a reasonable specification for accuracy and actually maintain and often improve on the accepted standards, while keeping up a high productive rate.

That this has been done is a tribute to the skill and persistence of all who were associated with the Survey and brings out the point that there is no great difficulty in sustaining a good class of work, and that there is, consequently, no point in performing low order work with the equipment and techniques currently available, when it is just as easy from an observational point of view, to produce higher order work.

Very frequently, when it is claimed that third order work takes only, say, half the time necessary for similar work to second order standard, the claim is probably traceable in large measure to the differing quality of station marking. Permanence and easy location of stations are of vital importance.

The quality of station marking is a vexed subject, particularly when considerations of size and composition of the mark are weighed against future vandalism and ignorance, cost, available materials and other factors. Nevertheless, as nearly all surveying authorities find it difficult to initiate and sustain a station mark maintenance programme, it is in the community interest that careful consideration be given to this most important aspect of our work.

It can be argued that a survey station of any order, badly marked, has an initial and future potential cost out of all proportion to its community value, and reflects little credit on its originator.

To those who are in difficulties in controlling the quality of marking, there is good insurance value in a 35 mm photograph or photographs of all works undertaken in establishing a survey station. Other fringe benefits accrue also from these photographs.

Horizontal Control For 1:100,000 Mapping

Two methods of survey have been initiated for control in areas where the Division of National Mapping has the responsibility of directly producing 1:100,000 maps. These are:

(a) Second order ground traverses using electronic distance measuring equipment.

(b) Trilateration using airborne electronic distance measuring equipment, (Aerodist).

As a general rule, traversing has, and will be confined to areas where the topography, ease of movement and scarcity of timber and scrub allows this type of survey to achieve a rate and cost of production of map control points comparable with that achieved by airborne methods. There are quite firm indications that in favourable terrain, the cost and rate of control per square kilometre undertaken by ground traversing is quite favourable, as compared with Aerodist.

Aerodist has quite definite cost advantages in timbered terrain and areas where access and movement by ground vehicles is difficult or impracticable.

However, as it is extremely difficult to equate the methods over the varying types of terrain that each method might encounter in the course of a field season, it is equally difficult to distinguish any narrow zone on either side of which, between the extremes of topography, where one method can be accepted as being superior to the other.

Ground Traversing

Ground traversing in the normal situation leaves a much greater density of marked ground stations in its wake, which in the long term, is of a greater benefit to the community. Against this background, the time scale of the 10 year programme intrudes, and all methods must ultimately be judged according to the capacity to provide the requisite ground control in a period not exceeding 8 years.

In some areas of Australia, ground traversing can again be advantageous in supplementing existing triangulations, where these are not providing sufficient coverage, or alternatively are not strategically placed for control of a block of map sheets. In these areas, the introduction of Aerodist is not economically or practically sound.

It should be interposed here, that as from about 12 months ago, the great bulk of Australia will be re-photographed on a standard flight plan. This plan can operate in all areas where the range in elevation of the terrain is less than 2000 feet approximately. Beyond this range, special flight planning is necessary.

One of the many virtues of this, is that control surveys may precede photography, although this should be avoided as far as possible, but it ensures that control points are placed in the common side laps of mapping photography, and as far as possible adjacent to sheet corners. This may be considered an unnecessary nicety, but it leads to orderly compilation procedures, particularly in the region of projection zone boundaries.

Ground traversing so far has been run along parallels of latitude at 30 minute spacing, all work closing on portions of the geodetic survey. This system provides ground control at each corner of a 1:100,000 map sheet and provides additional points along the north and south edges. As an example of what can be achieved, one party of 13 traversed to strict second order standards, approximately 1100 miles in the desert area near the coast in Western Australia in the 1966 field season.

All ground stations are additionally marked by some device on the ground that lends itself to an unambiguous identification of the point on spot photography.

Spot photography is a well established technique in the Division, and for 1:100,000 mapping purposes, it is intended that no control point shall be used for photogrammetry unless it has been successfully transferred to the mapping photography from spot photography.

The zoom stereoscope has been found to be very effective for this transfer procedure.

It should be noted that pre-marking of ground control stations for the RC9 aerial photography contracts is not considered to be practicable. The very large areas to be photographed each year would demand a maintenance and inspection programme of the markers that would be prohibitively expensive both in cost and manpower.

Aerodist

This technique has been extensively described in Australian and Canadian literature, as set out references, (2), (3), (4), (5), (6).

Quite a large stock-pile of observational data has been obtained in readiness for the 1:100,000 mapping programme, and this is now being finally adjusted by use of the variation of co-ordinates programme on the CDC 3600 computer.

The reduction of Aerodist charts and preparation of data for the computer adjustment has been entirely a manual operation up-to-date, and with limited resources to undertake all this work, together with the inevitable problems requiring investigation that come to the surface over the space of three years experience, it is only now that we are rapidly closing on the day when published co-ordinates will be available.

Some degree of automation is close to being introduced in the form of 2 punched tape digitizer with the input from a chart reader, where the operator follows the primary trace with cursors and adds other information from the secondary codes to the tape through an electric typewriter keyboard. This equipment, as with many new items of equipment, has been troublesome in initial performance, but when in full-time operation, should speed up chart reduction quite considerably, and subsequent publication of results.

As with the national geodetic adjustment, no interim co-ordinates will be published, as these only add confusion to an already exacting task in recording and dissemination of survey information in a country the size of Australia.

The Aerodist system is now a working tool which can be relied on to undertake a forecast programme of work in a field season. The field party is now supported by a caravan which doubles as a mobile workshop and field office. It is equipped with a portable power supply and a range of test equipment and spare parts sufficient to enable maintenance to be carried out in the field as required. This includes provision for frequency measurement, so that a continual check can be kept on drifting A+ A- corrections, and index error, which has shown out over a period as a factor to be checked and allowed for in measurement reductions.

Average progress in a normal configuration i.e, a double-braced quadrilateral, is 2.5 quadrilaterals per week. This figure is based on terrain types which allow the remote stations to be occupied by ground vehicle. In this type of terrain, it is the practice, as far as possible, to have the stations established along roads and other places where ease of access and re-occupation are ruling factors.

Where the programme moves into inaccessible country, the reconnaissance, station marking and site preparation has been effected with a helicopter. This practice leaves a clear-cut operation for the following measuring party operating with a twin-engined fixed wing aircraft, and a helicopter for the positioning of the remote stations. With this type of organization the measuring progress will keep pace with that expected in open country, and importantly keep aircraft and skilled operator utilization up to a high level. There is a sizeable investment in the making of skilled operators, and this fact combined with the high capital cost of equipment, justifies the additional transportation costs in the interests of full utilization, and the requirement of an upper time-limit on the acquisition of ground control.

In the measurement configurations used to date there have been 3 basic designs:

(a) 30 minutes of arc x 30 minutes of arc quadrilateral, double braced.

(b) 1 degree of arc x 1 degree of arc quadrilateral, double braced.

30 minutes x 30 minutes (a)

This has been used in an area in Queensland, where the State has also a future requirement for 1:50,000 and 1:25,000 mapping. Overlaid on this pattern is what amounts to a separate pattern obtained by measuring directly, the braced 1 degree x 1 degree quadrilaterals.

1 degree x 1 degree (b)

This was the initial approach to the density of control required for 1:100,000 scale mapping, and was based on indications arising from photogrammetric test work in the Division, and being the first field operation attempted, was also a test bed for the field performance of the equipment. At this stage the horizon camera has not been introduced.

1 degree x 1 degree etc. (c)

With the decision setting out the priorities and scales for mapping of different areas of the continent, a further modification of the design was introduced whereby a centre point was included in each 1 degree square. These points were supplemented with a further point to be fixed along the mid-side of each 1 degree square by an airborne trilateration technique. This, in effect, provides control at 30 minute intervals in latitude and longitude, or at the corners of each 1:100,000 map sheet.

Trilateration

The trilateration technique used embraced the simultaneous exposure of a 70 mm camera in the near vertical and a Wild HC-1 horizon camera in a combined mount with the instant of exposure recorded on the Aerodist chart recorder during 3-channel operation. The mount is manually levelled. Theoretically, with the space co-ordinates of the aircraft at the instant of exposure available, and the ability to fix the nadir of the near vertical simultaneous photograph from the horizon camera exposures, this is the answer. However, this has not transpired in practice, primarily due to the fact that spectroscopic film, as recommended by the makers of the horizon camera, is not a practical proposition for field use in Australia, and the substitutes, combined with hazy horizons, do not provide the necessary consistently acceptable results.

There are other approaches to this problem and these are being investigated. One method under consideration involves the installation of a normal or wide 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.

Control Density

This matter is still under investigation and will not be decided until actual production tests of a large group of 1:100,000 map sheets have been completed.

In the area in Queensland covered by the 30 minute figures, there is an opportunity to extend the horizontal control photogrammetrically by use of differing densities of ground control provided by Aerodist methods. There is some second order ground traverse control available to provide a check on residuals arising from differing combinations of density.

This test, under production conditions, will provide an excellent guide to the density of control required where only 1:100,000 maps are intended, and how far it is economically justified to go in providing control for future larger scale mapping.

These decisions will undoubtedly be influenced by the photogrammetric methods to be adopted, and here is a vital decision which must take account of many factors.

Any decision must always be taken in respect of our collective national ability to complete the mapping of Australia in 10 years. Any other benefits that can be achieved through added refinements or greater accuracy than is necessary, without significant cost increases or depletion of resources and which allow the time scale to be met should be considered, but not at the expense of disrupting the programme.

Range of Measurement

Using the line crossing technique, occasional lines of 200 – 250 km 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, if tests show that larger quadrilaterals than those already mentioned can be used.

Accuracy Attainable by Aerodist Survey

The determination of station elevations has a 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 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 seem better still. However, local operation over a three year period indicates that about ±4 metres would be a more appropriate figure.

Investigations carried out indicate that quite large blocks of Aerodist trilateration adjusted to perimeter control (assumed free of error) will result in average coordinate 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. From results available, it would seem that a coordinate precision can be expected of approximately ±1 to 2 metres relative to surround control.

Vertical Control

Airborne surveying techniques already developed, together with those under development and those that will surely come with advancing technology, are supplanting the hitherto classical surveying methods in many aspects of data acquisition for mapping purposes. These methods have come to stay, and although frequently the initial capital cost is high, the improvements in accuracy will eventually provide direct measurements of an order of accuracy that will certainly suffice for topographic mapping of many scales and make the necessity to derive data, less and less necessary.

These methods will require the provision of basic data networks as a reference, and in the field of vertical control, the third order levelling network being built up over Australia, is providing this basis.

In the 1:100,000 mapping programme, the approach of using airborne survey measurements as vertical control for individual models, is being pursued in many areas.

Airborne Profile Recording

The equipment in current use is the Canadian airborne profile recorder which is operated on contract to the Division. This equipment utilizes the radar principle which has a cone of radiation at the aircraft of order of 1.5 degrees. The radar profile is presented as pen traces on a paper chart and is related to the terrain by exposures of a vertical 35 mm frame camera, which are automatically recorded as an event on the paper chart.

Apart from the inherent uncertainties in establishing which point on the ground is "heighted" by radar in rough terrain, there is also the problem of correctly establishing the gradient of the isobaric surface along which the aircraft flies in the course of recording a profile. The standard length of a profile is presently about 100 miles, and these profiles are run along the common side laps of the mapping photography.

Although a practical every-day procedure has yet to be evolved for office processing for compilation purposes, it is now evident that provided due care is exercised in equipment operation and full recording of all pertinent data in the aircraft, and opportunities are taken or made to compare the profile with known ground heights, the profiles used intelligently will provide individual heights on each model not in error by more than 25 feet. When the model is levelled to more than 3 heights, the overall result will be much improved.

This method, while not perhaps completely removing the necessity of bridging for height, should reduce bridging to a very small number of models in isolated instances. The use of profiles would have its greatest application where photogrammetric adjustments are made for planimetric position only.

Laser Terrain Profiler

The laser technique, for so long confined to the laboratory, is now being developed, inter alia, to make use of its properties in the field of measurement. Already, a laser device has been constructed overseas to measure distances on the ground with an accuracy of 1 x 10^-6. Its potential has also been put to use in the construction of laser airborne profile recorders overseas for various purposes, including installation in a satellite.

Considerable research in the laser field has been conducted by the Weapons Research Establishment (W.R.E.), Department of Supply, and in 1966, the Department of National Development sponsored a project within the W.R.E. for the development of a laser terrain profiler for use by the Division of National Mapping.

Design studies have now reached the stage where the decision has been made on the type of laser to be used, and specifications for all other components are sufficiently advanced to enable ground trials to start in September 1967, followed by airborne trials later on in the year.

The laser will be a C.W. Argon Ion type operating on a wavelength of 4880A. The transmitter and receiver will be a Cassegrain optical system with a field of view of both the transmitter and receiver of 10^-4 radius, which is of the order of 20 seconds of arc. The height computer will provide an output directly in metres and in the finally engineered recording apparatus, it is hoped that the height record will be imaged on to the same roll of film as the terrain image from strip camera.

The laser will sample a spot on the ground of one or two square feet, and height measurements will be made at the rate of 50 per second, and the resolution of the laser in height measurement, will be better than 0.3 metres.

From this brief outline, it can be seen that the potential of this system is a very substantial increase in accuracy over the radar-type profile recorders, although the problem of determining the gradient of the isobaric surface used as a reference still remains. However, as the pencil beam of laser light is so small, the opportunities for calibrating on existing height control points on the ground are very much greater; and undoubtedly errors due to this source can be effectively minimized.

Johnson Elevation Meter

This equipment has lived up to its reputation of providing heights to √D feet where D is the distance in miles run. This figure relates to one-way measurement. It can be improved to about 0.6 √D with two-way measurement.

So far, this equipment has been used mainly for making comparisons with airborne profile records (A.P.R.), and the results of these have, on occasions, considerably improved the accuracy in profile heights in the absolute sense. It has also had a useful role in providing profiles of airstrips tied to the levelling network for use as datum surfaces for APR operations, and this it can achieve in a very expeditious manner.

In the flat country, where the road system is reasonably extensive, it has demonstrated that a network of heights can be obtained which are adequate for contouring for 1:100,000 purposes. Again, in this operation, all work is connected to the third order level network.

In Victoria, elevation meter heights have been obtained to supplement existing vertical control for use in photogrammetric block adjustment of a very large area of eastern Victoria being undertaken by the Victorian Department of Lands and Survey for a portion of the 1:100,000 mapping programme.

From the equipment and methods outlined, will primarily come the vertical control used on the 1:100,000 programme by the Division of National Mapping. These methods may be supplemented on occasions, by forms of barometric heighting using helicopters for transport, and in some instances, height control may be derived photogrammetrically, but as far as possible, the accent will be placed on the use of direct field measurements.

The Future

There are many interesting speculations concerning future methods to be employed in topographic mapping, and no doubt, the coming decade will see sweeping changes come into mapping techniques through the effects of automation.

Progress towards automation in surveying and mapping procedures is something that the profession will have to live with, although there are widely differing views on the impact of automation. Already automated stereoplotters with their potential for production of orthophotographs are on the market, and will improve in performance and capability from year to year.

One subject worthy of consideration by all engaged in topographic mapping concerns some means of establishing the nadir of aerial photographs at the time of exposure. There are 3 methods of achieving this which come to mind:

(a) Stabilizing the entire camera within its mount.

(b) By a remote sensing apparatus, determine the dis-levellment of the camera at the instant of exposure.

Although there has been quite an amount of investigation into these problems, a solution is still required, which enables the camera to operate in normal commercial-type aircraft. Modern Technology offers several methods of achieving this goal, but an answer must surely come from communication between scientists and those engaged in aerial surveying.

Finally, in Australia we are now fortunate to have a geodetic framework for mapping, which is modern, accurate, easily recoverable and for which information is freely available. The levelling network is now fairly extensive and in the next few years will encompass this country.

Not only the governmental mapping authorities, but all those who have direction of large projects, be they surveyors, engineers or other project executives, should use their endeavours to have all large area surveys connected to the national schemes.

These connections will not only provide checks on the work being undertaken, but would make worthwhile contributions to survey co-ordination and the benefits that will flow from this work.

References

1. Commonwealth of Australia Gazette No.84, 6th October, 1966.

2. Lines J.D. - "Aerodist in Australia 1963-64", Australian Surveyor, Vol.21. No.2, 1966.

3. Lambert B.P. and other officers of Division of National Mapping - "Aerodist Surveys - Report on operations carried out by the Australian Division of National Mapping since 1963"; presented at the I.A.G. Symposium on Electromagnetic Distance Measurement, Oxford, England 1965.

Lambert B.P. - "The Use of Aerodist for filling in between Tellurometer Traverse Loops" - to be presented at Commonwealth Survey Officers Conference, Cambridge, England, August 1967.

5. Turner L.G. - "Aerodist Operations in Australia", Document E/Conf. 52/L44.

6. Anderson Major E.U. - Operation of Aerodist distance measuring equipment in Papua - New Guinea, Document E/Conf. 52/L53.

7. Numerous reports on Aerodist operations published by Department of Energy, Mines & Resources (formerly Dept. of Mines & Technical Surveys), Canada.

Tuttle A.C. - "Aerodist in geodetic surveying in Canada". Document E/Conf.52/L95.

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