XIII Congress of the International Society for Photogrammetry Helsinki, 1976 Commission V lnvited Paper

K. B. ATKINSON University College London London, WClE 6BT, United Kingdom A Review of Close-Range Engineering Photogrammetry

Recent applications of close-range photogrammetry to engineering, and some of the problems encountered, are discussed.

INTRODUCTION bonnell, 1974) is the third in a comprehen­ T IS POSSIBLE to gain a certain degree ofsatis­ sive series of reports on developments in ar­ I faction from the achievements ofCommis­ chitectural photogrammetry which have sion V ofthe International Society for Photo­ taken place since Carbonnell (1969) gave an grammetry during recent years. Non­ invited paper to this Commission at the XIth topographic photogrammetry is being International Congress of Photogrammetry in applied in an ever increasing number of Lausanne. Attention has since been focussed ways. More and more measurement prob­ upon medical applications of photogram­ lems are being solved by photogrammetric metry and the Commission V symposium in techniques. In 1968, Commission V directed 1974 was devoted to this topic (Karara and

ABSTRACT, There is little doubt that photogrammetry will be applied to an increasing extent in the solution of measurement problems in engineering. Already there are indications that the number and vari­ ety ofapplications is growing. An attempt has been made to review recently published examples, the majority of which have been re­ ported from the United Kingdom. They include the recording and measurement ofdam displacements, unstable geological structures, snow cover, soil, structural and hydraulic models, constructional problems, and box girder load tests. The paper concludes by indicat­ ing some of the problems which beset these applications of close­ range photogrammetry. Their solution would help to further the wider acceptance of these techniques. our attention towards the needs of architec­ Herron, 1974). An International Exploratory ture, archaeology, and the conservation of Committee on Biomedical Photogrammetry what is worthwhile of our man made herit­ has since been seeking ways of furthering age. The result was the establishment ofthe this development. International Committee on Architectural It would appear that engineering or indus­ Photogrammetry which has since fostered trial applications of photogrammetry might developments in this field with conspicious offer a further rewarding field of activity. success. Their most recent publication (Car- Those who are in a pessimistic frame ofmind

PHOTOGRAMMETRIC ENGINEERING AND REMOTE SENSING, 57 Vol. 42, No.1, January 1976, pp. 57-69. 58 PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING, 1976

may say that several attempts to apply close­ significant application in shipbuilding, range photogrammetry to engineering prob­ which was described by Newton (1974), has lems have met with only limited success and now been joined by the establishment of a their widespread adoption has not occurred. photogrammetric unit associated with the Lacmann (1950) described many examples in construction ofoil rigs. Both aspects will be his text book, several of which date from the fully explained in another invited paper* to early part ofthe 20th century. To this writer's be given to the XIIIth International Congress. knowledge, applications in the motor and These developments have resulted from co­ chemical industries were eventually discon­ operation between the University ofNewcas­ tinued, despite the proven success of the tIe upon Tyne and the British Ship Research photogrammetric methods (Farrand, 1965). Association in the first case, and between the This appears to have been because, very of­ University and Longdin and Browning (Sur­ ten, the engineer chooses to use those veys) Ltd. in the second. methods and equipment with which he is Because they are the subject ofan invited familiar, unless the alternative offers distinct paper**from the United Kingdom, it is ap­ advantages. Ifphotogrammetry is just as good propriate to mention here the impact of co­ as, but no better than, an established tech­ herent optical methods on measurement nique, it is unlikely to be adopted because it problems in engineering. Developments requires expertise and equipment which are have already been reported by Gates (1975) relatively unusual. However, there are cir­ and Burch and Forno (1975) of methods cumstances in which photogrammetry may which are being evolved at the National be applied to advantage and this paper enu­ Physical Laboratory and, in the latter case, in merates some ofthem and indicates possible collaboration with the Transport and Road lines of development. Research Laboratory. Burch and Forno were specifically concerned with measurements of RECENT ENGINEERING ApPLICATIONS. WITH deflections of a large beam under load by PARTICULAR REFERENCE TO WORK IN THE UNITED using techniques which, hitherto, had not KINGDOM In an attempt to limit the coverage of this paper, it was decided to review engineering applications of photogrammetry which have been published since January, 1972 or which are known to be in progress. The majority of examples which will be cited have originated in the United Kingdom but similar.work has taken place in a number of other countries. January, 1972 was chosen as a suitable date because most work reported at the XIIth International Congress ofPhotogrammetry in Ottawa would have been prepared or pub­ lished before that date. Readers in search of examples of earlier work could consult, for example, FIG. 1 A scheme to double the carriageway ofa Bulletin de la Societe Franc;aise de road near Plymouth, Devon involved a section of Photogrammetrie, 22 (1966) which contains excavation in rock. The southern extremity is articles on the application of close-range shown in one ofthe survey photographs, together photogrammetry in the hydro-electric power with adjoining property and a retaining wall and industry;joumal ofjapan Society for Photo­ slope stabilisation measures. The determination grammetry, Special Volume 2 (1966) which ofquantities for a further extensive retaining wall contains the papers given at the Commission could be made from the results of a photogram­ V symposium ofthat year; and reviews by the metric survey which was carried out by Univer­ author (Atkinson, 1968 and 1972b), both of sity College London and Cartographical Serv­ which include sections on, and references to, ices (Southampton) Ltd. for Devon County close-range engineering photogrammetry. Council. The period under review has seen further gradual development of non-topographic Invited papers, Commission V, International Con­ photogrammetry in the United Kingdom. gress of Photogrammetry, 1976. * Close-range photogrammetry as an aid to mea­ Engineering applications form a substantial surement of marine structures, by 1. Newton. section of this development. Perusal of the ** Three-dimensional location and measure­ Directory which was compiled and edited by ment by coherent optical methods, by J.W.C. Dowman (1974) confirms this impression. A Gates. A REVIEW OF CLOSE-RANGE ENGINEERING PHOTOGRAMMETRY 59 been easy to apply away from the stability of conducted a photogrammetric study ofjoint laboratory surroundings. influence on weathering ofsteep chalk faces; An area in which commercial photogram­ it has determined the joint disposition and metric companies have been able to contrib­ material volume after slide failure in Oxford ute to the application of close-range tech­ Clay; and it has monitored a deep, steep rock niques is thatofdams and rock outcrops, both cutting on the Trans-Pennine (M62) motor­ natural and excavated. Fairey Surveys Ltd. way. carried out a survey of Edinburgh Castle An interesting variation on surveys of this Rock (Cheffins and Rushton, 1970) in 1969 type and magnitude of structure is reported using a Wild RC5A camera, and since that from Poland (Butowtt et ai., 1974) where time several other companies have been in­ studiesoferosion ofa numberofscarps on the volved in similar surveys in places as Widely banks of the River Vistula have been carried separated as Muckle Flugga in the Shetland out. A modified Fairchild air survey camera Isles and Plymouth in Devon (Figure 1). The (cf. Cheffins and Rushton, 1970), carried on a engineering contractors need these surveys launch, was used for photography, and con­ because ofthe instability ofthe rock surfaces. toured plots and profiles were derived from Rock bolting, construction ofretaining walls, the photography. rock blasting and related safety aspects, and Collaboration between the Institute of prefabrication ofstructures all require the in­ Hydrology and The City University has led to formation provided by, typically, a 1:50 scale the use of terrestrial photogrammetry in an survey with contours or isometrons* related analysis of snow distribution (Blyth et ai., to a vertical datum plane. A dam is a structure 1974). A research catchment on the eastern of similar size and its behaviour is of the ut­ slopes of Pumlumon Fawr in Wales con­ most concern; photogrammetric surveys of tained a test area, 300 m x 100 m in extent and dams have been reported in the United King­ ranging in height between 460 m and 525 m. dom and elsewhere. The Building Research This site was photographed, both free of Station and Hunting Surveys Ltd. were snow and under snow cover ofabout 100 mm jointly responsible for the photogrammetric depth, from each end of a 100 m base. En­ study of the constructional displacements of couraging results for the determination of the rockfill dam at Llyn Brianne in mid-Wales snow depth when compared with site checks (Moore, 1973). The dam is 90 m high and may lead to the use of photogrammetric 200 m wide. A Wild P30 phototheodolite was monitoring in connexion with provision of used to provide stereoscopic coverage, at hydrological data for use in flood warning eight different stages ofconstruction, ofboth systems. Photogrammetry may also be used upstream and downstream shoulders. in an examination ofthe physical processes of Ground control was established on the valley snowmelt. sides, together with 80 targeted points at var­ Fairey Surveys Ltd. have, over the years, ious levels on the dam. Three dimensional shown a refreshing readiness to attempt the co-ordinates were required of these pre­ solution of a variety of measurement prob­ marked points and in fact three dimensional lems. Reference already has been made to displacements were determined to an aver­ Edinburgh Castle Rock (Cheffins and age accuracy of0.05 m. Profile lines through Rushton, 1970) and, in a recent paper, Chef­ the dam also were obtained. Brandenberger fins (1975) has described other applications and Erez (1972) and Brandenberger (1974) of non-topographic photogrammetry in also have reported on dam surveys which in­ which his company has been involved. Two volved photogrammetric techniques. Bran­ of the applications which are described in denberger achieved greater accuracies at that paperare ofinterest at this juncture. The Outardes, Quebec than Moore reported from first concerned the determination of distor­ Wales but ofcourse so much depends on the tions in the area adjacent to the static vent ofa configuration of the site and on the equip­ BAC 1-11 aircraft. Close-range stereometric ment which happens to be available for the photography ofan area just in excess of 1 m2 survey. was analysed to provide a regular grid of The Building Research Station has also depth values ofthe area ofinterest. The cam­ eras if= 150 mm) were about 1.5 m away from the fuselage and the overall repeatability of * O.C. Gibbins has used the tenn isometron for depth measurements from four different sets lines which pass through points having the same depth or distance from a vertical plane. See Gib­ ofphotography ofthe same subject, when an­ bins, a.c., 1970. Searching for hydroelectric alysed photogrammetrically, was ± 0.36 mm powerschemes in Tasmania with special reference root mean error (or 1 part in 4200 of to the Dove River. Cartography, 7(2): 59-68. the object distance). 60 PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING, 1976

The second engineeringapplication which the centrifuge in order to take survey photo­ Cheffins (1975) described concerned the def­ graphy. Synchronised with the two cameras inition of the complex shapes of a steel re­ is a powerful high speed flash unit which pro­ spirator mould. Photography and plotting vides 100 J for 2.5 fJ-S. The accuracy of meas­ both were achieved with a modified William­ urements is claimed to be about ± 0.2 mm son multiplex unit (described by Cheffins which is considered to be acceptable. and Clark, 1969) which allows the reprojec­ Butterfield et ai. (1970 and 1972) and An­ tion ofa distortion free 1:1 scale stereomodel. drawes and Butterfield (1973) have made use During the period under consideration, offalse parallax techniques in their studies at Griin and Stephani (1974) published an ac­ the University of Southampton in soil count ofa photogrammetric survey of a Ger­ mechanics and in simple channel flow in man atomic reactor and they reported that hydraulic experimentation. The fluid surface standard errors of± 1 mm in plan and ± 2mm was enhanced by paper tape punchings so as in height were achieved when a maximum to improve stereoscopic pointing. Similar object distance of 10 m was involved. Some work is reported from the University ofBris­ years ago, Fairey Surveys Ltd. were active in tol in Dowman.(1974) and from Switzerland a similar field and, although no results were by Schmid (1973). published, a personal communication from In this section of the paper, it has been Cheffins (1967) gave some details. In the possible briefly to review photogrammetric course of designing reactor vessels for nuc­ activities in branches of engineering which lear power stations, the Central Electricity include marine structures, dam displace­ Generating Board constructed and tested ments, scarp erosion and rock face instability, model pressure vessels. During the testing Sl10W depth measurement, aircraft vent de­ procedures, it was necessary to measure sur­ formation and respirator mould shape, atomic face distortions which took place in the vicin­ reactor geometry, and soil and hydraulic ity of 1 m diameter panels which contained model behaviour. These applications of several nozzles. A panel was located on the close-range photogrammetry have involved underside ofa vessel which was 3 m in diame­ research establishments, commercial com­ ter and more than 13 t in weight. The vessel panies, and university departments. Some of was installed in a pit for safety reasons and for the British survey companies already have some of the trials it was subjected to high received specific mention but others also are temperatures as well as to high pressures, involved in close-range applications, notably which· made direct measurement of a test Meridian Airmaps Ltd. who have carried out panel difficult ifnot impossible. Stereoscopic tunnel surveys, bridge surveys for restoration photography was taken by phototheodolite at work, landslip analysis and displacement two successively occupiedcamerastations on measurement of earth retaining banks. the floor of the pit and with the camera axis Photarc Surveys Ltd. has been formed since pointing vertically upwards. Later, a pair of 1972 and this company has a special interest suitably modified Kelsh plotter projectors in close-range engineering photogrammetry. were used to acquire photography. Co­ There will be more to say about the contribu­ ordinate analysis was carriedout on a Wild A8 tion of the commercial survey companies in with machine co-ordinates transformed to a the concluding section of this paper. How­ control system in the object space. ever the author would now like to tum to One of the most fruitful fields of applica­ some projects with which he has been in­ tion for close-range photogrammetry lies in volved at University College London. the analysis of models. In engineering, a number of interesting examples have been BUILDING AND STRUCTURAL ENGINEERING reported. Particularly noteworthy is the cen­ trifugal testing of soil models, initiated by Figures 2 and 3 show examples of precast Professor A.N. Schofield at the University of concrete panels which were contoured at 2 Manchester Institute of Science and mm vertical intervals in order to show the Technology. Photogrammetric aspects of extentofdeformation from an ideal plane sur­ these tests were reported by E1-Beik (1973). face. Figure 4 shows a building constructed They began in 1969 and serve to obtain pre­ of precast units and Figure 5 illustrates the cise measurements of surface movement of deviations of panels from the vertical in that soil models which rotate in a centrifuge at building. This kind of information on man­ speeds up to 45 ms-1• Tests are monitored by ufacturing and erection tolerances was pro­ closed circuit television and, at appropriate vided for the Building Research Station be­ orcritical stages ofa test, a pairofZeiss (Jena) tween 1961 and 1963 ata time when solutions UMK 10/1318 cameras can be lowered into were being sought for these construction A REVIEW OF CLOSE-RANGE ENGINEERING PHOTOGRAMMETRY 61

FIG. 2 & 3. Examples of precast concrete panels which were photographed and con­ FIG. 5. Sample measurements of the build­ toured in order to assess flatness. ing (Fig. 4) which show departures from the vertical in the YZ plane. The depth(Z) scale is in mm at 1:1. The height (Y) element is dia­ grammatic.

gradual movement or settling of the tower, over a long period, produced numerous frac­ tures in the fabric. The Ministry of Public Building and Works (now a part of the De­ partment of the Environment) were anxious to strengthen the tower against further struc­ tural deterioration or possible failure and, in order to assess the position and to decide upon remedial measures, large scale draw­ ings of the tower elevations were needed, showing all masonryjoins and fractures. Plot­ ting at 1:50 scale conveniently allowed all FIG. 4. A buildingat Aldershot, Hampshire joins to be shown. Small fractures could, which had just been erected using precast panels. however, only be represented by a single line while both sides oflarger gaps could be plot­ ted. At first sight, the fractures appear to problems in the building industry (Atkinson have been exaggerated or drawn more heav­ and Newton, 1968). ily but this appearance derives either from Huby's Tower, Fountains Abbey (Figure 6) the double line representing both sides ofthe is a ruin, approximately 47 m high and 18.5 m fracture or from the irregular pattern con­ wide at the base. It was built about 1500 and trasted with the regularcourses ofstonework. has been in a ruinous condition since the A full account ofthis survey was given to the dissolution of the monasteries by King symposium of ISP Commission V in Paris Henry VIII in 1539. The four walls have no (Atkinson and Proctor, 1970). internal or external structural support and a Model structures have been subject to 62 PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING, 1976

which now provides partial cover for the Farahabad sports stadium near Teheran. The + - roofconsists ofa numberofsuspension cables whichare connectedatone endto a reinforced concrete spatial beam and atthe otherend to a + flexible cable which is in tum supported by two pin-ended steel pylons. A family of pre­ stressing cables is arranged orthogonally to the suspensioncables. Theobjectofthe inves­ + + tigation, carried out in the Space Structures Research Centre ofthe University ofSurrey, was to check the preliminary design and to provide information for the final design ofthe cable roof. The detailed shape ofthe net was determined at one stage during testing by employing close-range photogrammetry (At­ kinson, 1972a). Stereometric photography of the model and five premarked control points was taken with an Officine Galileo camera mounted on the laboratory ceiling above the model (Figure 7). From observations in a Thompson-Watts Mk.2 plotter, the co­ ordinates of all points of cable intersection were determined with a standard deviation of ± 0.1 mm in plan and ± 0.3 mm in elevation. The photogrammetric technique provided a simple and elegant solution to a measure­ ment problem which would have been dif­ ficult to solve by other means.

A PHOTOGRAMMETRIC WRIGGLE SURVEY In recent years, construction has been tak­ ing place ofa new twin tunnel underthe River .. +...... + Mersey between Liverpool and Wallasey. FIG. 6. Amuch reduced copy ofthe plotted The consulting engineers, Mott, Hay and An­ west elevation of Ruby's Tower, Fountains derson, were in possession ofsuggestions for Abbey, Yorkshire. The distance between the provision of a photogrammetric wriggle neighbouring grid intersections is 5 m. The survey made by the Swedish photogrammet­ original plot was at a scale of1:50. Structural rist, Dr. (now Professor) K. Torlegard. In order weaknesses and masonry cracks are clearly to test the feasibility ofthis scheme, we were shown and are most numerous just below the lower Latin inscription.

photogrammetric analysis in recent years. The work of Professor K. Linkwitz and his colleaguesatthe UniversityofStuttgartis well known in connexion with the digital model­ ling of roof structures for the Montreal Expo 67, the Munich Olympic Stadium and, re­ cently, the Mannheim garden exhibition (Linkwitz, 1967; Linkwitz and Preuss, 1971 and 1974; Happold and Liddell, 1975). At University College London, we have been involved in two modest experiments, one of whichanalysedthe effectofwinddeformation on model high rise flats and the other deter­ FIG. 7. One photograph of a stereopair ofthe mined the geometry ofa 1:50 scale model ofa structural model ofthe Farahabad stadium roof. prestressed cable roof. This experimental in­ Five control points may be seen, as well as the vestigation, fully described by Nooshin and 207 cable intersections which were co­ Butterworth (1974), related to a cable roof ordinated. A REVIEW OF CLOSE-RANGE ENGINEERING PHOTOGRAMMETRY 63 invitedto undertake a trial surveyofa length of and Technology. Their investigations into tunnel which already had been surveyed by box girder behaviour, sponsored by the conventional ground survey methods. Highway Engineering Computer Board of A wriggle survey determines the extent to the Department of the Environment, have which a tunnel has departed, duringconstruc­ benefitted from the use ofclose-range photo­ tion, from its designed position. The Second grammetry as one of several measurement Mersey Tunnel is lined with precast units techniques. We have been involved in two which contain the road deck supports. If, for projects, the first of which has successfully example, these deviate from their designed been completed while the second is in prog­ position by more than the allowed tolerances ress at the time of writing (J uly 1975). there will be difficulty in constructing the The first project concerned the testing to road deck while maintaining the statutory failure of two steel box girder models which height ofvehicle clearances. A 50 m length of incorporated high strength friction grip tunnel was photographed in January 1970 splices. Harding (1974) gives exhaustive de­ with one stereopair, using an Officine Galileo tails of the tests. Their aim was to provide camera (Figure 8). Analysis on a Thompson­ data on the behaviour of large scale joints in Watts Mk.2 plotter resulted in a root mean complete structures so that existing design square difference between 151 photogram­ data on small elemental behaviour can be metric measurements and field survey values studied in relation to overall joint behaviour. of5.5 mm. This degree ofaccuracy is satisfac­ The box sections were manufactured from tory for the wriggle survey and commercial mild steel. Each section was 3 ft (0.91 m) high companies have subsequently undertaken by 8 ft (2.44 m) wide. Two of the three sec­ production surveys. If successful, a serious tions used in each test were 14 ft 3% in (4.37 obstacle in road and rail tunnel construction m) long while the third was 12 ft 5 in (3.78 m) will be overcome since photography does not long. These sections were connected by occupy a tunnel for the long periods required high strength friction grip boltedjoints. Only by ground survey. A very detailed account of one ofthejoints was under test; the other was the methods and results ofthis wriggle survey significantly overstrong. has been published by Proctor and Atkinson The main purpose of the tests was the ob­ (1972). servation ofjoint behaviour, but the observa­ tion of both spliced and unspliced panel Box GIRDER DEFORMATION MEASUREMENT buckling behaviour was of importance. For Since 1973, there has been collaboration this secondary purpose, 150 (test 1) and 100 between the photogrammetrists of Univer­ (test 2) strain gauge elements were sity College London and engineers in the positioned on web and flange panels to Civil Engineering Structures Laboratories at monitor panel strains while close-range neighbouring Imperial College of Science photogrammetry was used to measure out­ ofplane movement ofboth web and compres­ sion flange panels. Dial gauges also were used to measure overall box deflection and twist. A single Officine Galileo if= 150 mm) cam­ era was used in these experiments (130 x 180 mm format). It was mounted on an overhead crane with the camera axis pointing vertically downwards. An operator was able to sit with the camerato load plates and release the shut­ ter and camera and operator were moved by the crane controls (Figure 9). It was possible to tiltthe cameramountaboutx, y, z axes prior to each photograph so that tilts were reduced to a minimum and complete coverage was FIG. 8. A survey photograph of the length of ensured. Eight camera stations were oc­ the Second Mersey Tunnel which was subject cupied at each load stage of a test providing to an experimental wriggle survey. Co­ two strips offour photographs and, therefore, ordinates were determined on alternate rings (every 8 ft (2.44 m)). Points were observed on six stereomodels for analysis. The camerasta­ both premarked and natural surfaces. Concen­ tions were positioned 3.1 m vertically above tric bull's eye targets and paint markings were the web on each side of the box girder by used for premarking. The tunnel diameter was using a plumb bob so that the lateral overlap 31 ft 7 in (9.63 m). between the two strips of photography ran 64 PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING, 1976

~. FIG.9. A general view ofthe box girder FIG. 11. A view of a splice in the com­ test arrangement. The camera and cam­ pression region of the web showing era operator are suspended from the combined compression and shear slip. crane above the area ofinterest to the left Several premarked points are visible. of, and including, the bolted joint. Con­ trol bars can be seen immediately above the compression flange to the left ofthe with web depth so that all targets could be tie-down reaction beam. The left hand imaged from the relevant camera stations. end ofthe box is supported on hydraulic The cross dimensions were chosen to suit the jacks. scale of photography for a given depth. The control system consisted ofbars which longitudinally along the upper surface ofthe crossed the box girder (Fig. 9) along lines box girder. The six stereomodels covered an joining corresponding principal points of area which included the two compression photographs in the two strips. The bars were flange panels to the left of the bolted splice mounted on posts which were prestressed to which is clearly visible in Figure 9. All points the floor. Crosses were stuck to the bars at for which co-ordinates were required were known intervals. Additional posts were also premarked with photographically repro­ constructed outside the box and in line with duced crosses. The flange targets were stuck these bars to provide points for relative orien­ on the flange surface. Web points consistedof tation which were in focus. As these points crosses which were stuckonto studs and they, were fixed throughout a test, they were used in tum, were welded to the web plates (Fig­ to compute relative displacements ofthe box. ures 10 and 11). The lengthofthe studs varied Lighting was provided by two flash units

FIG. 10. The tie-down reaction beam (top left) and premarked points on rods welded to the web. The maximum length ofrods was 75 mm. Photogrammetric ob­ servations were made to all premarked points. The photograph shows buckles in FIG. 12. A view ofthe bifurcated box girderbridge web panels at failure. model. A REVIEW OF CLOSE-RANGE ENGINEERING PHOTOGRAMMETRY 65

principal distance with radial distance. In order to calculate correction terms for ob­ served co-ordinates, the approximate dis­ tance of each target will have to be deter­ mined. Control points on the bridge will take the form ofmarks on plumb wires, each 1 mm in diameter and 1 mm long. These wires are hung on the stable mounting frame of the bridge. They will be surveyed from four con­ crete pillars erected specially for the purpose. Control points will be placed at the front and back ofeach overlap area. The survey ofthese marks presents several problems and it will require much careful planning to keep the FIG. 13. The underside of the box girder horizontal and vertical scale factors equal. bridge model, showing some of the pre­ The nature ofthe box girder is such that the marked points whose relative movements are direction of deformation of any point can be to be determined. fairly well predicted. This gives rise to the possibility of using false parallax techniques which were attached to the camera and three on pairs of photographs taken from the same auxiliary flash units which were mounted point before and after deformation. Since around the box. The use ofa short flash dura­ there will be approximately 30 camera sta­ tion overcame any problems which might tions, it will not be possible to relocate the have arisen from camera movement. camera exactly. A fairly simple adaptation of the false parallax principle is being investi­ The photography was observed on a Hilger gated which requires that the photographs and Watts stereocomparator. Relative and ab­ need only be taken from approximately the solute orientations were computed analyti­ same pointand with roughly the same orienta­ cally. Accuracies in the horizontal plane were tion." good but movements in depth were deter­ ROCK MECHANICS mined less satisfactorily (average root mean Collaboration over an eight year period be­ square error at height control was 0.34 mm). tween the Rock Mechanics Group at Imperial However, their determination was not so im­ College of Science and Technology and the portant in these tests. Scale control atthe base Department of Photogrammetry and Survey­ of the box and triangulation of the control ing, University College London, led to sev­ points which were provided would have im­ eral applications of photogrammetry. Wick­ proved this accuracy. ens and Barton (1971) have explained how A second deformation measurement proj­ close-range photogrammetry was used to de­ ect is now in progress. The author is indebted termine roughness profiles of model joints to Mr. P.I. Scott of University College Lon­ (Figures 14 and 15). Similar model material don for the note which follows. was then used in the construction of two­ "A 1:12 scale model ofa bifurcated box gird­ dimensional open pit excavations. Stages in er bridge (Figure 12) is to be load-tested at these excavations and associated displace­ Imperial College ofScience and Technology. ments in the structure were analysed by false Measurements will only be required in the localised areas where the failure occurs but their location will not be known until the failure has taken place. Close-range photo­ grammetry thus provides the means ofrecord­ ing all points andofproviding co-ordinates for only the required few when their location is known. The model bridge carries approxi­ mately 4000 premarked targets (Figure 13). Relative movement between them is to be de­ termined to better than 0.3 mm. This require­ ment has necessitated the use ofa small object distance of 2.3 m. The shape of the bridge dictates the large depth of field of 1.8 m to 3.4 m. The majority of the related research is thus centered around the variation of the principal distance with object distance throughout the field. A photographic calibra­ FIG. 14. One photograph of a stereopair, taken tion method has been devised which evalu­ in orderto determine surface roughness ofmodel ates this variation as well as the change in joint material. 66 PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING, 1976

FIG. 16. The ultimate stage in model excava­ tion. The measurement of displacements around an excavation was carried out by false parallax methods. Movements of 0.1 mm were detected.

observed and recorded in the steromodel. Details of geological interpretation and joint selection, together with case histories and ac­ curacies, will be found in Ross-Brown et al. (1973). The co-ordinates ofthe points on the joints are used in a subsequent plane fitting procedure. This involves setting up the nor­ mal equations for a linear least squares fit and solving these nonnal equations. From the re­ sulting symmetric coefficient matrix the three direction cosines may be determined and subsequently transformed by means of the transformation matrix previously ob­ tained. The planes may then be described in FIG. 15. The block of20 samples which appears tenns of dip and dip direction or azimuth. in Fig. 14 reproduced with overlying roughness A related development of this application profiles. The profile data were digitised and has occurred in work carried out by the computer plotted. The information on roughness Transport and Road Research Laboratory was used to determine suitable materials for the with a Wild P32 in Colombia (Kempson and two-dimensional open pit excavations. Heath, 1972). The Laboratory investigated a number of landslides in order to detennine parallax techniques which Wickens and Barton slope angles, sizes ofcracks and fissures, and (1971) also explained (Figure 16). This work earth volumes. Analysis of the photography fonned part of the experimental background was undertaken by Meridian Ainnaps Ltd. to geologicaljointmapping ofopen pit mines. Practical photogrammetric techniques were CONCLUSIONS AND SUGGESTIONS evolved and modified for use in a mine situa­ There is a great deal ofactivity and interest tion. They were based on the Wild P30 photo­ in engineering photogrammetry although equipment (Ross-Brown and At­ much of the work is being carried out in a kinson, 1972). Analysis ofthe survey photog­ research context. When that research work raphy (Figure 17), some ofwhich has been car­ reaches a conclusion, there may be no further ried out by a commercial company (Meridian employment for the photogrammetric tech­ Ainnaps Ltd.), involves a conventional rela­ niques which have been used as a measuring tive orientation in a plotting instrument but tool. Other projects, though entirely com­ absolute orientation is a computational pro­ mercial, are often unusual or "one off' and cedure. This makes it possible to employ design, erection, setting out survey, control principal distances and photographic tilts or monitoring are peculiar to that project. In which some plotting instruments cOl,lld not cases such as these, there is still an enonnous accept. The orientation ofeach joint is deter­ need for education ofthe engineer so that he mined by four points whose co-ordinates are is aware of the full possibilities of photo- A REVIEW OF CLOSE-RANGE ENGINEERING PHOTOGRAMMETRY 67

FIG. 17. A survey photograph, taken at Atalaya mine in Spain, for geological joint mapping. Areas within a stereomodel are interpreted and marked on paper printenlargements. In anyone area, photogrammetric observations are made to determine the orientation ofidentifiable joints or planes. grammetry. There is also a place for liaison photogrammetric unit. Earlier in this paper, and co-operation between the university de­ the demise ofunits in the motor and chemical partment with research capabilities, the en­ industries was mentioned. Now we note the gineering consultant or contractor, and the establishment of a unit in the north of Scot­ commercial survey organisation. Methods of land associated with oil rig construction. An­ close-range photogrammetry can then be de­ other solution to this type of problem would vised and tested and, if found to be suitable, be for the survey company to offer the serv­ the engineer can use the methods himselfor ices of a mobile analysis unit. The appear­ sub-contract the work to a commercial survey ance of the Zeiss (Oberkochen) stereocord company. These survey companies have, on G-2 prompts this suggestion, for the provision occasion, been reluctant to involve them­ ofco-ordinate information ofthe kind which selves in non-aerial photogrammetry be­ is required in oil ~ig fabrication could be pro­ cause, to do so, would be uneconomical in vided by an organisation with this type ofequip­ their estimation. It would be safe to say that ment. Proctor (1973) summarised these prob­ this view has changed and there should be a lems when he noted that "most ofthe equip­ reasonable return on those photogrammetric ment and expertise is to be found amongst, surveys which involve either repeated pho­ and has been developed for, the tography (over months or years) from the profession. There is no scientific reason why same survey stations, as in construction work, this should remain so but there are practical or where there are a large number of photo­ problems, equipment and education being graphs to be analysed. Unfortunately, com­ foremnst, in extending photogrammetric panies have been guilty ofextremely expen­ practice to all user disciplines." sive quotations for photogrammetric analysis, Mention of equipment brings us back to a possibly because they did not really want the frequently ridden hobby horse! Strictures on work, but the effect on the engineers and on the non-compatibility ofcamera and plotting the development ofthe subject was not help­ instrument principal distances have been ful. Several government research establish­ made frequently (for example, Atkinson, ments in the United Kingdom possess photo­ 1974). Now the problems posed by the in­ grammetric cameras and it should be possible strument manufacturers have been joined by for them to seek quotations for the analysis of the photographic material manufacturers' re­ their photography. It is rarely possible for an luctance to supply glass plates. Constructive engineering problem to warrant an exclusive suggestions for the equipment manufacturers 68 PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING, 1976

include the provision ofan improved means Atkinson, KB., 1972b. Special applications of of detennination of instrument height. In photograrnmetry. Chartered Surveyor, 104(10): providing control for the photogrammetric 495-497. survey ofa high structure, probably the most Atkinson, KB., 1974. The Wild P31 terrestrial camera. Survey Review, 22(174): 375-378. critical observation lies in measuring the Atkinson, KB. and Newton, I., 1968. Photogram­ height ofthe theodolite trunnion axis above a metry. Chapter 6 in Photography for the scien­ levelled ground mark (Atkinson and Proctor, tist (Ed. C.E. Engel). Academic Press, London 1970). The Same criticism is valid if camera and New York. 632 pages: see 275-276. co-ordinates and exterior orientation data are Atkinson, KB. and Proctor, D.W., 1970. A photo­ being detennined, as Erlandson and Veress grammetric survey ofbuilding deformation over (1974) have pointed out. They also call for a a long period: Huby's Tower, Fountains Abbey. manufacturer's calibration, to the same accu­ Bulletin de La Societe Frane

trifugal testing ofsoil models. Photogrammetric olympischen Dacher Munchen. Ibid., 39(4): Record, 7(41): 538-554. 147-156. Erlandson, J.P. and Veress, S.A., 1974. Contempo­ Linkwitz, K., and Preuss, H. D., 1974. "Ober die rary problems in terrestrial photogrammetry. photogrammetrischrechnerische Ermittlungdes Photogrammetric Engineering, 40(9): 1079­ Membranzuschnitts fur das Schwimmhallen­ 1085. provisorium in Munchen. Ibid., 42(5): 171-174. Farrand, R, 1965. Photogrammetry applied to pipe Moore, J.F.A., 1973. The photogrammetric mea­ systems of chemical plant. Photogrammetric surement of constructional displacements of a Record, 5(26): 100-112. rockfill dam. Photogrammetric Record, 7(42): Gates, J.W.C., 1975. Position and displacement 628-648. measurement by holography and related tech­ Newton, I., 1974. Dimensional quality control of niques. Ibid., 8(46): 389-407. large ship structures by photogrammetry. Ibid., Grun, A. and Stephani, M., 1974. Photogrammet­ 8(44): 139-153. rische Vermessung der Brennkammer des Atom­ Nooshin, H. and Butterworth, J.W., 1974. Experi­ reaktors Garching/Munchen. Bildmessung und mental study ofa prestressed cable roof. Paper Luftbildwesen, 42(1): 12-18. given at the International Conference on Ten­ Happold, K and Liddell, W.I., 1975. Timberlattice sion Roof Structures, London. 17 pages. rooffor the Mannheim Bundesgartenschau. The Proctor, D.W., 1973. Where stands photogram­ Structural Engineer, 53(3): 99-135. metry today? Photogmmmetric Record, 7(42): Harding, J.E., 1974. The behaviour of high 706-711. strength friction grip boltedjoints in box girder Proctor, D.W. and Atkinson, K.B., 1972. Experi­ construction. Imperial College, London. CES­ mental photogrammetric wriggle survey in the LIC Report BC 2. x + 54 pages. 184 figs. Second Mersey Tunnel. Tunnels and Tunnel­ Karara, H.M. and Herron, RE., (Eds.), 1974. Bio­ ling, 4(2): 115-118 and 145. stereometries '74. American Society of Photo­ Ross-Brown, D.M. and Atkinson, K.B., 1972. Ter­ grammetry, Falls Church, Virginia. 640 pages. restrial photogrammetry in open-pits: 1. Kempson, L.L. and Heath, W., 1972. An appraisal Tmnsactions/SectionA ofthe Institution 0 fMin­ ofphotogrammetric methods for landslide sur­ ing and Metallurgy, 81: 205-213. veys in Colombia. Transport and Road Research Ross-Brown, D.M., Wickens, E.H. and Markland, Laboratory Technical Note TN 735. 26 pages. J.T., 1973. Terrestrial photogrammetry in open­ Lacmann, 0., 1950. Die Photogrammetrie in ihrer pits: 2. Ibid., 82: 115-130. Anwendung auf nicht-topographischen Ge­ Schmid, W., 1973. Die Photogrammetrie im was­ bieten. S. Hirzel Verlag, Leipzig. 220 pages. serbaulichen Versuchswesen. Vermessung, Linkwitz, K., 1967. Ein kontinuierliches digitales Photogrammetrie, Kulturtechnik, 71(4): 132­ Modell, dargestellt am Beispiel des Deutschen 137. PavilIons Montreal. Bildmessung und Luftbild­ Wickens, KH. and Barton, N.R, 1971. The applica­ wesen, 35(3): 95-100. tion of photogrammetry to the stability of exca­ Linkwitz, K. and Preuss, H.D., 1971. Die photo­ vated rock slopes. Photogrammetric Record, grammetrische Vermessung der Modelle der 7(37): 46-54.

ASP Needs Old Magazines

Because of an unexpected demand for journals and student requests, the supply of some back issues of PHOTOGRAMMETRIC ENGINEERING has been depleted. Consequently, until further notice, National Headquarters will pay to the Regions-or to individual members-$l.oo for each usable copy of the following issues sent to Headquarters, 105 N. Virginia Ave., Falls Church, Va. 22046:

Year Numbers May 1973-Vol. XXXIX No.5 January 1974-Vol. XL No.1 February 1975-Vol. XLI No.2 March 1975-Vol. XLI No.3