404 'l'HOTOGRAMMETRIC ENGINEERING

tion, they afford a convenient, inexpensivesource of information concerning mapaccuracy. Highway engineers should takethe lead in the development of data on mapaccuracy as a means of increasing the use­fulness of photogrammetric mapping.

In concl usion, the au thor acknowledgesthe contributions to this study made byRex H. Fulton, Senior Highway Engineerand George P. Katibah, SupervisingPhotogrammetrist, of the California Di­vision of Highways. Over a period of sev­eral years Mr. Fulton has been responsiblefor many of the improvements in Californiamapping specifications as well as for usingthe stereoplotter in map checking. Mr.Katibah has supplied technical knowledgeand supervision necessary for mapaccuracy investigation and plotter check­ing.

REFERENCES1. Sharp, John V. PHOTOGRAMMETRIC ENGI­

NEERING, June 1951, Page 358.2. Thompson, Morris M. and Davey, C. H.

Sttrveying and Mapping, January-March1953, Page 40.

3. Thompson, Morris M. Surveying and Map­ping, April-June 1956, Page 164.

4. A.S.T.M. Manual on Quality Control ofMaterials, Special Publication 15-C.

5. Harman, William E., Jr. MANUAL OFPHOTOGRAMMETRY, Page 262.

6. Altenhofen, Robert E. PHOTOGRAMMETRICENGINEERING, June 1951, Page 370.

7. Struck, Luis. PHOTOGRAMMETRIC ENGI­NEERING, March 1952, Page 84.

8. Trorey, Lyle G. PHOTOGRAMMETRIC ENGI­NEERING, December 1951, Page 831.

9. Pennington, John T. MANUAL OF PHOTO­GRAMMETRY, Page 639.

10. Pryor, William T. Highway Resea1'Ch BoardBulletin 157, Page 23.

Photogrammetry and Road Locationin the U. S. Forest Service*

C. E. CARNAHAN,

Division of Engineering,U. S. Forest Service, Ogden, Utah

ABSTRACT: The author describes the photogrammetric road location anddesign procedures developed an.d used by the United States Forest Service.The processes cover all phases of road sttrvey operations from recon­naissance through final design. A brief discussion is given to the testsconductr;d by the Forest Service to determine what first order photogram­metric equipment would best serve their needs. The work evolving aroundthe Stereo planigraph C-8 is broken down into seven distinct operations toobtain the necessary information at the required accuracy for the successfuldesign of a road construction project.

I N RESPONSE to the mounting demand forational Forest timber products, the

Forest Service of the V. S. Department ofAgriculture is engaged in a steadily ex­panding construction program for forest

development roads. Ten years ago, theForest Service was building less than J ,000miles of road annually. Recent years havewitnessed a considerable increase in roadconstruction financed from appropriations

* Presented at 2nd Annual Surveying and Mapping Conference, University of Utah, Jan. 17,1958.

ROAD LOCATION IN THE U. S. FOREST SERVICE 405

and an even greater increase in the mileageof roads constructed yearly by purchasersof National Forest timber. Roads in bothcategories are located by Forest Serviceengineers. 3,580 miles of National Forestdevelopment roads were completed in thefiscal year ending June 30, 1957. Projectsaggregating 3,500 miles or more will becommenced during the current fiscalyear. Meanwhile, surveys are being madeand plans prepared for 3,700 miles of con­struction and major reconstruction sched­uled for the fiscal year commencing nextJuly 1.

This accelerating activity together withthe acute shortage of civil engineers madenecessary searching for a new approach toour road survey problems. Early in 1956,after careful consideration of several alter­natives, the Regional Forester for the Inter­mountain Region approved the establish­ment of a Photogrammetric Section in theDivision of Engineering for that region.Its function is to locate and design roadsthrough the use of aerial photographs andfirst-order photogrammetric equipment.

The heart of the Photogrammetric Sec­tion is the Stereoplanigraph C-8, manu­factured by Zeiss Aerotopograph of M u­nich, Germany. It was installed at Ogden inFebruary 1957. Two technicians fromGermany accompanied the equipment toinstall, adjust and calibrate it, and to con­duct a training program for section per­sonnel.

The seq uence of operations in vol ved inphotogram metric road location varies wi thand is dependent upon the working dataavailable. For illustrative purposes it isassumed that a project is located in anarea for which standard U. S. GeologicalSurvey topographic maps and 1: 20,000scale resource photography are available.The steps followed in locating, designingand determining costs of the road to bebuilt are as follows:

Action is taken to obtain an exact-scalecopy of the original topographic manu­script on which is shown all existing fieldcontrol, both horizontal and vertical, there­by assuring that all basic data are madeknown to the Section. Using these topo­graphic manuscripts, preliminary loca­tions are traced, care being exercised tokeep within the grade and curvature limita­tions of the standard of road design desired.At the conclusion of such work it may befound that five or six possible routes have

been delineated for consideration.By utilizing the available 1: 20,000

photography and third-order photo­grammetric plotting instruments, thesepreliminary lines are transferred very care­fully to the photographs, thus producinga "bird's-eye" view of the several possibleroutes for the proposed road. The photo­grammetrist has now made available to thedesign engineer a delineation of preliminarylocations from which a vast amount ofinformation can be obtained. A report isprepared describing each line in detail withregard to grade, per cent of sidehill slopesneeded to hold the road, type and densi tyof cover, and all unusual topographicfeatures which might be encountered.

At this point the soils analyst begins tostudy the survey material. The Inter­mountain Region is exceptionally fortu­nate in that the staff has a soils specialistwho has made an intensive study of photointerpretation as it pertains to soils, andwho is capable of determining probablesoil structure from the appearance oftopographic features and of cover typeswith correlation with soil classificationmaps. He prepares a report on each tracedline in which he points out the advantagesand disadvantages expected to be encoun­tered. His report also indicates areas inwhich heavy rock work is apparent, thosein which unstable soils are likely to be en­countered, the intensity of drainage dis­section, probable subsoil conditions, etc.All of these can be predicted to a certaindegree of accuracy. The importance ofsuch preliminary data cannot be over­stated.

If timber survey maps are available, theroutes are also d~lineated on them, thusfurther completing the picture for con­sideration by personnel involved in TimberManagement. It quickly tells the story ofthe type, age and density of timber whichwill be reached by adding the road to theTransportation System.

Following the preparation of maps,photos and analyses, the material is su b­mitted to the Forest Supervisor underwhose jurisdiction the proposed projectfalls. He, with his Forest Engineer andother members of his staff, select the mostappropriate route. By this means there is afeeling of certainty that the road will belocated where it will best serve multipleuse of the forest. Without question this isthe most important service that this part

406 PHOTOGRAMMETRIC ENGINEERING

of the Engineering Division can provide.It is mandatory at this point in the

operations for the Forest Engineer orother competent person to traverse theselected route by foot or horseback to gainfirst-hand information, to verify the gradesand interpretations, and to set reflectorsignals for use as reference poin ts whichwill register on the large-scale photographywhich must still be obtained for surveypurposes. This step is called "verificationof route reconnaissance," since, up to thistime, route study has been conducted with­out putting a survey crew in the field.

The reflector signals set during groundreconnaissance serve an extremely impor­tant function in photogrammetric surveys.First, as a positive and unquestionable tiebetween the paper and ground locations.The signals are normally located at con­venient intervals along the route wherethey will serve as reference points andalso on all section, quarter section, H.E.S.and other land monuments which deter­mine the location of private land holdings.Signals are set on any field establishedelevation, traverse PI or triangulationstation which may be encountered alongthe route. By marking these points priorto acquisition of the survey photography,the accuracy of all operations is consider­ably increased, and the data are more easilyutilized.

Before proceeding with a discussion ofthe actual processes involved in determin­ing the road location and design, there willbe briefly considered the background thatmotivated the Forest Service to invest inhigh-order photogrammetric equipment.

The original efforts of the Forest Servicewere developed around Kelsh plotter-typeequipment since this was available andthe personnel were well versed in its opera­tion and capabilities. Several projects wereundertaken-each with a variation in theprocedural approach and with varyingdegrees of success. The procedure wasessentially one of drawing a large-scaletopographic strip map using the the Kelshplotter and designing the road on thisbase. Regardless of the technique em­ployed, there were continual limitationscharacteristic of the machine, such as thegraphical analysis or radial plot which isrequired to tie stereo-models together andto ground-control, its inability to extendcostly field-control, and the human tend­ency to commit errors.

In 1955, Zeiss-Aerotopograph of Munich,Germany, offered the Forest Service theuse of its RMK 21/18 camera and a C-8Stereoplanigraph for test purposes, to de­termine the capability of this equipment toperform precise photogrammetric surveys.The Forest Service accepted the offer andselected an area on the Tahoe NationalForest in California for use as a provingground. The tests conducted there were tostudy advanced photogrammetry as ap­plied to cadastral surveys and engineeri ngsurveys of bridge sites, dam sites, adminis­trative building sites, and road location.The Tahoe Project was photographed atvarying scales with the RMK 21/18camera during the summer of 1955. Field­control operations were executed in thesame season by U. S. Forest Serviceengineers. Subsequently, during the win­ter, Mr. J. E. King went to Munich, Ger­many to present the problems and to ob­serve the work in progress. At this pointacknowledgement is made of the excep­tionally fine work and cooperation by boththe Forest Service personnel in Californiaand the Zeiss-Aerotopograph organiza­tion; largely because of their combinedefforts the recent advances were madepossible.

For the benefit of those interested in thecadastral survey application of these tests,attention is invited to Mr. King's report,"Photogrammetry in Cadastral Survey­ing" which was given before the combinedmeetings of the American Society ofPhotogrammetry and the American Con­gress of Surveying and Mapping in Den­ver, Colorado, October 2, 1956.1

An analysis of the data obtained throughthese tests (with emphasis on accuracy)formed the basis for purchasing one RMK21/18 camera and two StereoplanigraphC-8 instruments. One instrument is locatedat the Eastern Photogrammetric Section,Alexandria, Viginia, and the other at theIntermountain Regional headquarters ofthe U. S. Forest Service in Ogden, Utah.The latter instrument is devoted entirely toroad location operations, and is the equip­ment which is the subject of today's dis­cussIOn.

In considering the subject of equipmentand why it was selected for the Region'sroad location operations, first to be con-

1 Published in PHOTOGRAMMETRIC ENGINEER­ING, Vol. XXIII, No.3.

ROAD LOCATION IN THE U. S. FOREST SERVICE 407

sidered is the aerial camera, the RMK21/18; this has a focal-length of 8i" anda 7X 7" format. The accuracy tests towhich this camera has been subjected ratethe resolving power of the lens at roughlytwice that normally found in 8t" lens; itsmeasured distortion is 3 microns as com­pared to 25 microns frequently encoun­tered in conventional cameras. It has anew, extremely fast shutter which willoperate up to 1/1,000 of a second, thusmaking it capable of practically stoppingimage motion at the instant of exposure.The intervalometer which actuates theshutter has automatic features incor­porated within it which make virtually im­possible the photographer losing the de­sired overlap in line of flight. All of thesefeatures add up in the long chain of itemsaffecting accuracy. But of all these preci­sion characteristics, the most important isthat the camera is matched to the Stereo­planigraph C-8 and becomes a companioninstrument in every sense of the word.

Perhaps the question most frequentlyasked is, "what makes a first-order instru­ment so much better than those of lowerorder?" This will be answered briefly:

1. The resolving power of the optics in­corporated in the instrument.

2. The accuracy of its tracks, ball­bearings, spindles, and motion trains.

3. The accuracy of the camera lens­distortion compensating device.

4. The magnification ratio of its viewingsystem.

Some of you may think of the accuracyof photogrammetric equipment in terms ofC-factor, which is defined as the ratio ofthe photographic flight height to thesmallest contour interval which can bedrawn. Those in the Forest Service in­volved in applying photogrammetry toroad location do not use this concept; in­stead the thought is in terms of point meas­urement, and it is recognized that manythings enter into the picture when deter­mining C-factors which do not apply to theSection's work.

The drawing of contours causes a loss inaccuracy since the operator does not havethe opportunity to make repeated meas­urements. Further, the ability of operatorsto move a floating mark over the surfaceof the ground varies radically between indi­viduals.

The C-factor was originally conceiven

for use as a tool in planning a photogram­metric topographic mapping job, andvalues were assigned to various plotting in­struments for this purpose. However, it isfel t that it fails to reprcsen t the trueaccuracy of the instrument.

The preference is the determination ofthe root mean-square error of the measuredvalues made on precision grid plates. Thesemu'st be oriented in the machine and multi­ple readings made on each grid intersectionby the operator so as to minimize thehuman element of error. The recordedvalues of these measurements are thenanalyzed for the accuracy obtained. Usingthis means that by checking the C-8Stereoplanigraph in Ogden on a projectiondistance of 418 mm., its accuracy wasdetermined to be 1; 17,800 of this dis­tance. This means that measurements canbe made to 0.0235 mm. on 8t' photog­raphy; this val ue in terms of feet-on-the­ground will vary with the scale of thephotography. As an example, assume it is"decided to obtain photography for a surveyjob with the RMK 21/18 camera flying atan elevation of 3,500 feet. By simple ratioit is determined that under ideal conditionsthe machine is capable of measuring to 0.2ft. in vertical displacement. This figure,as has been stated, applies strictly to themachine capabilities.

The actual field tests of the equipmentreflect the accumulated errors of all theequipment, materials and processes utilizedto .produce the diapositives being used.The test therefore becomes realistic andtrue and not one of comparison.

The photography obtained on the TahoeNational Forest Project, for use in theMunich test for profiling and cross sec­tioning, was 1; 2,400 (200 ft. per inch). Thecontrol was established photogrammetri­cally from 1; 27,000 scale photography, alsoflown at this time.

The area of the test is described by thefield crew as follows: "Topographic fea­tures consist of ridges, gulches, andslopes of varying steepness with densecoverage of manzanita 8 to 12 feet inheight. There are many madrone, oak, pineand fir trees scattered throughout thearea. Poison oak, bitter brush and severeerosion in some of the gulches combine tomake this a di fficul t test for the plotter."A ground check traverse was run over 31centerline stations and 6 cross sectionss~lected at random from the centerline and

408 PHOTOGRAMMETRIC ENGINEERING

cross-sections established and measured bythe plotter. The extent of the field checkwas reduced due to the heavy line clearingwhich was found necessary. Thirteen man­days of work were required to check thislimited area, which gives an indication ofthe cover density. The average error of the100 points tested was ±0.51 ft.; however,50 per cent of the points that were checkedshowed an error of 0.3 ft. or less, and 65per cent were with 0.5 ft.

A recent test made on the Ogden ma­chine in which 1: 15,000 scale photographswere used resulted in an average error of0.4 ft. for 15 points tested. Interferingground cover in this instance was sparse;this eliminated the problems presented bythe dense brush cover encountered in theTahoe test.

Accuracy, although of prime importancein the Forest Service concept of photo­gram metric surveying, is not the only con­sideration. It is well known that the areasin which the Forest Service has to operateare remote, rugged and often difficult toreach. All field surveys and operations aredifficult and costly. With the Stereo­planigraph, control can be extendedthrough inaccessible areas by a processknown as Stereotriangulation or "Bridg­ing," thus making possible establishingfield control where the least cost and effortare required. This is a point of major im­portance.

Still another extremely valuable char­acteristic of the Stereoplanigraph is theplane-coordinate grid system incorporatedwithin it, and an attendant tabulating de­vice which records the x, y and z coordi­nates at any point measured with the float­ing mark. The existence of these featurespoint immediately to the possibilities ofthe mathematical treatment of data and itstransformation to the State Plane Coordi­nate System for ground application.

The techniques used to arrive at finaldesign may vary to a certain extent fromproject to project depending upon thebasic information available; however, theyhave now reached the point where there isconfidence in proceeding in a positive man­ner.

The Section's concept of photographicneeds for road location operation is iden­tical to that of the Ohio Department ofHighways. Two scales of photography areused-one for reconnaissance and controlpurposes and the other for design. The

small-scale or control-photography is nor­mally about 1: 20,000 scale. Design photog­raphy may vary from 1: 2,500 to 1: 10,000depending upon the class of road desired,ruggedness of terrain, and density of cover.The very nature of ational Forest areas,rugged terrain, relative low land values,inaccessibility, etc. permit the use ofsmaller scales of ph0tography; and con­currently a lower order of accuracy thanwould be required in heavily populatedand industrialized areas.

Assuming that a given project has beenauthorized for survey. and that the re­connaissance has been completed, and theroute selected, the procedure to be fol­lowed in actually completing a survey willnow be discussed.

First-Field control (both horizontaland vertical) is established to meet therequirements of the Stereoplanigraph andthe small-scale photography. Full use ismade of all existing 3rd-order or bettercontrol which has been established andpublished by the various responsiblegovernment agencies. This may necessitatea Ii ttle search and a bi t of correspondence;however, the effort is usually well re­warded. It may be found that level linesand triangulation stations already estab­lished will serve the needs and that actuallythese already have been photo identified,thus entirely eliminating field operations.Usually, however, some additional fieldwork will be required, and this is accom­plished by using the T-2 theodolite forhorizontal and vertical-angle measurementand self-leveling level for vertical control.

At this point acknowledgement is madeof the exceptionally fine cooperation re­ceived from the Geological Survey throughmaking its field control data available. Thecooperation between these two agenciesis an excellent example for all Governmentagencies.

Second-If the reflector signals previ­ously mentioned were not set during veri­fication of the reconnaissance, they mustnow be established. The importance of thisoperation cannot be overestimated; uponit much of the accuracy of the survey isdependent.

Third-Both control photography andlarge-scale aerial photographs for designuse are obtained with the RMK 21/18camera. The ratio of these scales sho,uld notexceed 3 to 1 due to differences in image,size and conClirrent reading acuity. The

ROAD LOCATION IN THE U. S. FOREST SERVrCE 409

camera is owned by the Forest Service.National Forest aerial photography isordinarily contracted but owing to dif­ficulty attending advance scheduling ofroad survey photography, only the planeand pilot are contracted. The photographeris a Forest Service employee who is wellinformed on the requirements of the photo­gram metric equipment and in addition isacquainted with the forest areas in whichthe work is being done. His all-aroundknowledge is exceptionally helpful.

Fourth-With the small-scale photog­raphy and the field-control which hasbeen obtained, the photogrammetrist us­ing the Stereoplanigraph establishes andextends supplemen tary con trol by theStereotriangulation or "Bridgi ng" tech­nique. Roughly, this may be described asthe precise orientation of the initial modelto established field control, to which isadded a succession of 3-dimension modelsto form a network of control similar inappearance to a field triangulation scheme.As this network is extended it interceptsother field established con troJ. Because themachine bridges along a Rat plane whichdoes not correspond with the earth's sur­face, and because of minute systematicerrors in even this precision equipment, itis reasonable to expect deviations from thetrue values of field established control at asystematic rate. It is apparent then thatbecause these deviations are systematic,this elimination or correction can be ac­complished mathematically. The differ­ences in the recorded and ground-co­ordinates form the basis for this mathe­matical adjustment. This adjustment ap­plies not only to the limited field-controlpoints but to all machine measured pointsas well. I t is obvious that this is a tre­mendous undertaking in mathematics.

At the ti me this paper is wri tten, theSection is cooperating with various com­putor companies, Zeiss Corporation andthe Coast & Geodetic Survey in the estab­lishment of a fully electronic computeradjustment. It is believed that its suc­cessful conclusion will surely be realizedquickly. In making this spacial bridge, inthe same operation, there is establishedsupplementary control for the large-scaledesign photography. This is merely a mat­ter of selection and measurement of pointscommon to the two scales of photography.

Fifth-A second bridge is made, thistime utilizing the large-scale (or design)

photography and the supple men tary con­trol established by the first "bridge." Thetechnique is identical to the first, and itspurpose is to further refine the machine­established control, minimizing any read­ing errors which may have been committed.

Sixth-Following the second adjust­ment the diapositives are re-oriented intheir corrected positions and work mayproceed wi th design and cross-sectioni ngand the establishment of coordinate valueson reference points which had been markedduring the verification of the reconnais­sance. I t should be pointed out that atthis point the work may be "farmed-out"to second-order plotting equipment fordrawing strip topography.

I t is believed that the characteristics ofthe C-8 can best be used in point measure­ment whether it be control extension,profiling, or cross sectioning. If contoursare to be drawn in any amount the workshould be delegated to less expensiveequipment. Whatever the innovations in­troduced at this point-the final step is toactually profile and cross-section the de­signed road with the Stereoplanigraph inorder that the most accurate measure­ments possible may be recorded for trav­erse and earth work q uan ti ty determi na­tions. In this way, the errors inherent incontour lines and the attendant errors ofinterpolating and scaling data from themare avoided. Currently attempt is beingmade to design a section of road wi thou tstrip topography; its possibilities causeenthusiasm.

Seventh-The final step is drawingplans and profiled and completing the cal­culations necessary prior to issuing con­struction con tracts. The possi bili ties ofusing electronic computing equipment toachieve final results from data furnishedvia the Stereoplanigraph was at the ou tsetexciting, but not considered particularlynecessary. As the work has progressedand methods have developed, this com­bination of Stereoplanigraph and elec­tronic computer has become more andmore important. Although the actualcomputation work has not yet been com­pleted by electronics-the plans are madeand the procedure established except forminor details. In a few short weeks thisphase of the operation should be wellestablished and in good operating order.

What is gained by turning to photo­grammetry for road location surveys? (1)

410 PHOTOGRAMMETRIC ENGINEERING

a location acceptable to the majority ofinterested people; (2) a twelve-monthsurvey season instead of the usual six; (3)a saving of engineering manpower forpreparing contract specifications and su­pervision of construction; (4) the surveydollar stretched over many more milesthan it previously covered; (5) already asaving of many miles of high-cost con­struction and avoidance of difficult areasby being able to "look ahead" throughphotogrammetry.

Although the equipment represents aconsiderable monetary investment, there isevery reason to believe that in two to fouryears it will pay for itself in actual cashsavings to the U. S. Forest Service. -Thisestimate is based on a double-shift capacityof 150 miles of road design per year at asavings of approximately $300.00 permile or $45,000 per year. Each operationis directed toward saving time and moneyand the elimination of interference fromadverse weather conditions, limited fieldseason and personnel shortages.

The efforts are only just beginning tomaterialize into production. The opera­tion has been shifted into second gear-so

to speak-with 16 hours of operation perday; there is hope of having seven roadsurveys completed within the next 6 or 8months. A standing invitation is given tovisit the laboratory to observe the work,discuss procedures and exchange a fewideas. Like topographic mapping, en­gineering surveys have taken to the airand tedious computations are being com­fortably consumed by the electronic com­puter.

The Forest Service engineers have ac­cepted these new tools and are enthusiasticabout their possibilities. It is hoped thatthey will help others to accept and usethese tools successfully.

The operations and procedures de­scribed in this paper are in use by theForest Service, U. S. Department ofAgriculture at the Region 4 headquartersof the Forest Service at Ogden, Utah. Dr.Richard E. McArdle is Chief of the ForestService and Mr. Floyd Iverson is Re­gional Forester for Region 4. Mr. ArvalAnderson is the Assistan t Regional Fores­ter in charge of the Division of Engineeringand the author is Chief of the Carto­graphic Section of that Division.

Mapping of the Ungava Peninsula*

s. G. GAMBLE, Chief Topographical Engineer,Dept. of ~Mines and Technical Surveys,

Ottawa, Canada

ABSTRACT: The mapping of the mineral wealthy Ungava Peninsula ofnorthern Quebec has provided the Topographical Survey of Canada toassess new survey and photogrammetrical techniques on a large practicaltest. The environment is described with particular emphasis on mappingconditions and the initial planning and surveys are outlined. In theUngava operation the application of electronic instruments was impres­sive and included the use of Shoran for geodetic control, radar altimetersfor vertical control and horizontal scale, tellurometers for secondary surveycontrol and for establishing scale on selected flights. Experience gainedfrom the Ungava project should reduce the cost of mapping in the northernareas of Canada.

T HE mapping of the Ungava Peninsulaof northern Quebec has proved to be

one of the most important mapping proj-

ects ever undertaken by the TopographicalSurvey of Canada. The reason is not thatthe potential mineral wealth of the Penin-

* By permission of the Director, Surveys and Mapping Branch, Department of Mines &Technical Surveys, Ottawa, Can., presented at 24th Annual Meeting of the Society, Hotel Shore­ham, Washington, D. C., March 26, 1958.