The Digital Terrain Model-

Theory & Application *

C. L. MILLER, Asst. Prof. of Surveying
Director, Photogrammetry Lab.
AND
R. A. LAFLAMME, Research Asst.
Dept. of Civil & Sanitary Engineering,
Massachusetts Inst. of Technology
Cambridge, Mass.

ABSTRACT: In order to realize the full potential value of an integrated

system of photogrammetry and electronic computers as applied to en-
gineering problems, the concept of a digital terrain model has been de-
veloped oy the Photogrammetry Laboratory at the Nlassachusetts Institute
of Technology. A band or area of terrain is represented in numerical or
digital form from data taken from a contour map, or directly from the
stereo plotter and stored on computer input material. The stored digital
terrain model may be used to obtain numerical solutions to many types
of terrain analysis problems by processing through an electronic digital com-
puter according to programmed instructions. Such an approach enables
the engineer to numerically evaluate an unlimited number of possible
location, design, and other geometric solutions to the problem presented.

INTRODUCTION
electronic digital computer can be
T HE
applied to two major areas of applica-
tion important to the photogrammetric
engineer. The first is concerned with the
reduction of the raw data (the photograph)
to obtain the basic photogrammetric out-
put (the spatial location of points). The
computations associated with analytical
space resection and intersection, aero-
triangulation, transformations, and adjust-
ments would be examples of the first area
of computer application.
The second area of application concerns
the computations associated with the solu-
tion of engineering, scientific, and military
problems which involve the use of photo-
grammetric output data. The determina-
tion of highway earthwork, the determina-
tion of the change in the shape of a glacier,
and the supplying of guidance instructions
for a low level missile would be examples of a
computer utilizing photogrammetric data. C. L. MILLER

* Presented at the Society's 24th Annual Meeting, Hotel Shoreham, Washington, D. C.

March 27, 1958.
433
434 PHOTOGRAMMETRIC ENGINEERING

As a technical specialist, the photogram- quirement, a machine data form counter-

metric engineer is primarily concerned with part of the engineer's topographic map or
the first area of application. However, graphical cross section is needed.
since these two areas are closely related, Before the second requirement can be
one giving direction to the other, they stated, certain characteristics of the mod-
cannot be entirely separated. Stated more ern computer should be examined. As
generally, an entirely new solution to a everyone has heard, computers perform
problem often comes from an integration arithmetic operations at fantastic speeds.
of the technological advances made in However, few people, even engineers,
many related or unrelated fields; also, comprehend the significance of such com-
consideration of the over-all problem from puting speed on the impact the computers
raw data to final answer will often lead to should have and will have in the approach
more efficient solutions. to future problems. Such an impact is
During the last two years, the Photo- gradually being felt in the field of highway
grammetry Laboratory of the M.LT. Civil engineering where the initial efforts were
Engineering Department has been con- to have the electronic computer do the
ducting research on the subject of "new identical work formerly handled by man-
approaches to highway engineering ual methods. Now that the "power under
through the application of photogram- the hood" is being felt, radical changes in
metry, automation instrumentation, and approaches can be expected.
electronic computers." The program is One way to take advantage of the speed
sponsored by the Massachusetts Depart- of the electronic computer is to numerically
ment of Public Works in cooperation with evaluate perhaps a hundred solutions to a
the Bureau of Public Roads. The purpose problem where only one or two were con-
of this paper is to describe one develop- sidered previously. Another use for the
men t by the project-the theory and speed is to take advantage of more exact
application of the digital terrain model. and rigorous mathematical formulations of
Although it is primarily an advance in the the problem instead of being limited to the
use of the photogrammetric output, the simple approximations dictated by manual
implications to the practice of photo- methods. Such changes in approach are
grammetry are important. economically justified as long as cost re-
flects a correspondingly more economical
THE TERRAIN DATA PROBLEM
solution. For example, if a computer can
A convenient representation of the sur- determine a mile of earthwork in ten
face of the earth is a common requirement minutes at say $10 in computing cost, it
for many engineering, scientific, and mili- certainly makes sense to spend 16 hours or
tary problems. The use and nature of $1000 in computing time if $10,000 can be
terrain data storage and presentation saved in construction costs by a more eco-
forms such as the topographic map, the nomical location.
physical model, the three dimensional We now come to our second requirement
stereo model, and the profile and cross- for the efilcient storage of terrain data. To
section are all common examples. These take full advantage of the computer, our
are all essentially analog forms of terrain digital terrain data should permit the ef-
data and are designed primarily for human ficient consideration of a large number of
interpretation and utilization in mental possible solutions to the problem at hand
and manual processes. Obviously, many of and preferably more exact or correct
the mental and manual processes for which solutions.
the human utilizes such terrain data in The digital terrain model system dis-
obtaining numerical answers to problems cussed in the remainder of the paper is one
can be efficiently handled by the elec- possible approach to realizing more of the
tronic digital computer. potential value of photogrammetry and
A fundamental requirement for using a electronic computers in such fields as civil
computer efficiently is to have the terrain engineering.
data in a form which the machine under-
stands. In the case of the electronic digital THE DIGITAL TERRAIN MODEL
computer, this form is of course digital data
THE CONCEPT
on computer input material such as
punched cards, punched tape, or magnetic The digital terrain model (DTM) is
tape. Therefore, re-stating the first re- simply a statistical representation of the
 DIGITAL TERRAIN MODEL-THEORY AND APPLICATION 435

continuous surface of the ground by a number of possible sets of selected points.

large number of selected points with The only requirement the set must meet
known xyz coordinates in an arbitrary co- is that it be stored systematically to facil-
ordinate field. Storing the DTM data on itate recovery by the computer. This
computer input material makes it avail- could be accomplished by specifying that
able to the computer for an analysis of a the points be stored sequentially in order
wide variety of terrain problems, and also of increasing x (or y or z). Such a sequence
for the evaluation of an unlimited number might be used if a system of points such as
of independent solutions to each type of used by the plane tahle topographer was
problem. selected.
By requiring that the points be located
THE COORDINATE SYSTEM
along a system of parallel scan lines, a
The origin and direction of xy horizontal more practical sequence of data results
axes of the DT M coordinate system and (Figure 1). Such scan lines will normally be
the z datum may be selected at will with lines of constant x (y scan lines) or lines of
due regard to convenience and the re- constant y (x scan lines). The distance be-
quirements of the particular problem at tween or spacing of the scan lines may be
hand. This coordinate system is independ- variable or constant, and the frequency of
ent but should be related to an estab- the poi n ts along the scan lines may be
lished system such as state plane-coordi- variable or constant. \iVith constant spac-
nates and mean sea-level datum. By arbi- ing of the scan lines and constant fre-
trarily specifying the DTM coordinates of quency of points along the scan lines, a
two points of known state plane-coordi- sq uare or rectangular grid would resul t.
nates, the state plane-coordinates of all Such a system would permit (1) the high-
DT M points may easily be computed, est degree of automation in the data pro-
when required, using simple rotation and curement phase, (2) the simplest process-
translation of axes formulae. Normally, ing of the data by the computer.
the DTM datum will be mean sea-level or With a variable frequency of points
a parallel datum. For a given project, a along the scan lines, the selected points
number of different but related DTM co- might be on (a) equal increments of z or
ordinate systems might be used for differ- contour line crossings, (b) increments
ent parts of the project area. In general, corresponding to a constant product of yz
the engineer has rather complete freedom increments, (c) terrain control points such
and considerable flexibility in selecting the as high and low points and slope breaks.
most convenient DTM coordinate system. Here again, the engineer has complete
freedom in selecting the system of points
DIGITAL REPRESENTATION SYSTEMS
best suited to the problem at hand. His
The surface of the ground may be choice might consider the type of terrain,
represented, in the DTM, by anyone of a available data procurement equipment,

DATA BASELlliE "x _ .-'

./ / Ltl/I, £x, j~
FIG. 1. This illustrates the arrangement of the Baseline, Scan lines, and Data Points. Data
Points are shown at each contour line, but more or less points may be used as dictated by system
requirements.
436 PHOTOGRAMMETRIC ENGINEERING

and the nature and requirements of his ap- the scan line may be obtained. By setting
plication of the DTM. In any of the sys- the differentiated form equal to zero, the
tems, the density of the points for a given location of the high and low points may be
type of terrain will depend on the accuracy obtained. By integrating the polynomial,
requirements associated with the applica- the area under the curve may be obtained.
tion. Different applications will require The following equations result.
different accuracy levels or degrees of per- dz
missive terrain approximation. In some Slope -
dy
= A + B [ (y - Yl) + (y - yz) .]
engineering problems, a series of DT M's
with progressively higher densities of + C[(y - Yl)(Y - yz) + (y - y,)(y - y,)
points for smaller areas might be advisable + (y - yz) (y - Y3)]
corresponding, for example, to the various Areay"..-y, = (y, - yz) [Zl - AYI - By,yz - CY1YZY3]
stages in the location and design of a high-
way. ()"z _ yzZ)
+ ---- 2
[A - B(y, + yz)
THE MATHEMATICAL TERRAIN MODEL
+ C(YlYZ + y,y, + y,y,) ]
The DT M, like the topographic map, (y,,3 _ yz3)
uses a sample of data to represent the con- + --:3-- [B - C(YI + yz + Y3)]
tinuous surface of the ground. The DT M
uses a sample of the infinite number of + ~~- yz') [C]
surface points, and the topographic map 4
uses a sample of the infinite number of
Interpolation between the scan lines
contour lines of the surface. Just as the
may be accomplished in a similar manner
engineer must interpolate on the topo-
by evaluating the polynomials across the
graphic map, the computer will have to
scan lines and parallel to the scan lines
interpolate with the DTM. And in both
which pass through the desired point from
cases straight line interpolation will often
points previously interpolated on the
be quite satisfactory. polynomials along the scan lines. An alter-
Since with the high-speed electronic
nate approach, and one which shows prom-
computer, much more sophisticated inter-
ise, is to evaluate the equation of a series
polation with the DTM is quite practical,
of surfaces which will fit the DTM points.
it is proposed that the actual model uti-
Considerable work is being performed on
lized by the computer be a mathematical
this subject but it has not yet reached a
model of the surface generated with the
stage far enough advanced for reporting at
data furnished by the DTM. Instead of
this ti me.
connecting each successive pair of points
The mathematical terrain model can be
with a straight line, a third degree poly-
justified if the number of DTM points
nomial will be generated by the computer
necessary to represent an area of interest
(Figure 2). A scan line will then consist of a
can be greatly reduced. A single third de-
number of continuous curves, each good
gree polynomial might, for example,
for a specified range of y val ues. The represent a given profile as accurately as
equation for the section between yz and Y3
would 50 straight line interpolations.
is of the form: (Figure 2) The speed and efficiency of the
z = z, + A (y - Yl) + B (y - Yl) (y - yz) electronic computer permits one to think
+ C(y - Yl)(Y - yz) (y - Y3) in terms of representing the surface of a
project area by thousands or literally tens
where the constants are computed by the of thousands of mathematical equations.
following equations in terms of the yz's of
the four points bracketing the range of DTM DATA INSTRUMENTATION
usefulness of the individual polynomial:
The most important contribution of
(zz - Zl) (Z3 - Zl) - A ()13 - Yl) photogrammetry to the DTM concept is
A = -- B = ::....::---:.:--..,.-"'''--~
(yz - )'1) (Y3 - Yl) (Y3 - yz) the practicality of obtaining the tre·
mendous amount of terrain data neces-
~-~-A~-~-B~-~~-~
C = ..:-c.._2-,--~-:-:-,'--'----:-':--~ sary. It would be quite impractical to ob-
(y, - y,)(y< - yz)(y, - Y3) tain the coordinates of, for example,
By differentiating the equation of the 10,000 points in a one square mile area in
polynomial, the slope at any point along any other way. Using photogrammetry,
 DIGITAL TERRAIN MODEL-THEORY AND APPLICATION 437

REPRESENTATION OF TERRAIN PROFILE BY THIRD DEGREE POLYNOMIAL

z ' 1000 + 0.31 y - 0.9 X 16~2 + 0.5 X 16 y'

6

1040r----,r--r----r--,---r---,---,----,r--r--r--,---,

960

940

'no
900

0+00 2+00 12+00

FIG. 2. This illustrates how terrain which would be reprcscnted by many points, using straight
line interpolation, may be reprcsented by fewer points using a third degree polynomial.

the data procurement problem is reduced lated work on the procurement of digital
to one of obtaining the coordinates of data from maps to operate a numerically
many points from the stereo model in a controlled milling machine for the produc-
fast, efficient, and accurate manner. tion of physical terrain models.
One approach to this problem is to pre- A second manual approach to obtaining
pare the standard topographic map to the DTkf data would be to scale the data
photogrammetric methods. The DTM data directly from the stereoplotter manuscript.
are readily obtainable from the contour Any of the systems of data along scan
map by graphically plotting the scan lines, lines could be selected. \Vi th some plotters
manually scaling and recording the data, it is possible to plot continuous pro/iles
and manually punching the results. Such across the stereo model. This is being ac-
an approach has the advantage that a complished in the M.LT. Photogrammetry
minimum additional investment in instru- Laboratory by means of a Nistri coordina-
mentation is required-the price of a sim- tometer and coordinatograph unit, operat-
ple scale. For limited and special use of the ing on either the Kelsh or Balplex plotter.
DTM system, such an approach might be These plots may then be digitized manu-
economical. However, the totally manual ally or automatically with a scanning and
approach is of course the slowest and most output system.
costly in manpower expenses. It would not The /irst real step in automation of the
be practical for extensive use of the DTM data procurement is obtained by digitizing
concept. Two men, one scaling and one re- one or more of the x, y, and z scanning mo-
cording, will take about 200 points per tions of the stereoplotter. If a rigid grid
hour on the average. Hence, our example system is being used, it would only be nec-
of 10,000 points would take 50 hours at a essary to digitize the z axis, the x and y
cost of 100 manhours plus the cost of positions being furnished by a projected or
punching and verifying the results. plotted grid. If the x increments are con-
Semi-automatic scanning and output stant, and/or plotted, digitizing of the y
instrumentation for operation with a map and z axes would be sufficient. Complete
as the source of data is quite possible and flexibility and automation isonly achieved
is the subject of current research and de- by digi tizing all three axes.
velopment at M.LT. Several different ap- A basic requirement for a scanning sys-
proaches to the problem are being investi- tem is that the scanning motions and the
gated and will be reported on at a later measurements they represent be converted
date. ]. A. Stieber, of the U. S. Naval to a form which can be digitized. This may
Training Device Center, is doing some re- be accomplished by converting the linear
438 PHOTOGRAMMETRIC ENGINEERING

FIG.3. The Nistri Coordinatometer shown on the Balplex Plotter. This unit is used to
obtain the x, Y, and z coordinates of points. rt is also L1sed on the Kelsh Plotter.

motions to equivalent shaft rotations by a static type, meaning that the operator
such means as a lead screw, rack and must stop for each point, and press a
pinion, continuous wire or chain, or odom- button to accomplish the automatic read-
eter type wheel. There are a number of out and recording. Two such systems have
commercial components for converting been designed and are being developed in
shaft rotations to digital output. A system the M.LT. Photogrammetry Laboratory,
of measuring and digitizing based on the one using electronic counters (Figure 4)
use of a diffraction grating might also be and the other, high speed relays. Although
used. A scanning, measuring, and digitiz- there are several commercial readout sys-
ing system of unique design based on send- tems available which are adaptable to
ing sound waves down a wire, has been de- photogrammetry, the M.LT. systems are
veloped by the Physics Department at being built to obtain a number of special
M.LT. for photographic data reduction experimental features for research work.
and shows potential for stereoplotter ap- The second basic type of readout sys-
plication. As a matter of convenience, it is tem which might be used would permit
desirable to have the scanning axes parallel readout "on-the-f1y" meaning that the
to the DTM data coordinate axes in order operator could scan the model with a con-
to directly read DT M coordinates. There- stant or variable but continuous drive. The
fore, either the stereo model or the scan- coordinates of points at equal y, z, or yz
ning unit axes should be capable of being increments would automatically be read
rotated about a vertical axes. out and recorded. Although the initial cost
For experimental work, the M.LT. would be greater than for a static readout
Photogrammetry Laboratory is using the system, the increased speed would eco-
previously mentioned Nistri unit as a nomically justify such a system in many
three-dimensional scanning unit for double cases. The actual data recording unit can
projection plotters (Figure 3). The unit is take many commercial forms such as a
rotated in stereo model space to align the tape punch, card punch, electric type-
axes. wri ter, magnetic tape recorder, or combi-
An odometer type scanning unit for use nation of such units.
with maps or plotters is being built. The The final degree of au tomation would be
Nistri unit may also be used for map scan- achieved by substituting an automatic
ning. scanning system for the human operator.
Two basic types of readout systems may Vlie can also expect to see radical changes
be used. The slowest but simplest would be in the stereoplotter itself as more advan-
 DIGITAL TERRAIN MODEL-THEORY AND APPLICATION 439

FIG. 4. The console, in the center of the photograph, gives a visual readout of the coordinates
obtained from the Nistri Coordinatometer. The readout is by means of electronic counting tubes
which also drive the Bendix tape punch, shown in the lower right-hand corner of the photograph.

tage is taken of the technological develop- or the path of a microwave. A series of

ments in automatic instrumentation. mathematical equations must be written
From the above paragraphs, it can be seen to represent the geometry of such surfaces
that the photogrammetric engineer has in a three dimensional coordinate system
many possibilities for increasing the effi- which coincides with or is related to the
ciency and speed of data procurement in DTM coordinate system. When this has
conjunction with the DT M system. The been accomplished, a program is prepared
M.LT. research staff is concerned with the to give the computer a set of instructions
investigation of the over-all field of auto- to follow in relating the two surfaces and
matic instrumentation in photogrammetry computing the answers of interest.
with particular emphasis on highway en- Although each application of the DT M
gineering applications and is exploring will have unique requirements and special
many new possibilities in this direction. controls on the necessary computer pro-
Since many other private, institutional, grams, there are a number of problems
and governmental groups are working in which are common to many applications.
the same and related fields, many new de- For example, the coordinates of the inter-
velopments are anticipated. section of a given plane and the surface of
the model is fundamental to the solution of
ELECTRONIC COMPUTER OPERATIONS
many engineering problems. The inter-
With the DT M data or the mathemati- section of a vertical plane and the model is
cal terrain model data stored on computer a terrain profile; a horizontal plane and the
input material, they may be read into in- model, a terrain contour; and a sloping
ternal computer storage in blocks as re- plane and the model, perhaps the limits of
quired. Hence, the represented surface is a highway cut or fill. Such surfaces inter-
essentially available "on demand" at secting the model might also commonly be
electronic speeds for operations by the cylindrical or conical. Once the intersec-
computer. Most applications of the DT M tions of the two surfaces have been com-
are concerned with mathematically re- puted, a common problem is to determine
lating some spatial surface of interest to the enclosed areas or volume. Hence, it is
the DTM represented surface. The spatial possible to write general computer pro-
surface of interest may be that of a pro- grams which can be adapted easily to a
posed highway, the surface of a reservoir, wide range of applications.
440 PHOTOGRAMMETRIC ENGINEERING

Three basic electronic computer pro- to select the highway profile or grade line
grams have been written at M.LT. for to meet specified optimization conditions
applying the DTM to highway engineering. and according to controls on gradient,
The Phase I program gives a general curvature, sight distance, design practice,
horizontal alignment solution for the prob- and similar geometric control specifica-
lem of relating any alignment to the DTM tions. AI though the mathematical formu-
data. The input data are the coordinates lation and programming of such problems
and radius of curvature at each point of will be quite complex and costly, they are
intersection-the P.L-and an origin of entirely practical and economically justi-
centerline stationing. The input data can fied.
be referenced to a coordinate system dif-
ferent from that of the DT M data and MULTIPLE VARIABLE EVALUATION
usually will be stated plane coordi na tes. The
The usefulness of the DT M approach
output of the Phase I program is the D.T M
can be greatly extended by simultaneously
coordinates and centerline stationing of
considering additional variables related to
selected points along the alignment, usu-
the problem at hand. This can be accom-
ally at stated intervals such as every 50
plished by attaching classification infor-
feet. The program also serves as a general
mation and quantitative data describing
solution to the profile problem, furnishing
other variables at each point. For example,
a terrain profile defined by the intersection
the total data associated with each terrain
of a series of plane and cylindrical vertical
point in the M.I.T. experimental work and
surfaces ,with the model.
automatically recorded by the output in-
The Phase II program gives a solution
strumentation take the form:
to the vertical alignment problem and re-
lates the profile reference line of the high- pppp xxxxxx YYYYY ZZZZZ CCNN
way surface to the DTM data. The input
to Phase II consists of the station, eleva- comprised of four identification digits, six
tion, and length of parabolic curve associ- digits of the x coordinate, five of y, five of
ates with each P.I. The output is the refer- z, a classification and two digits of quanti-
ence elevations at the same points com- tative data. Six different formatsof the out-
puted in Phase I. put data are presently possible by means of
After the reference line for the proposed a six position switch for different IBM
highway surface has been completely card, Remington Rand card, and Bendix
fixed in three dimensional space and re- tape punch data formats.
lated to the DT M data, the Phase III pro- The classification and quantitative data
gram generates the highway surface, com- associated with a given point can take
putes intersections of the surface and the many forms and includes a number of dif-
model, and determines the enclosed areas ferent types of variables. For example, the
and volumes. classification might be the type of soil at
In the present program, the surface gen- the point, and the quantitative digits the
erated or the cross-section templet can be depth of overburden to bedrock. Such data
composed of a series of 20 planes, all of might be obtained by airphoto analysis
which are variable from job to job, and and geophysical methods. The computer
six of which are variables selected by the program using such DT M data would com-
computer. Although these programs have pute the volumes of each type of material
immediate practical application to a vari- included in the construction requirements.
ety of problems, they represent only the Alternately, the quantitative data might
crude beginning of the programs which will represent the unit right-of-way cost and
ultimately be used with the DTM. the computer would be expected to deter-
The programs described above require mine the relative land acquisition cost for
that the engineer specify to a great extent each alignment evaluated. The solution of
the shape and location of his surface of in- the ul timate forms of such problems will be
terest. In many cases it will be possible based on the digital cost model previously
and desirable for the computer to assist in proposed by the senior author. In solving
selecting the location of the engineering the problems, the computer will evaluate
surface, according to limits and controls the most economical solution considering
set by the engineer. An example would be a all cost and benefit variables. Admittedly,
problem in which the computer is required such problems become quite complex and
 DIGITAL TERRAIN MODEL-THEORY AND APPLICATION 441

have immediate practical restrictions but RESUME OF ApPLICATIONS

certainly should be the subject of active A number of application examples have
research.
already been mentioned to illustrate char-
acteristics of the DTM approach. A re-
PRESENTATION AND UTILIZATION sume of these and several other civil en-
OF THE OUTPUT DATA gineering areas of application follows.
The output of the electronic computer a. Location, design, and quantity anal-
will be a digital or numerical form of data. ysis for highways, railroads, canals,
Quite often the result of interest will be in levees, dams, dredging, airports,
the form of a numerical answer and no fur- building develop men ts, etc.
ther transformation of the data will be b. Quantity determinations for borrow
necessary. However, in many cases the pits, quarries, open pit mines, coal
engineer will require an analog form of and ore piles, and other types of man-
data presentation for human study. A made and natural cuts and embank-
graphical plot will be the most commonly ments.
used analog form. c. Surface change studies related to
Continuous line plotters are now avail- settlement, erosion, silting, etc.
able for graphically plotting the results of d. Clearance studies for airport ap-
DTM problems. The potential applica- proach zones, microwave systems,
tions of the DTM will call for a family of radar and missile installations, etc.
such plotters. For example, the output of e. Terrain analysis problems associated
the mathematical terrain model approach with reservoirs, drainage problems,
would call for a system which would per- transmission lines of all types, etc.
mit plotting of the continuous third or In essence it may be said that whenever
higher degree polynomial equations de- two surfaces of interest are to be related
fining surface profiles. and computations are required, the DTM
Note that the computer analysis of the offers a possible approach. Photogram-
mathematical terrain model could furnish metry will usually offer the only practical
output data for the plotting of contour approach to the problem of obtaining the
lines either in terms of the coordinates of a required data, and the electronic com-
large number of points on each contour or puter makes it possible to consider com-
in terms of a series of equations defining puting problems which would require an
each contour line. Such output could be army of men with desk calculators.
automatically plotted to obtain contour
maps of the DT M area.
ACKNOWLEDGMENTS
An electronic computer program has
been written by the M.LT. Computation The research described in this paper is
Center for plotting and recording a contour sponsored by the Massachusetts Depart-
map of a matrix of digital values with the ment of Public Works and is directed by
Type 740 Cathode Ray tube output re- Anthony N. DiNatale, Commissioner of
corder of the IBM type 704 computer. Public Works; E. J. McCarthy, Chief
This program is being extensively used at Engineer; Patrick F. Cox, Deputy Chief
M.LT. in scientific data reduction and Engineer; Joseph O'Neil, Supervisor of Re-
illustrates a number of possibilities in the search; and Charles Whitcomb, Location
engineering field. In addition to two di- and Surveys Engineer. The program is
mensional continuous plots, the output conducted in cooperation with the U. S.
data can be used for controlling three-di- Bureau of Public Roads under the super-
mensional cutting machines for carving sision of Charles Hall, District Engineer,
physical models of the DTM area. and John Swanson, Regional Engineer.
In all of the previous examples, the Encouragement and technical advice from
equations defining the engineering sur- the office of H. A. Radzikowski, Chief, Di-
faces of interest will furnish the data for vision of Development, Bureau of Public
the plotting or carving of such surfaces Roads, is gratefully acknowledged. The
along with the DT M surface. There are following firms have actively assisted the
many other possible forms and uses for the program in arranging for instrumenta-
output of the DTM system in solving en- tion facilities: I BM Corporation, Bendix
gineering problems. Computer Division, Remington Rand,
442 PHOTOGRAMMETRIC ENGINEERING

a.M.L Corporation of America, Bausch 9. Miller, C. L. and R. D. Schurz, "Educa-

and Lomb. Finally, the authors pay trib- tional and Research Activities in Photo-
ute to their colleagues on the project staff grammetry at M.LT." PHOTOGRAMMETRIC
including D. R. Schur;,:, E. P. Gladding, ENGINEERING, March 1958.
10. Miller, C. L., "Impact of the Expanded
and T. H. Kaalstad. Highway Program on Photogrammetry,"
PHOTOGRAMMETRIC ENGINEERING, Decem-
REFERENCES
ber 1956.
Addjtional material on the subject paper will II. Roberts, P.O., "Using New Methods in
be found in the following publications of the Highway Location," PHOTOGRAMMETRIC
M.l.T. Photogrammetry Laboratory. ENGINEERING, June 1957.
1. No. 107-Preliminary Report, Integrated
Aerial Photogrammetric and Electronic Information on developments by other
groups which are related to the subject matter
Computer System
include:
2. No. 109-Digital Readout Systems and
Components for Photogrammetric Instru- 12. ROSENBERG, P., "I nformation Theory and
men ta tion Electronic Photogrammetry," PHOTO-
3. No. 110-A Study of Directional Scanning GRAMMETRIC ENGINEERING, September
Techniques 1955.
4. No. 111-Digital Terrain Model Approach 13. Esten, R. D., "Automatic Contouring,"
to Highway Earthwork Analysis PHOTOGRAMMETRIC ENGINEERING, March
5. No. 112-An Automatic Digital Output 1957.
System for Double Projection Stereoplot- 14. Spooner, Dossi, and Misulia, "Let's Go
tel's Over the Hill-Potential Benefits of Profile
6. No. 113-The Skew System for Highway Scanning the Stereo-Model," PHOTOG RAM-
Earthwork Analysis METRIC ENGINEERING, December 1957.
7. o. 114-Electronic Computer Program- IS. Schwidefsky, K., "New Aids for umerical
ming of the Skew System Photogrammetry," Photogrammetria, Vol.
8. TO. liS-Earthwork Data Procurement by XIV, No. I, 1957-58.
Photogrammetric Methods 16. Stieber, j. A., "The Master Terrain Model
Previous papers by the project staff which System," paper presented at 1957 Joint
discuss some of the concepts presented in th is Eastern Computer Conference. To be pub-
paper include: lished in proceedings of conference.

by twenty times, and was manufacturing optica

glass at the rate of 40,000 pounds per month
At the close of ''''orld War I, B&L had pro-
duced 450,000 pounds, or 65% of all the optical
glass used by the military forces. Twenty years
later with the outbreak of World War II,
Bausch & Lomb was again ready for military
production. During the war, the Glass Plant
B&L GLASS PLANT CELEBRATES 40TH ANNI- manufactured high quality optical glass at
VERSARY about 1200% of the pre-war production.
This is an anniversary year for the Bausch Since 1918 B&L has continued to produce its
& Lomb Optical Company of Rochester, New own optical glass, and at the present time is the
York. I t marks 40 years of the successf ul manu- only manufacturer in the Western Hemisphere
facture of high grade optical glass for scientific, producing optical elements from sand to fin-
professional and industrial purposes. The story ished product. The requirements for the optical
of optical glass in the United States begins at products made by Bausch & Lomb, including
Bausch & Lomb early in the century, when both ophthalmic goods and scientific instru-
William Bausch began experiments by mixing ments, numbers more than 120 different types
small batches of ingredients in an oil fired fur- of optical glass.
nace which had been constructed next to the It has manufactured its own optical glass
molding plant. for years. Because the research experts have
Beginning in 1917, the manufactur of optical selected only the best raw ingredients and chem-
glass in this country has been closely allied with icals and the engineers and craftsmen control
national defense. Until 1917 all of the optical every operation in the processing, the Bausch &
glass used in the United States was imported Lomb Optical Company has earned a reputa-
from Germany. By the end of 1917, the Bausch tion for the manufacture of optical glass of the
& Lomb Glass Plant had multiplied its volume very highest quality.