Photogrammetry ebook free download

de Gruyter Textbook
Karl Kraus
Photogrammetry



Karl Kraus

Photogrammetry
Geometry from Images
and Laser Scans
Second Edition

Translated by
Ian Harley
Stephen Kyle

w Walter de Gruyter
DE

G Berlin · New York


Author
ο. Univ.-Prof. Dipl.-Ing. Dr. techn. Karl Kraus
formerly
Institute of Photogrammetry and Remote Sensing
Vienna University of Technology
Vienna, Austria
Translators

Prof. Ian Harley
Dr. Stephen Kyle
University College London
London, Great Britain

This second English edition is a translation and revision of the seventh German
edition:
Kraus, Karl: Photogrammetrie, Band 1, Geometrische Informationen aus Photographien und Laserscanneraufnahmen. Walter de Gruyter, Berlin · New York, 2004
First English edition:
Kraus, Karl: Photogrammetry, Volume 1, Fundamentals and Standard Processes.
Dümmler, Köln, 2000.

© Printed on acid-free paper which falls within the guidelines
of the ANSI to ensure permanence and durability.

Bibliographic information published by the Deutsche

Nationalbibliothek

The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie;
detailed bibliographic data are available in the Internet at .

ISBN 978-3-11-019007-6
© Copyright 2007 by Walter de Gruyter GmbH & Co. KG, 10785 Berlin, Germany.
All rights reserved, including those of translation into foreign languages. No part of this book
may be reproduced in any form or by any means, electronic or mechanical, including photocopy,
recording, or any information storage and retrieval system, without permission in writing from
the publisher.
Printed in Germany.
Coverdesign: +malsy, kommunikation und gestaltung, Willich.

Printing and binding: Hubert & Co. GmbH & Co. KG, Göttingen.


Foreword to the second English edition
The first edition of Volume 1 of the series of textbooks "Photogrammetry" was published in German in 1982. It filled a large void and the second and third editions were
printed soon afterwards, in 1985 and 1990. The fourth edition was published in English
in 1992, translated by Peter Stewardson. The following three editions were published
in German in the years 1995, 1997, and 2003, making seven editions in all. The English
edition was re-printed in 2000.
Volume 1 was additionally translated into several languages, including Serbocroatian,
by Prof. Joksics, Technical University Belgrade; Norwegian, by Prof. Oefsti, University of Trondheim; Greek, by Dr. Vozikis and Prof. Georgopoulos, National Technical
University of Athens; Japanese, by Prof. Oshima and Mr. Horie, Hosei University;
Italian, by Prof. Dequal, Politecnico Torino; French, by Prof. Grussenmeyer and O.
Reis, Ecole Nationale Superieure des Arts et Industries de Strasbourg; Hungarian by
Prof. Detreköi, Dr. Melykuti, S. Mihäli, and P. Winkler, TU Budapest; Ukrainian by
S. Kusyk, Lvivska Politechnika; and Turkish, by Prof. Altan, Technical University of
Istanbul.
This second English edition is a translation of the seventh, German, edition by Dr. Ian
Harley, Professor Emeritus, and Dr. Stephen Kyle, both of University College London.
They not only translated the text, but they also made valuable contributions to it; their
comments and suggestions led to a clearly improved edition. Compared to the first
English edition there are major changes. Analogue and analytical photogrammetry are
reduced significantly, most importance is given to digital photogrammetry, and, finally,
laser scanning is included. Terrestrial as well as airborne laser scanning have gained
great importance in photogrammetry. Photogrammetric methods are, with small adaptations, applicable to data acquired by laser scanning. Therefore, only minor additions
to photogrammetry were necessary to cover the chapter on laser scanning. Compared
to the previous German edition there are, especially, updates on digital cameras and
laser scanners.
The original German version arose out of practical research and teaching at the Vienna University of Technology. Volume 1 first introduces the necessary basics from
mathematics and digital image processing. It continues with photogrammetric acquisition technology with special consideration of photo-electrical imaging (CCD cameras).

Particular attention is paid to the use of the Global Positioning System (GPS) and Inertial Measurement Units (IMU) for flight missions. The discussion on photogrammetric
processing begins with orientation methods including those based on projective geometry. The orientation methods which are discussed for two images are extended to
image blocks in the form of photogrammetric triangulation.


vi

Foreword

In the discussion of stereo-plotting instruments most attention is given to digital softcopy stations. In addition to automatic processing methods, semiautomatic methods,
which are widely used in practice, are also explained. This textbook first treats digital orthophoto production, and then includes three-dimensional virtual worlds with
photographic texture.
This selection and arrangement of material offers students a straightforward introduction to complex photogrammetry as practised today and as it will be practised in the
near future. It also offers practising photogrammetrists the possibility of bringing themselves up-to-date with the modern approach to photogrammetry and saves them at least
a part of the tedious study of technical journals which are often difficult to understand.
For technically oriented neighbouring disciplines it provides a condensed description
of the fundamentals and standard processes of photogrammetry. It lays the basis for
that interdisciplinary collaboration which gains ever greater importance in photogrammetry. Related, non-technical disciplines will also find valuable information on a wide
range of topics.
For the benefit of its readers, the textbook follows certain principles: didactics are put
before scientific detail; lengthy derivations of formulae are put aside; theory is split into
small sections alternating with practically-oriented passages; the theoretical basics are
made clear by means of examples; and exercises are provided with solutions in order
to allow self-checking.
This series of textbooks is a major contribution to photogrammetry. It is very sad that
Prof. Kraus, who died unexpectedly in April 2006, cannot see it published. At that time
the translation was already in progress. Final editing was performed by Dr. Josef Jansa
and Mr. Andreas Roncat from the Vienna Institute of Photogrammetry and Remote
Sensing. Thanks are also due to the many people at the Institute of Photogrammetry
and Remote Sensing who did major and minor work behind the scenes, such as drawing

and editing figures, calculating examples and exercises, making smaller contributions,
proofreading, composing the I4TgX text, etc. This book, however, is truly a book by
Prof. Kraus.
Karl Kraus was born in 1939 in Germany and became Professor of Photogrammetry
in Vienna in 1974. Within these 32 years of teaching, counting all translations and
editions, more than twenty textbooks on photogrammetry and remote sensing bearing
the name Karl Kraus were published. Many examples and drawings in this textbook
were supplied by the students and collaborators of Prof. Kraus in Vienna. With deep
gratitude the entire Institute of Photogrammetry and Remote Sensing looks back at the
time spent with Karl Kraus and forward to continuing the success story of this textbook.

Norbert Pfeifer
Professor in Photogrammetry
Institute of Photogrammetry and Remote Sensing
Vienna University of Technology

Vienna, Summer 2007


Notes for readers
This textbook provides an introduction to the basics of photogrammetry and laser scanning. References to Volume 2, Chapters B, C, D, and E, refer to
Kraus, Karl: Photogrammetry, Volume 2, Advanced Methods and Applications, with contributions by J. Jansa and H. Kager. 4 th edition, Diimmler,
Bonn, 1997, ISBN 3-427-78694-3.
Volume 2 is a completely separate textbook and is currently out of print. It covers
advanced topics for readers who require a deeper theoretical knowledge and details of
specialized applications.



Contents

Foreword
Notes for readers

ν
vii

1

Introduction
1.1 Definitions
1.2 Applications
1.3 Some remarks on historical development

2

Preparatory remarks on mathematics and digital image processing . . .
2.1 Preparatory mathematical remarks
2.1.1 Rotation in a plane, similarity and affine transformations . . . .
2.1.2 Rotation, affine and similarity transformations in threedimensional space
2.1.3 Central projection in three-dimensional space
2.1.4 Central projection and projective transformation of a plane . . .
2.1.5 Central projection and projective transformation of the straight
line
2.1.6 Processing a stereopair in the "normal case"
2.1.7 Error theory for the "normal case"
2.2 Preliminary remarks on the digital processing of images
2.2.1 The digital image
2.2.2 A digital metric picture
2.2.3 Digital processing in the "normal case" and digital projective
rectification

40

Photogrammetric recording systems and their application
3.1 The basics of metric cameras
3.1.1 The interior orientation of a metric camera
3.1.2 Calibration of metric cameras
3.1.3 Correction of distortion
3.1.4 Depth of field and circle of confusion
3.1.5 Resolving power and contrast transfer
3.1.5.1 Diffraction blurring
3.1.5.2 Optical resolving power
3.1.5.3 Definition of contrast
3.1.5.4 Contrast transfer function

47
47
47
55
56
58
63
63
64
68
68

3

1

1
2
3
10
10
10
14
21
24
29
31
33
35
36
38


χ

Contents

3.2

3.3

3.4

3.5

3.6

3.1.6 Light fall-off from centre to edge of image
Photochemical image recording
3.2.1 Analogue metric image
3.2.1.1 Glass versus film as emulsion carrier
3.2.1.2 Correcting film deformation
3.2.2 Physical and photochemical aspects
3.2.2.1 Colours and
filters
3.2.2.2 The photochemical process of black-and-white
photography
3.2.2.3 Gradation
3.2.2.4 Film sensitivity (speed)
3.2.2.5 The colour photographic process
3.2.2.6 Spectral sensitivity
3.2.2.7 Resolution of photographic emulsions
3.2.2.8 Copying with contrast control
3.2.3 Films for aerial photography
Photoelectronic image recording
3.3.1 Principle of opto-electronic sensors
3.3.2 Resolution and modulation transfer
3.3.3 Detector spacing (sampling theory)
3.3.4 Geometric aspects of CCD cameras
3.3.5 Radiometric aspects of CCD cameras
3.3.5.1 Linearity and spectral sensitivity
3.3.5.2 Colour imaging
3.3.5.3 Signal-to-noise ratio
Digitizing analogue images
3.4.1 Sampling interval
3.4.2 Grey values and colour values

3.4.3 Technical solutions
Digital image enhancement
3.5.1 Contrast and brightness enhancement
3.5.1.1 Histogram equalization
3.5.1.2 Histogram normalization
3.5.1.3 Compensation for light fall-off from centre to edge
of image
3.5.1.4 Histogram normalization with additional contrast
enhancement
3.5.2 Filtering
3.5.2.1 Filtering in the spatial domain
3.5.2.2 Filtering in the frequency domain
Image pyramids/data compression
3.6.1 Image pyramids

70
71
71
72
73
77
77
79
81
82
84
87
89
91
91

93
93
97
100
102
103
103
104
105
106
107
107
109
110
Ill
114
114
119
120
122
122
125
128
128


Contents

3.7

3.8

4

xi
3.6.2
Aerial
3.7.1
3.7.2

Image compression
cameras and their use in practice
Flight planning
Metric aerial cameras
3.7.2.1 Large format, metric film cameras
3.7.2.2 Digital cameras with CCD area sensors
3.7.2.3 Digital 3-line cameras
3.7.3 Satellite positioning and inertial systems
3.7.3.1 Use of GPS during photogrammetric flying missions
and image exposure
3.7.3.2 Accurate determination of exterior orientation
elements by GPS and IMU
3.7.3.3 Gyro-stabilized platforms and particular features of
line cameras and laser scanners
3.7.4 Image motion and its compensation
3.7.4.1 Compensation of image motion in aerial film cameras
3.7.4.2 Image motion compensation for digital cameras with
CCD area arrays
3.7.4.3 Image motion compensation for digital line cameras .
3.7.5 Effective illumination in aerial photography

3.7.6 Survey aircraft
Terrestrial metric cameras and their application
3.8.1 "Normal case" of terrestrial photogrammetry
3.8.2 Stereometric cameras
3.8.3 Independent metric cameras
3.8.4 Semi-metric cameras
3.8.5 Amateur cameras
3.8.6 Terminology and classification
3.8.7 CCD cameras
3.8.8 Planning and execution of terrestrial photogrammetry

Orientation procedures and some methods of stereoprocessing
4.1 With known exterior orientation
4.1.1 Two overlapping metric photographs
4.1.2 Metric images with a three-line sensor camera
4.2 With unknown exterior orientation
4.2.1 Separate orientation of the two images
4.2.2 Combined, single-stage orientation of the two images
4.2.3 Two-step combined orientation of a pair of images
4.3 Relative orientation
4.3.1 Relative orientation of near-vertical photographs
4.3.2 Relative orientation and model formation using highly tilted
photographs

129
131
131
137
137
144

146
147
147
148
153
155
157
158
159
159
161
163
164
165
166
169
170
170
172
173
180
181
181
183
184
185
188
189
193
193

197


xii

Contents

4.4

4.5

4.6

5

4.3.2.1
Gauss-Helmert model of relative orientation
200
4.3.2.2
A combined, single-stage relative orientation
201
4.3.3 Alternative formulation of relative orientation
201
4.3.4 Relative orientation of near-vertical photographs by
y-parallaxes
205
4.3.4.1
Mountainous country (after Jerie)
206
4.3.4.2

Flat ground (after Hallert)
209
4.3.5 Critical surfaces in relative orientation
210
4.3.6 Error theory of relative orientation
213
4.3.6.1
Standard deviations of the elements of orientation . . 2 1 3
4.3.6.2
Deformation of the photogrammetric model
215
Absolute orientation
219
4.4.1 Least squares estimation
219
4.4.2 Error theory of absolute orientation
226
4.4.3 Determination of approximate values
228
Image coordinate refinement
230
4.5.1 Refraction correction for near-vertical photographs
230
4.5.2 Correction for refraction and Earth curvature in horizontal
photographs
233
4.5.3 Earth curvature correction for near-vertical photographs . . . . 2 3 5
4.5.4 Virtual (digital) correction image
237
Accuracy of point determination in a stereopair

238

Photogrammetric triangulation

246

5.1
5.2

246
248
248
256

5.3

5.4
5.5

Preliminary remarks on aerotriangulation
Block adjustment by independent models
5.2.1 Planimetrie adjustment of a block
5.2.2 Spatial block adjustment
5.2.3 Planimetrie and height accuracy in block adjustment by
independent models
5.2.3.1
Planimetrie accuracy
5.2.3.2
Height accuracy
. . .

5.2.3.3
Empirical planimetric and height accuracy
5.2.3.4
Planimetric and height accuracy of strip triangulation
Bundle block adjustment
5.3.1 Basic principle
5.3.2 Observation and normal equations for a block of photographs .
5.3.3 Solution of the normal equations
5.3.4 Unknowns of interior orientation and additional parameters . .
5.3.5 Accuracy, advantages and disadvantages of bundle block
adjustment
GPS- and IMU-assisted aerotriangulation
Georeferencing of measurements made with a 3-line camera

259
259
265
267
267
269
269
270
273
274
274
276
277


Contents

5.6
5.7
6

xiii
Accounting for Earth curvature and distortions due to cartographic
projections
Triangulation in close range photogrammetry

280
282

Plotting instruments and stereoprocessing procedures
286
6.1 Stereoscopic observation systems
286
6.1.1 Natural spatial vision
286
6.1.2 The observation of analogue and digital stereoscopic images . .288
6.2 The principles of stereoscopic matching and measurement
295
6.3 Analogue stereoplotters
297
6.4 Analytical stereoplotters
300
6.4.1 Stereocomparators
300
6.4.2 Electronic registration of image coordinates in the
monocomparator
302

6.4.3 Universal analytical stereoplotter
304
6.5 Digital stereoplotting equipment
306
6.6 Computer-supported manual methods of analysis
307
6.6.1 Recording in plan
308
6.6.2 Determination of heights
310
6.6.3 Recording of buildings
312
6.6.4 Transition to spatially related information systems
316
6.7 Operator accuracy with a computer assisted system
317
6.7.1 Measurement in plan
317
6.7.1.1 Point measurement
317
6.7.1.2 Processing of lines
318
6.7.2 Height determination
319
6.7.2.1 Directly drawn contours
319
6.7.2.2 Relationship between contour interval and heighting
accuracy
320
6.7.2.3 Contours obtained indirectly from a DTM

321
6.7.2.4 Measurement of buildings
322
6.7.3 Checking of the results
323
6.8 Automatic and semi-automatic processing methods
323
6.8.1 Correlation, or image matching, algorithms
323
6.8.1.1 Correlation coefficient as a measure of similarity . . .324
6.8.1.2 Correlation in the subpixel region
326
6.8.1.3 Interest operators
330
6.8.1.4 Feature based matching
331
6.8.1.5 Simultaneous correlation of more than two images . . 332
6.8.2 Automated interior orientation
334
6.8.3 Automated relative orientation and automated determination of
tie points
335


xiv

Contents

6.8.3.1

Near-vertical photographs with 60% forward overlap taken over land with small height differences . . . 336
6.8.3.2
Near-vertical photographs with 60% forward overlap taken over land with large height differences . . . 337
6.8.3.3
Arbitrary configurations of photographs and objects
with very complex forms
337
6.8.3.4
Line-based (edge-based) relative orientation
338
6.8.3.5
Tie points for automated aerotriangulation
339
6.8.4 Automated location of control points
339
6.8.5 Inclusion of epipolar geometry in the correlation
341
6.8.5.1
Epipolar geometry after relative orientation using
rotations only
342
6.8.5.2
Epipolar geometry in normalized images
343
6.8.5.3
Epipolar geometry in original, tilted metric
photographs
345
6.8.5.4
Derivation of normalized images using the elements

of exterior orientation
346
6.8.5.5
Epipolar geometry in images which have been
oriented relatively using projective geometry
347
6.8.5.6
Epipolar geometry in three images
349
6.8.6 Automated recording of surfaces
350
6.8.7 Semi-automated processing for plan
352
6.8.7.1
Active contours (snakes)
353
6.8.7.2
Sequential processing
355
6.8.8 Semi-automatic measurement of buildings
360
6.8.9 Accuracy and reliability of results obtained by automated or
semi-automated means
363
6.8.10 Special features of the three-line camera
364

7

Orthophotos and single image analysis

7.1
7.2

7.3

7.4

366

Perspective distortion in a metric image
367
Orthophotos of plane objects
373
7.2.1 With vertical camera axis
373
7.2.2 With tilted camera axis
376
7.2.3 Combined projective and affine rectification
378
Orthophotos of curved objects
380
7.3.1 Production principle
380
7.3.2 Orthophoto accuracy
384
Analogue, analytical and digital single image analysis
393
7.4.1 Analogue, analytical and digital orthophoto analysis
393
7.4.2 Analytical and digital analysis of a tilted image of a flat object . 393

7.4.3 Analytical and digital single image analysis of curved object
surfaces
394


Contents

7.5
7.6
8

XV

Photo models
Static and dynamic visualizations

Laser scanning
8.1 Airborne laser scanning
8.1.1 Principle of operation
8.1.2 Analysis and processing
8.1.2.1 Georeferencing
8.1.2.2 Derivation of terrain models
8.1.2.3 Generation of building models
8.1.3 Comparison of two paradigms and further performance
parameters of laser scanners
8.2 Terrestrial laser scanning
8.2.1 Principle of operation
8.2.2 Georeferencing
8.2.3 Connecting point clouds
8.2.4 Strategies for object modelling

8.2.5 Integration of laser data and photographic data
8.3 Short range laser scanning

Appendices
2.1-1
2.1-2
2.1-3
2.2-1
4.1-1
4.2-1
4.3-1
4.6-1

396
399
400
400
400
404
404
407
411
413
419
419
420
422
423
426
428

432
Three-dimensional rotation matrix
432
Mathematical relationship between image and object
coordinates (collinearity condition)
436
Differential coefficients of the collinearity equations
438
Derivation of Formula (2.2-5) using homogeneous coordinates . 440
Estimation by the method of least squares
441
Direct Linear Transformation (DLT) with homogeneous
coordinates
444
Differential coefficients for the coplanarity equations
445
The empirical determination of standard deviations and
tolerances
447

Completion of the references

449

Index

451

Chapter 1
Introduction
1.1 Definitions
Photogrammetry allows one to reconstruct the position, orientation, shape and size of
objects from pictures; these pictures may originate as photochemical images (conventional photography) or as photoelectric images (digital photography). Laser scanner
images, a third group, have arrived in recent years; laser scanner images have distance
information associated with every picture element. The results of a photogrammetric
analysis may be:
• numbers—coordinates of separate points in a three-dimensional coordinate system (digital point determination),
• drawings (analogue)—maps and plans with planimetric detail and contour lines
together with other graphical representation of objects,
• geometric models (digital)—which are fed in to information systems,
• images (analogue and/or digital)—above all, rectified photographs (orthophotos) and, derived from these, photomaps; but also photomontages and so-called
three-dimensional photomodels, which are textured CAD models with textures
extracted from photographs.
That branch of photogrammetry which starts with conventional photographs and in
which the processing is by means of optical-mechanical instruments is called analogue
photogrammetry. That which is based on conventional photographs but which resolves
the whole process of analysis by means of computers is called analytical photogrammetry. A third stage of development is digital photogrammetry. In that case the light
falling on the focal plane of the taking camera is recorded not by means of a lightsensitive emulsion but by means of electronic detectors. Starting from such digital
photographs, the whole process of evaluation is by means of computers—human vision and perception are emulated by the computer. Especially in English, digital photogrammetry is frequently called softcopy photogrammetry as opposed to hardcopy
photogrammetry which works with digitized film-based photographs1. Photogrammetry has some connection with machine vision, or computer vision, of which pattern
recognition is one aspect.
'See PE&RS 58, Copy 1, pp. 49-115, 1992.


2

Chapter 1

Introduction

In many cases interpretation of the content of the image goes hand in hand with the
geometrical reconstruction of the photographed object. The outcome of such photointerpretation is the classification of objects within the images according to various
different characteristics.
Photogrammetry allows the reconstruction of an object and the analysis of its characteristics without physical contact with it. Acquisition of information about the surface of
the Earth in this way is known nowadays as remote sensing. Remote sensing embraces
all methods of acquiring information about the Earth's surface by means of measurement and interpretation of electromagnetic radiation 2 either reflected from or emitted
by it. While remote sensing includes that part of photogrammetry which concerns itself
with the surface of the Earth, if the predominant interest is in geometric characteristics,
one speaks of photogrammetry and not of remote sensing.

1.2

Applications

The principal application of photogrammetry lies in the production of topographic
maps in the form of both line maps and orthophoto maps. Photogrammetric instruments function as 3D-digitizers; in a photogrammetric analysis a digital topographic
model is formed, which can be visualized with the aid of computer graphics. Both the
form and the usage of the surface of the Earth are stored in such a digital topographic
model. The digital topographic models are input in a topographical information system as the central body of data which, speaking veiy generally, provides information
about both the natural landscape and the cultural landscape (as fashioned by man). A
topographic information system is a fundamental subsystem in a comprehensive geoinformation system (GIS). Photogrammetry delivers geodata to a GIS. Nowadays a very
large proportion of geodata is recorded by means of photogrammetry and laser scanning.
Close range photogrammetry is used for the following tasks: architectural recording;
precision measurement of building sites and other engineering subjects; surveillance
of buildings and documentation of damage to buildings; measuring up of artistic and
engineering models; deformation measurement; survey of moving processes (for example, robotics); biometric applications (for example, computer controlled surgical
operations); reconstruction of traffic accidents and very many others.

If the photographs are taken with specialized cameras, photogrammetric processing is
relatively simple. With the help of complex mathematical algorithms and powerful
software, however, the geometric processing of amateur photographs has now become
possible. This processing technology is becoming more and more widely used, especially now that many people have their photographs available on their computers and,
in addition to manipulation of density and colour, are frequently interested in geometric
processing.
2

See DIN 18716/3.


Section 1.3

Some remarks on historical development

3

1.3 Some remarks on historical development3
Technologies arise and develop historically in response both to need and to the emergence and development of supporting techniques and technologies. With the invention
of photography by Fox Talbot in England, by Niepce and Daguerre in France, and
by others, the 1830s and 1840s saw the culmination of investigations extending over
the centuries into optics and into the photo-responses of numerous chemicals. Also
at that time, rapid and cost-effective methods of mapping were of crucial interest to
military organizations, to colonial powers and to those seeking to develop large, relatively new nations such as Canada and the USA. While the practical application of new
technology typically lags well behind its invention, it was very quickly recognized that
cameras furnished a means of recording not only pictorial but also geometrical information, with the result that photogrammetry was born only a few years after cameras
became available. Surprisingly, it was not the urgent needs of mapping but the desire
accurately to record important buildings which led to the first serious and sustained
application of photogrammetry and it was not a surveyor but an architect, the German
Meydenbauer 4 , who was responsible. In fact it is to Meydenbauer that we should be

grateful, or not, for having coined the word "photogrammetry". Between his first and
last completed projects, in 1858 and 1909 respectively, on behalf of the Prussian state,
Meydenbauer compiled an archive of some 16000 metric images of its most important
architectural monuments.
Meydenbauer had, however, been preceded in 1849 by the Frenchman Laussedat, a
military officer, and it is he who is universally regarded as the first photogrammetrist
despite the fact that he was initially using not a camera but a camera lucida, working
on an image of a facade of the Hotel des Invalides in Paris. The work of both of
these scientists had been foreshadowed by others. In 1839 the French physicist Arago
had written that photography could serve "to measure the highest and inaccessible
buildings and to replace the fieldwork of a topographer". Earlier than this, in 1759,
Lambert, a German mathematician, had published a treatise on how to reconstruct
three-dimensional objects from perspective drawings.
The effective production of maps using photogrammetry, which was to become a technological triumph of the 20 th century, was not possible at that time, nor for decades
afterwards; that triumph had to wait for several critical developments: the invention
of stereoscopic measurement, the introduction of the aeroplane and progress in the development of specialized analogue computers. For reasons which will become clear to
readers of this book, buildings provided ideal subjects for the photogrammetric techniques of the time; topographic features most certainly did not. Without stereoscopy,
measurement could be made only of very clearly defined points such as are to be found
on buildings. Using cameras with known orientations and known positions, the threedimensional coordinates of points defining a building being measured photogrammet3

Permission from the publishers to use some of the historical material from "Luhmann, T., Robson,
S., Kyle, S., Harley, I.: Close Range Photogrammetry. Whittles Publishing, 2006" is gratefully acknowledged. That material and this present section were both contributed by one of the translators, Ian Harley.
4
A limited bibliography, with particular reference to historical development, is given at the end of this
chapter.


4

Chapter 1

Introduction

rically were deduced using numerical computation. The basic computational methods
of photogrammetry were established long ago.
By virtue of their regular and distinct features, architectural subjects lend themselves
to this technique which, despite the fact that numerical computation was employed,
is often referred to as "plane table photogrammetry". When using terrestrial pictures
in mapping, by contrast, there was a major difficulty in identifying the same point on
different photographs, especially when they were taken from widely separated camera
stations; and a wide separation is desirable for accuracy. It is for these reasons that so
much more architectural than topographic photogrammetry was performed during the
19th century. Nonetheless, a certain amount of topographic mapping by photogrammetry took place during the last three decades of that century; for example mapping in the
Alps by Paganini in 1884 and the mapping of vast areas of the Rockies in Canada by
Deville, especially between 1888 and 1896. Jordan mapped the Dachel Oasis in 1873.
In considering the history of photogrammetry the work of Scheimpflug in Austria
should not be overlooked. In 1898 he first demonstrated double projection, which
foreshadowed purely optical stereoplotters. In particular his name will always be associated with developments in rectification.
The development of stereoscopic measurement around the turn of the century was a
momentous breakthrough in the history of photogrammetry. The stereoscope had already been invented between 1830 and 1832 and Stolze had discovered the principle
of the floating measuring mark in Germany in 1893. Two other scientists, Pulfrich in
Germany and Fourcade in South Africa, working independently and almost simultaneously 5 , developed instruments for the practical application of Stolze's discovery. Their
stereocomparators permitted stereoscopic identification of, and the setting of measuring marks on, identical points in two pictures. The survey work proceeded point by
point using numerical intersection in three dimensions. Although the landscape could
be seen stereoscopically in three dimensions, contours still had to be plotted by interpolation between spot heights.
Efforts were therefore directed towards developing a means of continuous measurement and plotting of features, in particular of contours—the "automatic" plotting machine, in which numerical computation was replaced by analogue computation for resection, relative and absolute orientation and, above all, for intersection of rays. Digital
computation was too slow to allow the unbroken plotting of detail, in particular of contours, which stereoscopic measurement seemed to offer so tantalisingly. Only analogue
computation was fast enough to provide continuous feedback to the operator. In several
countries during the latter part of the 19th century, much effort and imagination was directed towards the invention of stereoplotting instruments, necessary for the accurate
and continuous plotting of topography. In Germany Hauck proposed such an apparatus.

In Canada Deville developed what was described by Ε. H. Thompson as "the first automatic plotting instrument in the history of photogrammetry". Deville's instrument had
several defects, but its design inspired several subsequent workers to overcome these,
5
Pulfrich's lecture in Hamburg announcing his invention was given on 23 rd September 1901, while
Fourcade delivered his paper in Cape Town nine days later on 2 nd October 1901.


Section 1.3

Some remarks on historical development

5

including both Pulfrich, one of the greatest contributors to photogrammetric instrumentation, and Santoni in Italy, perhaps the most prolific of photogrammetric inventors.
Photogrammetry was about to enter the era of analogue computation, a very foreign
idea to surveyors with their long tradition of numerical computation. Although many
surveyors regarded analogue computation as an aberration, it became a remarkably
successful one for a large part of the 20th century.
In Germany, conceivably the most active country in the early days of photogrammetry,
Pulfrich's methods were very successfully used in mapping; this inspired von Orel in
Vienna to design an instrument for the "automatic" plotting of contours, leading ultimately to the Orel-Zeiss Stereoautograph which came into productive use in 1909. In
England, F. V. Thompson was slightly before von Orel in the design and use of the
Vivian Thompson Stereoplotter; he went on to design the Vivian Thompson Stereoplanigraph, described in January 1908, about which Ε. H. Thompson was to write that
it was "the first design for a completely automatic and thoroughly rigorous photogrammetric plotting instrument". The von Orel and the Thompson instruments were both
used successfully in practical mapping, Vivian Thompson's having been used by the
Survey of India which bought two of the instruments.
The advantages of photography from an aerial platform, rather than from a ground
point, are obvious, both for reconnaissance and for survey; in 1858 Nadar, a Paris
photographer, took the first such picture, from a hot-air balloon 1200 feet above that
city, and in the following year he was ordered by Napoleon to obtain reconnaissance

photographs in preparation for the Battle of Solferino. It is reputed that balloon photography was used during the following decade in the American Civil War. The rapid
development of aviation which began shortly before the first World War had a decisive
influence on the course of photogrammetry. Not only is the Earth, photographed vertically from above, an almost ideal subject for the photogrammetric method, but also
aircraft made almost all parts of the Earth accessible at high speed. In the first half, and
more, of the 20th century these favourable circumstances allowed impressive development in photogrammetry, although the tremendous economic benefit in air survey was
not fully felt until the middle of that century. On the other hand, while stereoscopy
opened the way for the application of photogrammetry to the most complex surfaces
such as might be found in close range work, not only is the geometry in such cases often far from ideal photogrammetrically but also there was no corresponding economic
advantage to promote its application.
In the period before the first World War all the major powers followed similar paths in
the development of photogrammetry. After the war, although there was considerable
opposition from surveyors to the use of photographs and analogue instruments for mapping, the development of stereoscopic measuring instruments forged ahead remarkably
in very many countries; while the continental European countries broadly speaking
put most of their effort into instrumental methods, Germany and the Austro-Hungarian
Empire having a clear lead in this field, the English-speaking countries focused on
graphical techniques. It is probably true that until about the 1930s the instrumental techniques could not compete in cost or efficiency with the British and American
methods.


6

Chapter 1

Introduction

Zeiss, in the period following WWI, was well ahead in the design and manufacture
of photogrammetric instruments, benefiting from the work of leading figures such as
Pulfrich, von Orel, Bauersfeld, Sander and von Gruber. In Italy, around 1920, Santoni
produced a prototype, the first of many mechanical projection instruments designed
throughout his lifetime, while the Nistri brothers developed an optical projection plotter, shortly afterwards founding the instrument firm OMI. Poivilliers in France began

the design and construction of analogue photogrammetric plotters in the early 1920s.
In Switzerland the scene was dominated by Wild whose company began to produce instrumentation for terrestrial photogrammetry at about the same time; Wild Heerbrugg
very rapidly developed into a major player, not only in photogrammetric instrumentation, including aerial cameras, but also in the wider survey world. As early as 1933
Wild stereometric cameras were being manufactured and were in use by Swiss police for the mapping of accident sites, using the Wild A4 Stereoautograph, a plotter
especially designed for this purpose. Despite the ultra-conservative establishment in
the British survey world at that time, Ε. H. Thompson was able to design and build a
stereoplotter in the late 1930s influenced by the ideas of Fourcade. While the one such
instrument in existence was destroyed by aerial bombing, the Thompson-Watts plotter
was later based on this prototype in the 1950s.
Meanwhile, non-topographic use was sporadic for the reasons that there were few suitable cameras and that analogue plotters imposed severe restrictions on principal distance, on image format and on disposition and tilts of cameras.
The 1950s saw the beginnings of the period of analytical photogrammetry. The expanding use of digital, electronic computers in that decade engendered widespread interest
in the purely analytical or numerical approach to photogrammetry as against the prevailing analogue methods. While analogue computation is inflexible, in regard to both
input parameters and output results, and its accuracy is limited by physical properties,
a numerical method allows virtually unlimited accuracy of computation and its flexibility is bounded only by the mathematical model on which it is based. Above all,
it permits over-determination which may improve precision, lead to the detection of
gross errors and provide valuable statistical information about the measurements and
the results. The first analytical applications were to photogrammetric triangulation, a
technique which permits a significant reduction in the amount of ground control required when mapping from a strip or a block of aerial photographs; because of the very
high cost of field survey for control, such techniques had long been investigated. In
the 1930s, the slotted template method of triangulation in plan was developed in the
USA, based on theoretical work by Adams, Finsterwalder and Hotine. Up until the
1960s vast areas were mapped in the USA and Australia using this technique in plan
and one of the many versions of the simple optical-projection Multiplex plotters both
for triangulation in height and for plotting of detail. At the same time, precise analogue
instruments such as the Zeiss C8 and the Wild A7 were being widely used for analogue
triangulation in three dimensions.
Analytical photogrammetric triangulation is a method, using numerical data, of point
determination involving the simultaneous three-dimensional orientation of all the photographs and taking all inter-relations into account. Work on this line of development

Section 1.3

Some remarks on historical development

7

had appeared before WWII, long before the development of electronic computers. Analytical triangulation demanded instruments to measure photo coordinates. The first
stereocomparator designed specifically for use with aerial photographs was the Cambridge Stereocomparator designed in 1937 by Ε. H. Thompson. Electronic recording
of data for input to computers became possible and by the mid-1950s there were five
automatic recording stereocomparators on the market and monocomparators designed
for use with aerial photographs also appeared.
Seminal papers by Schmid and Brown in the late 1950s laid the foundations for theoretically rigorous photogrammetric triangulation. A number of block adjustment programs for air survey were developed and became commercially available, such as those
by Ackermann.
Subsequently, stereoplotters were equipped with devices to record model coordinates
for input to electronic computers. Arising from the pioneering ideas of Helava, computers were incorporated in stereoplotters themselves, resulting in analytical stereoplotters
with fully numerical reconstruction of the photogrammetric models. Bendix/OMI developed the first analytical plotter, the AP/C, in 1964; during the following two decades
analytical stereoplotters were produced by the major instrument companies and others.
Photogrammetry has progressed as supporting sciences and technologies have supplied
the means such as better glass, photographic film emulsions, plastic film material, aeroplanes, lens design and manufacture, mechanical design of cameras, flight navigation
systems. Progress in space technology (both for imaging, in particular after the launch
of SPOT-1 in 1986, and for positioning both on the ground and in-flight by GNSS) and
the continuing explosion in electronic information processing have profound implications for photogrammetry.
The introduction of digital cameras into a photogrammetric system allows automation,
nowhere more completely than in industrial photogrammetry, but also in mapping.
Advanced computer technology enables the processing of digital images, particularly
for automatic recognition and measurement of image features, including pattern correlation for determining object surfaces. Procedures in which both the image and its
photogrammetric processing are digital are often referred to as digital photogrammetry.
Interactive digital stereo systems (e.g. Leica/Helava DSP, Zeiss PHODIS) have existed
since around 1988 (Kern DSP 1) and have increasingly replaced analytical plotters.
To some extent, photogrammetry has been de-skilled and made available directly to a

wide range of users. Space imagery is commonplace, as exemplified by Google Earth.
Photogrammetric measurement may be made by miscellaneous users with little or no
knowledge of the subject—police, architects, model builders for example.
Although development continues apace, photogrammetry is a mature technology with
a history of remarkable success. At the start of the 20 th century topographic mapping
of high quality existed, in general, only in parts of Europe, North America and India.
Although adequate mapping is acknowledged as necessary for development but is still
lacking in large parts of the world, such deficiencies arise for political and economic
reasons, not for technical reasons. Photogrammetry has revolutionized cartography.


8

Chapter 1

Introduction

Further reading. Adams, L.P.: Fourcade: The centenary of a stereoscopic method
of photographic surveying. Ph.Rec. 17(99), pp. 225-242, 2001 · Albertz, J.: A Look
Back—Albrecht Meydenbauer. PE&RS 73, pp. 504-506, 2007 . Atkinson, K.B.:
Vivian Thompson (1880-1917): not only an officer in the Royal Engineers. Ph.Rec.
10(55), pp. 5-38, 1980 · Atkinson, K.B.: Fourcade: The Centenary—Response
to Professor H.-K. Meier. Correspondence, Ph.Rec. 17(99), pp. 555-556, 2002 .
Babington-Smith, C.: The Story of Photo Intelligence in World War II. ASPRS, Falls
Church, Virginia, 1985, reprint · Blachut, T.J. and Burkhardt, R.: Historical development of photogrammetric methods and instruments. ISPRS and ASPRS, Falls
Church, Virginia, 1989 · Brown, D.C.: A solution to the general problem of multiple
station analytical Stereotriangulation. RCA Data Reduction Technical Report No. 43,
Aberdeen, 1958 · Brown, D.C.: The bundle adjustment—progress and prospectives.
IAPR 21(3), ISP Congress, Helsinki, pp. 1-33, 1976 · Deville, E.: Photographic
Surveying. Government Printing Bureau, Ottawa, 1895. 232 pages · Deville, E.:

On the use of the Wheatstone Stereoscope in Photographic Surveying. Transactions
of the Royal Society of Canada, Ottawa, 8, pp. 63-69, 1902 . Fourcade, H.G.: On
a stereoscopic method of photographic surveying. Transactions of the South African
Philosophical Society 14(1), pp. 28-35, 1901. Also published in: Nature 66(1701),
pp. 139-141, 1902 · Fourcade, H.G.: On instruments and methods for stereoscopic
surveying. Transactions of the Royal Society of South Africa 14, pp. 1-50, 1903 ·
Fräser, C.S., Brown, D.C.: Industrial photogrammetry—new developments and recent
applications. Ph.Rec. 12(68), pp. 197-216, 1986 . von Gruber, 0., (ed.), McCaw,
G.T., Cazalet, F.A., (trans.): Photogrammetry, Collected Lectures and Essays. Chapman & Hall, London, 1932 · Gruen, Α.: Adaptive least squares correlation—a powerful image matching technique. South African Journal of Photogrammetry, Remote
Sensing and Cartography 14(3), pp. 175-187, 1985 · Harley, I.A.: Some notes on
stereocomparators. Ph.Rec. IV(21), pp. 194-209, 1963 · Helava,U.V.: New principle
for analytical plotters. Phia 14, pp. 89-96, 1957 · Kelsh, H.T.: The slotted-template
method for controlling maps made from aerial photographs. Miscellaneous Publications no. 404, U.S. Department of Agriculture, Washington, D.C., 1940 · Landen,
D.: History of photogrammetry in the United States. Photogrammetric Engineering
18, pp. 854-898, 1952 · Laussedat, Α.: Memoire sur l'emploi de la photographie
dans le leve des plans. Comptes Rendus 50, pp. 1127-1134, 1860 · Laussedat, Α.:
Recherches sur les instruments, les methodes et le dessin topographiques. GauthierVillars, Paris, 1898 (vol. 1), 1901 (vol. 2 part 1), 1903 (vol. 2 part 2) · Luhmann, T„
Robson, S., Kyle, S. and Harley, I: Close Range Photogrammetry. Whittles Publishing,
2006. 510 pages · Mason, K.: The Thompson stereo-plotter and its use. Survey of
India Departmental Paper No. 5, Survey of India, Dehra Dun, India, 1913 · Meier,
H.-K.: Fourcade: The Centenary—Paper by L.P. Adams. Correspondence, Ph.Rec.
17(99), pp. 554-555, 2002 · Meydenbauer, Α.: Handbuch der Messbildkunst. Knapp,
Halle, 1912. 245 pages · Poivilliers, G.: Address delivered at the opening of the Historical Exhibition, Ninth International Congress of Photogrammetry, London, 1960.
IAPR XIII(l), 1961 · Pulfrich, C.: Über neuere Anwendungen der Stereoskopie und
über einen hierfür bestimmten Stereo-Komparator. Zeitschrift für Instrumentenkunde
22(3), pp. 65-81, 1902 · Sander, W.: The development of photogrammetry in the
light of invention, with special reference to plotting from two photographs. In: von