PHOTOGRAMMETRY OVERVIEW

Beginning – Evolution - Current Status – Future preview.

And some of my professional retrospective.
_-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Preamble
Official definition - from the International Society for Photogrammetry and Remote Sensing (ISPRS): “the science and
technology of extracting reliable three-dimensional Geometric and thematic information, often over time, of objects and
scenes from image and range data.

Etymology - from 3 Greek radicals: Photon, meaning light, Gramma, meaning drawing, and Metron, meaning
measurement. Hence, the term comprises: measurement of what is drawn by the light. Photographs, are the most common
objects drawn by light
Layman's explanation - Land Surveying is a well-known technology for making maps by means of measurements on the
Earth's surface. Similarly, Photogrammetry is the science and technology of making maps by means of measurements on
film-based overlapping photographs. Furthermore, since the advent of precise scanners and digital cameras, it became
possible to do measurements on digital images.

Land Surveying Photogrammetry

Beginning
It was right after the perfection of photography by Niepce and Daguerre between 1829 and 1839, the introduction of
glass-negatives by Archer in 1850, and the invention of the photogoniometer by Ignacio Porro in 1855, that
Photogrammetry was conceived in 1864 by the French engineer Aimé Laussedat (titled the father of Photogrammetry).
Laussedat did a map of the city of Paris using photographs taken from kites, balloons and rooftops; he used a
photogoniometer, which became the earliest map drawing instrument. And he named his mapping method as metro-
photography. The term Photogrammetry was unveiled by A. Meydenbauer and S. Finsterwalder in Munich, Germany, in
1870.

Instrumental evolution
Further to photogoniometers, in 1895/96, the photo-theodolites introduced by F. Finsterwalder and Carl Koppe started to
be built and used for mapping. In 1893, F. Stoltz introduced the concept of the floating mark (index mark), enabling the
designing and building of a new era of instruments. That was the start of the optical-mechanical (analog)
Photogrammetric instruments. In 1900, Carl Pulfrich introduced his stereo-comparator, von Orel his stereo-autograph, and
Otto von Grubber published his research ‘single-and double-point resection in space', defining the best areas on the stereo
model to choose relative and absolute orientation points, which became known as Gruber points. From there on, several
companies became notorious in Germany, Italy, and Switzerland as experts in designing and building analog instruments
for use on both aerial (airborne cameras) and terrestrial (cameras on a tripod) photogrammetry. Those companies were
also experts in designing and building film metric cameras to cope with instrument advancements. The stereo viewing
system of analog instruments was then split among optical trains, anaglyphical (red and blue glasses), and polaroid
glasses. And further, with the coming digital era, liquid crystal glasses were introduced.
With the advent of digital computers, analytical instruments (computer-assisted) era did surge. Then, most of the
mechanical operations required to perform the orientations of the aerial or terrestrial photos, in relation one to each other
and in relation to the object, started to be supported by computer operations. Map depiction, which was made using
precise X - Y pen drawing tables, was replaced by special CAD software.
Orientations - are the reconstructions, on a photogrammetric instrument, of the geometry of the camera (interior
orientation) and its position at the moment of exposure. Practically, it establishes the interrelation among photos (relative
orientation) and the reference of the photos with the object (absolute orientation). As the photos (or digital images)
encompass 3 rotations, 3 translations and 1 scale factor, in relation one to each other, and in relation to the object (camera
attitude).
Technological evolution
Plane table phase (1850 to 1900), of 2D photogrammetric mapping. There were no corrections of photograph distortions
until the advent of the photo rectifiers around 1913.
Analog Photogrammetry (1900 to 1960), also called classical or conventional photogrammetry. It was remarkably
influenced by the invention of the airplane which facilitated taking projected overlapping photographs, which are ‘sine
qua non’ for artificial stereo vision. Analog instruments were designed to reconstruct the attitude of the camera's exposure
moment to form the stereo-model, required to make 3D thematic maps. During this period, the costliest project phase was
the establishment of ground control points (GCPs) to reference the stereo-model to the object surface. Nominally, the
absolute orientation. Theoretically, 3 GCPs are needed on the stereo model to perform the absolute orientation. Hence, the
photogrammetric technique of Aerial-Triangulation (AT) was developed to establish auxiliary and redundant control
points, at least 6 photogrammetric points per model (9 per photo). AT itself must be supported by GCPs established on the
field, and in accordance with the requested accuracy for the mapping project. In 1981, during my ‘forschungsarbeiten’ at
the University of Stuttgart, Germany, I derived empirical formulas to define where to place chains (ip, ih) of ground
control to support AT projects and preview expected final accuracies, as follows:
Using ground control points μxy ~ (0.65+0.25ip) μo μz ~ (1.05+0.36ih) μo
Using kinematics (airborne) GPS μxy ~ 0.8μo + 0.1μGPS μz ~ 1.2μo + 0.15μGPS
ip = interval among planimetric controls and ih = interval among height controls
μo = precision of the photogrammetric observations and μGPS = standard deviation of GPS observations.
μxy, μz = expected final accuracies for planimetry and height respectively.
Accuracy expectation as function of tie-points configuration
μxy ~ 0.7 to 1.3 μo (with 27 points per image) and μz ~ 1.6 to 2.9 μo. (with 9 points per image).
Also, the following graph shows the progress of Aerial-Triangulation methods and cost reduction according to its
technological progress.

Airborne (kinematic) GPS started to be used at

the middle 90's to determine ground coordinates
of the camera perspective centers. Due to lower
(over meters) accuracy it was initially used as
auxiliary control; until it was combined with
other international systems to form the Global
Navigation Satellite System (GNSS), increasing
the accuracy to centimeters. At this level of
accuracy, the above formulas to figure out
intervals (ip, ih) to place ground control chains
became obsolete.

Analytical Photogrammetry (1960 – mid-90's), Uuno Vilho (Uki) Helava (deceased) born in Finland and a researcher at
the Canadian Research Council, invented and patented the analytical plotter in 1957. Initially, the instruments were built
by an Italian-USA partnership for military clients. At the late 70's, digital computers became popular, and several Helava
systems became advantageous and popular in the photogrammetric mapping industry.
In all, the same processes performed mechanically by a human operator to form stereo models were replaced by software.
And big changes in the stockholders of photogrammetric instruments did start to happen. An Australian company succeed
converting analogue into analytical instruments by replacing their mechanical parts with a computer, preserving the
optical part for stereovision. The new era certainly improved project accuracy, reducing the time of projects execution and
training for the operators involved in the industry. New technical jargons started to be used. And, in the early sixties,
Geographic Information Systems (GIS) started to become widespread. Roger Tomlinson (deceased), a Canadian
Geographer, is acknowledged worldwide as the developer of computerized GIS; due to his pioneering work, in 1960, for
the Canada Land Inventory (CLI).
Satellite photogrammetry age - in May 1973, the SKYLAB was launched as the 1 st space station designed for scientific
research. It contained a metric camera with a resolution of 20 - 30 meters. I did use some of those photos for Aerial-
Triangulation (AT) to improve Earth's 1 satellite images accuracy. The Earth's 1 (current LANDSAT) imagery resolution
was between 80 and 100 meters.
At the 1988 ISPRS international congress in Kyoto, Japan, Dipl. Math. Hermann Klein (deceased) launched his company
Photogrammetry, and his version of PAT-M, and later, PAT-B running on personal computers (PCs). Both programs
were developed at the University of Stuttgart, Germany to run AT block adjustments. And became a worldwide favourite
for their envisioned purpose. I was a international individual representative from its start till its end in 2023. And was
very active in American and Asian countries.
In 1979, I and a remarkable computer program engineer, re-engineered 3 mapping production sectors of the ‘Instituto
Brasileiro de Geografia e Estatísticas’ (IBGE) Cartographic Branch. 1) the Photogrammetric Restitution Sector - using a
X-Y digital pen plotter to generate basic sheets for map depicting. Consequently, the Sector of mechanical plotting of
basic map sheets became obsolete. 2) the AT Sector – using PAT-M and magnetic tapes for data recording & backups.
Consequently, eliminating computer punched cards for data entry, and storage of printed reports. That was my start using
PAT programs.

Digital Photogrammetry (mid 90’s to present), - is entirely based on the analytical photogrammetry mathematical models.
However, with astonishing changes on instruments, cameras, and ‘modus operandi’. Photogrammetric instruments (now
photogrammetric workstations) became a computer equipped with sophisticated software, a stereovision system (normally
liquid crystal glasses) and a high-tech-mouse or handwheels. At the beginning, film-photographs (hard copy) had to be
converted into digital image (soft copy) by means of highly precise and expensive scanners. With the advent of digital
cameras, those scanners became obsolete. Commercial airborne digital cameras started to be used in 2000 permitting
obtaining highly resolution images directly from the air. Also, the transition from film photographs enhanced the
processing and integrity of thematic mapping. Furthermore, digital cameras contributed to eliminating spending on films,
laboratory processing and scanning of the hard copies. The integration of digital camera, kinematic GNSS, inertial
measuring systems like INS and IMU and reliable digital photogrammetric software was decisive to obtain high quality
Georeferenced data and progress towards the use of UAVs to dimmish use of airplanes, their required logistics and its
dependence on climate conditions. UAVs also simplified the use of multi direction flights, multi flying heights, multi
overlaps and use of mutual digital cameras. Hence, eliminating the high costs of obtaining the primary photogrammetric
tool; that is, photographs.

In 1984, Professor Friedrich Ackerman (deceased), University of Stuttgart – Germany, published the fundamentals and
performance tests for the procedure of least squares window matching, objectifying high precision parallax assessment,
point marking & transfer, and measurements. The points cloud concept. Further contributions to such technology were
made by other scientists, including Uki Helava. The methodology started to be fully used commercially by the beginning
of this century. From then on, measurements of AT points were replaced by automatic measurements (image correlation),
and AT became Auto-AT (AAT). Consequently, highly trained photogrammetrists were substituted by computer experts,
since precision measurements no longer depend on human visual acuity. Replacement of traditional instruments by
computers eliminated dependency on optical mechanics like lead screws, encoders, optical train, measuring mark size, or
a stable floor. Hence, no more instrument parts periodic calibrations.
In 1996, I designed the digital AT scheme for measurements and unique points coding procedures for a Canadian
Company developer of digital photogrammetric systems.

In 2016, I was asked to solve a ground control issue on a large AAT block adjustment project. Follows some info on the
final adjustment with PAT-B. I believe this is still a record in terms block size and final results.
 Imagery by a 79.8mm principal
distance digital camera.
Nominal photo scale 1/6250.
Precision of the AAT observations
σo = 4.69 μm
Final accuracy of the adjusted GPS
observations
E = 0.40m, N = 0.41m, H = 0.29m.
Better results in height then in
planimetry is due to a long shoreline
used as Z equality constrain.
Average of generated AAT
points per image, 35

Current Status
Due to its outstanding advances into the digital world, Photogrammetry is no longer serving only for basic vector mapping
(cartography) but also for digital thematic mapping, digital GIS prime layer, digital terrain model (DTM), digital surface
model (DSM), digital-orthophoto-maps. The best example of success on digital photogrammetry is Google Earth, a
platform involving satellite & digital imagery, Photogrammetry and GIS. Furthermore, a new technological lingo did
surge, ‘Geomatics - the science and technology of collecting, managing, and analyzing data about the Earth and its
phenomena’. The term encompasses several categories of Geosciences typically applied to environmental management
and infrastructure development.
Photogrammetry and LiDAR integration – both technologies are complementary; LiDAR providing precise 3D structural
measured data and photogrammetry providing Georeferenced colorful visual details from digital images.
In 1984, prior to the spread of LiDAR into the digital mapping industry, I manage to set a pilot project involving a Forest
Penetrating Radar (FPR) from a Canadian Company, Prof. Ackermann from the University of Stuttgart and a Brazilian
mapping company. Major objective of the project was to combine FPR data with overlapping photographs to map dense
forested areas like the Amazonian. Unfortunately, due to alleged national security, the pilot project was not authorized to
be realized.

Future preview
Photogrammetry keeps evolving! And AI was earlier incorporated for ground control and check points targeting. There
are authors already speculating about future GeoAI systems based on preliminary assessment, early experiences, research,
and predictions. Nowadays, data collection for mapping project is very precise. However, what about project integrity
(precision + accuracy)? A comprehensive distinction between precision and accuracy is a digital watch. It is precise
within less than a second per day. Nevertheless, if the watch is not related to the local time zone, it isn’t accurate.
Similarly, a digital mapping project that is not Georeferenced (absolute oriented) isn’t accurate. Thus, a photogrammetrist
must be involved in the project, to provide non-photogrammetrists with adequate guidance and assurance about the
project's integrity.
More about GeoAI (Digital Photogrammetry + GIS + AI enhancements). Such combination may provide advanced world
information in terms of reports, diagrams, graphs, images, location, voice info, and so on, stored on optical discs; a
medium that uses laser light to read and write encoded information.
As mentioned before, A.I. is already integrated into digital mapping, initially as intelligent targeting of ground control and
check points. Moreover, to convert labor-intensive photogrammetric mapping into a highly automated map depiction. And
to classify land cover type for environmental applications & urban planning, automated extraction (pattern recognition) of
features like forest, vegetation & crop types, roads, highways & railways, medicine, multi usage sites, animal habitats and
its behaviors, automated vehicles support. And further developments to come.

Further readings
 DIGITAL CAMERAS and IMAGERY RESOLUTION CONCEPTS
Activity | J. Carlos Maia | LinkedIn
 Essentials to achieve Ground Sample Data (GSD) Goal
Activity | J. Carlos Maia | LinkedIn
JCMaia-Jan/26