Cameras

([Link] Vol.1 § 3.1 - 3.4 - 3.6 -

3.7 ) not in this order

AA 2008 -2009 1
 Airborne and terrestrial analogue sensors
So far : theory based on idealized models Metric photograph

Pin hole camera

Exact central projection

Elements of Inner Orientation

Now: real instruments

Distorsions in lenses, films, etc…..

AA 2008 -2009 2
Airborne and terrestrial analogue sensors

Airborne cameras

Big format photograms (fixed)

Big blocks in normal geometry
Photos taken in motion
Resistant to high stress (pressure
and temperature jumps,vibrations)

Terrestrial cameras

Various medium formats

Photos taken with a still support
Normal or convergent geometry
Stereometric cameras (bi-cameras)

AA 2008 -2009 3
 CAMERA CLASSIFICATIONS
• METRIC CAMERAS developed for photogrammetric surveys, with
stable and measured I.O.
airborne and terrestrial

• SEMIMETRIC CAMERAS i.o. not completely stable

Close range
• NON METRIC CAMERAS (amateur, for low accuracy
survey)

DIFFERENT METHODS FOR INNER ORIENTATION

DETERMINATION :
• FIDUCIAL MARKS fixed on the image when is taken airborne and terrestrial

• RESEAUX CAMERAS : a thin glass plate bearing a calibra-

ted grid of regularly spaced crosses is imaged in every photo
close range
• FRAME CAMERAS : sharp image edge

AA 2008 -2009 4
 Terrestrial cameras
Cameras in fixed positions
E.O parameters once measured by
field survey

[Link]

Fiducial marks metric

Grid of targets semi-metric
Photo frame non-metric

Easily levelled on a tripod

Interchangeable with theodolite and targets

Horizontal or vertical tilt possibility at fixed angles

AA 2008 -2009 5
 Terrestrial metric cameras
Similar to airborne cameras
• fiducial marks (4)
• film flattening device (vacuum or pressure-plate)
• large format
• high quality optics
possible stereometric
• Examples:
WILD P31 (102x127 mm)
Zeiss UMK (130x180 mm)

Stereometric cameras can be

=
normal case AA 2008 -2009 6
 Terrestrial semi-metric cameras
Have a calibrated objective mounted on amateur body

Professional cameras with medium (60x60mm) format

adapted to photogrammetric use
• Reseaux (9 o 121 crosses) example: Rollei 6006
Defines the image reference system
Allows corrections of film deformations
Image coordinates are calibrated for each photogram

• Focusing at fixed distances

• Good quality optics
• Laboratory calibration certificate
• Today also in digital version
AA 2008 -2009 7
 Terrestrial non-metric cameras
or
amateur cameras
Common cameras

Can be used in photogrammetry only after self-calibration

procedure of all Orientation parameters (Int. and Ext.)

AA 2008 -2009 8
AERIAL CAMERAS
Exact geometry
Final requirements Image sharpness
High resolution
Central shutter
Working condition
Aerial cameras are mounted into an aircraft

Suppression of radio-interference, work with 28V DC

and follow safety flight regulations

• Big dimensions :
about 60x60x50 cm, weight about 70-90 kg

• Big format film 23 cm x 23 cm (9” x 9”) and magazine

with 120-150m film (500-600 photograms)

AA 2008 -2009 9
PRINCIPAL REQUIREMENTS IN AERIAL CAMERAS

• fixed principal distance c

Photos are taken at big distances (100m - Nx1000m) therefore there is no
problem with focusing distance: the objective is focused at infinity.

• fiducial marks (reperes) 4 or, better, 8, on the image plane (lightened) + 1

asymetrically placed inside or at the edge of the image, for the automatic
determination of image position

accurate construction of each component (optics, mechanical parts, [Link].

central shutters)

• automated device for flattening of the film, motion and drift corrections

• I.O. parameters by lab-calibration and components stability respect to

vibrations, big temperature changes, thermal expansion

AA 2008 -2009 10
AERIAL CAMERAS
Photogrammetric cameras classification according
field angle of the objectivs :

Angle with vertex in projection center O and rays

pointing to the ends of image diagonal

For the format 230 mm x 230 mm:

Reduced Normal WideAngle Super -

Normal N W WideAngle
SW
Field angle 30° 60° 90° 120°

Principal distance 600 300 150 90

c (mm)

Big scale survey

Medium and large scale Small scale
([Link] areas,
AA 2008
(more -2009
commonly used) 11
orthoimages)
AERIAL CAMERAS

c, Z, B and photograms overlapping are

geometrically constrained each other

Reduced Normal N WideAngle Super -

Normal W WideAngle
SW
Field angle 30° 60° 90° 120°

Principal distance 600 300 150 90

c (mm)
B/Z for 60% 1:6.6 1:3.3 1:1.6 1:0.95
overlapping

AA 2008 -2009 12
 Auxiliary data imaged with the photogram , digital, or analogical
To reconstruct [Link]. each photogram must bear:

• fiducial marks (to fix the principal point position)

• the principal distance
• camera number (numero di matricola) referring to the calibration certificate

And many others :

1 6 1 Circular level
Fiducial marks 1
3
13
2 2
(4/8 + 1) 5 Time
10
4
14 Course and
11 5
7
fine height
15 indicators
8
1 16 1
9
7
17 12 8 9
3
4 N camera
[Link] number 18 18 [Link].
[Link]. data N magazine
[Link] [Link]-scale 6
1 1 Project name
[Link] time, aperture Date
[Link] of image motion
AA 2008 -2009 13
[Link] wedge
 Schematic diagram of an aerial camera
Oculare Magazzino del film Indicatori trasporto film

OverlapCinederivometro
control Aspirazione Pressure
Piastra plate
pressa-pellicola
Cannocchiale di
navigazione Rotolo di pellicola
Viewfinder telescope Rollfilm
larghezza 24cm
Controllo di trasporto
Controle instruments
aspirazione Cono
Lens obiettivo
cone eand camera frame
Strumenti di controllo telaio della camera
e interruttori
Spherical level Elettronica di controllo
Livella sferica
28 Volt CC
K-clampOrientamento k
Obiettivo Supporto della camera
3 viti calanti
3 footwcrews
per φ ωfor φω Cavo di massa
Supporti anti-vibrazioni
Ventre dell'aereo
Otturatore Diaframma Portello di chiusura
Filtro

Auxiliary data imaged on the photogram

Geographic coordinates φ, λ
Projection center height
Angles ω, φ
Course and drift

If present: GPS/INS E.O. data

AA 2008 -2009 14
FURTHER REQUIREMENTS IN AERIAL CAMERAS
E.O. requirements
Possibility of reconstruct E.O. :

The camera

•Can be levelled with footscrew or gyros to mantain tilt angles ω and ϕ <
±5gon

•Can be rotated around its vertical axis to compensate drift (to be aligned to
the ground track of the aircraft) κ < ±3gon
3cm in 23cm images

•Navigation equipment to held the flying track, and the flying height

• overlap regulator (to mantain overlapping also in montainous terrains )

linked with automatic computation of ‘scatto’o ‘intervallo di scatto’

GPS/INS data : very important for navigation in poorly mapped regions

AA 2008 -2009 15
• can have a forward motion compensation device
• must have a central shutter so that exterior orientation is uniform over the whole
image This moreover allows to use shorter exposure time (1/150 ÷ 1/1000 s and less)
· The flash exposure of the fiducial marks must coincide with the mean moment of
exposure (mean time of shutter opening)
· The film must be kept flat with a vacuum pressure flattening device
· Must be built with materials having the same termal expansion coefficient.

Pilot - Co-pilot - Photographer

But now many operations are remote-controlled or automated, and the pilot alone
can operate the camera system.

Minimum interval between two successive shots : min cycle time 2 sec,this limits
the base length.
The camera can operate automatically, controlled by the overlapping regulator, or
manually.
There is a shot signal for the pilot to leave time to correct track, drift and attitude
before the next shot

AA 2008 -2009 16
 For checking drift and position are used:

• viewfinder (WILD)
• overlap control (ZEISS)
• special pointers remote-controlled by navigator-photographer

In the viewfinder the operator sees the apparent motion of the ground and:

Image frame (inquadratura)

Target point

Spherical level
Central line to check drift angle

Moving lines for overlap regulator

The lines should move at the same apparent

speed of the ground. In this case, the operator
control the preset overlap regulator along track.
Over mountainous areas apparent speed of motion can vary widely because of great
flying height variations. Moving lines extends over the full field of view.
AA 2008 -2009 17
(overlapping must be maintained in the highest parts)
The central line is used to check and adjust the drift angle, caused by wind velocity,
of the camera, by observing the apparent motion of a point close to this line.

The course is corrected by an angle

The aircraft flies along its nominal
k= -α, and the camera is rotated on
course, the side wind drives it along a
the aircraft of α . The real course thus
different course (drift α)
coincides with the planned direction
In the viewfinder a central point
and in the viewfinder a central point
doesn’t stay on the central line
lies on the central line
AA 2008 -2009 18
ANALOGUE AIRBORNE CAMERAS
CAMERE METRICHE AEREE

AA 2008 -2009 19
 Airborne camera RC30 (more)

Three objectives (8.8, 15, 30 cm).

Exposure intervals 1/100 s - 1/1000 s.
Max aperture f/4
Forward Motion Compensation
Film length 120m

Automatic flight planning, navigation and positioning systems with precise

positioning of projection centers.
ASCOT (Aerial Survey Control Tool)
Allows the automatic flight path following, giving to the pilot a real time control of
aircraft and camera positions respect to the planned course.

Directly gives center of projection coordinates to be used in A.T computations.

Taylored selection of imaged information (e.g. auxiliary and external data)

AA 2008 -2009 20
 RMK TOP (more)
Focal lenghts 15 cm e 30 cm.
Exposure intervals 1/50 s - 1/500 s
Forward Motion Compensation
Rubber vibration damper (better image
quality)

Possibility of flight control by GPS/INS system and of

precise projection centers positioning for successive A.T.
Can be totally and automatically controlled by the pilot
Taylored selection of imaged information (e.g. auxiliary
and external data).

Automatic control of drift and overlap

AA 2008 -2009 21
Will digital aerial cameras substitute analogue cameras ?

In the future, but analogue cameras are still widely

used :

World use of
analogue and
digital airborne
cameras

350

AA 2008 -2009 22
 INNER ORIENTATION DETERMINATION
A photogrammetric camera objective is an optical system realized combining different
kind of lenses. The deviation of its behaviour from an idealized model give rise to
various source of image distorsions

For an ideal optical system there are two principal planes that have peculiar properties:
principal planes H and H' (in object space and image space)

The intersections of H and H' with Idealized image formation geometry

the Optical Axis (OA) identify the
nodal points N and N'

• the ‘central rays’ passing

through the nodal points, form the
same angle τ with the optical axis
in the object space and in the
image space

N is the projection center in the

object-space and N′′ is the
projection center in the image-
space
AA 2008 -2009 23
INNER ORIENTATION

The optics of a metric camera is a thick asymmetric objective

Individual lenses are made of different types of glass to ensure the
greatest possible corrections of imaging errors

The aperture stop (diaframma), is not in the center of the objective, and limits
the effective bundle of rays forming the image points
entrance pupil EP
(and exit pupil in image space)

Problem : where is the projection center??

Optical axis

Aperture stop
Centers of EP and EP’ define the REAL projection centers, in
object and image spaces
AA 2008 -2009 24
INNER ORIENTATION
Real photogrammetric objective is very different from theoretical optical
model
REAL image
formation geometry

the optical axis OA

should contain the
centres of all lenses, but
small errors sum up

the reference optical axis is not OA, but a calibrated principal ray PRA,
(A= autocollimation) perpendicular to the image plane and passing through the
entrance [Link] O. Its extension through O’ intersects the image in the
‘principal point of autocollimation’ PPA

• τ′ is different from τ (because EP and EP′ do not coincide with H and H′)

• the mechanically realized principal distance s’ differs slightly from the optical
principal distance s’
AA 2008 -2009
• Image Plane is not rigorously perpendicular to the optical axis OA, nor to PR25A
 INNER ORIENTATION

MAIN EFFECT : angle τ′ is different from τ

Correction: define a mathematical projection center O’M wich lies at a perpendicular
distance c from PPA that reproduces the angles τ as closely as possible

In any case there exists a difference in position from real image point P’
and the theoretical position.

The distance is a residual error called optical distorsion .

We distinguish two components, one radial ∆ρ and the other ∆τ tangential,

perpendicular to the first one (radial, & tangential distortion).

The tangential distortion (often called ‘decentering distorsion’) is due to non

perfect centering of all the lenses forming the camera lens.

Usually the radial distortion ∆ρ is much bigger than the tangential one ∆τ,
that is normally around 5%.

AA 2008 -2009 26
INNER ORIENTATION

AA 2008 -2009 27
INNER ORIENTATION

DECENTERING DISTORSION
In high precision surveys the decentering or tangential distorsion must be taken
into account, that derives from a non perfect centering of the lenses composing
the optical system.

The effect is that radial straight lines become curves

AA 2008 -2009 28
INNER ORIENTATION

The following relation holds between radial distance ρ of a

point P’ respect to PPA and radial distorsion ∆ρ (in P’):
inner orientation equation
ρ = c tgτ + ∆ρ
ctgτ = ρ teor
AA 2008 -2009 29
 Camera calibration

Radial distorsion curves are measured during calibration procedures in

a laboratory using a ‘photo-goniometre’

For fixed τ, the corresponding ρ’ = ρobs = ρteor + ∆ρ

are measured, using a calibrated grid (0.1 µm)
AA 2008 -2009 30
A complete calibration procedure consists of determining the position
of PPA and the radial distorsion along the four semi-diagonals of the
photogram. (To take care of asymmetry) ctgτ = ρ teor
The differences ρ oss − ctgτ = ∆ρ are the optical radial distorsions

Usually these four curves are

non identical.

The asymmetry is greatly

reduced by choosing
another axis as reference:
the principal axis of
symmetry PRS.

PRS pass through O and

intersects the image plane
IP in the Point of Best
Symmetry
PBS
AA 2008 -2009 31
The reference optical axis becomes PRS in place of PRA and:
• (ξ , η ) refers to PBS
• the new projection center is O′M , and is at distance c from PBS
• radial distorsion curves are referred to PBS

In order to model and reduce as better as

possible the radial distorsions, during the
calibration procedure a new (calibrated!)
principal distance c’=c+ ∆c is computed, so as
the distorsion curves are as near as possible
to the ρ axis (nearly zero values)

In fact, the effect of varying the

principal distance is to rotate the
curves around the origin.
AA 2008 -2009 32
• Older photogrammetric objectives have mean radial distorsion up to ∆ρ ≤ 30 µm

• Modern objectives are nearly distorsion-free ∆ρ ≤ 2 µm

So small distorsions are corrected only in high precision photogrammetry, taking into
account both radial and tangential distorsion, together with other photogram
deformations.

• How radial distorsion curve is used:

given (measured) ξ and η, it is possible to compute ρ: ρ= (ξ − ξ 0 )2 + (η − η0 )2

ρ teor = ρ corr = ρ oss − ∆ρ

then ∆ρ can be deduced from distorsion curve
∆ρ = ρ oss − c tan τ = ρ oss − ρ teor

∆ξ =
(ξ − ξ0 ) ∆ρ ∆η =
(η − η0 ) ∆ρ
∆ρ is decomposed into the components ∆ξ,
∆ξ ∆η:
∆η ρ ρ

These are corrections to be added to image ξ corr = ξ oss − ∆ξ

measured coordinates ξ and η. ηcorr = ηoss − ∆η
AA 2008 -2009 33
• Modello di Fraser, 1999, 10 parametri:
• c, X0,Y0: parametri di O.I. standard
• K1, K2, K3: distorsione radiale
• P1, P2: distorsione tangenziale
• b1,b2: distorsione affine

ξ = Dξ - c NX / D η = D η - c NY / D

• Dx= X0 + x (K1 r2 + K2 r4 + K3 r6 ) +
• + P1(r2 + 2x2) + 2 P2 x y + b1 x + b2 y

• D η = Y0 + η (K1 r2 + K2 r4 + K3 r6 ) +
• + P2(r2 + 2y2) + 2 P1 ξ η

AA 2008 -2009 34
 CALIBRATION CERTIFICATE
For the physical definition of the image coordinate system fiducial marks are used,
imaged in every photograph, that define the fiducial centre FC.
But the image coordinates have to be referred to PPA or PBS.

Photogrammetric cameras commonly have PPA, PBS and FC at distances smaller

than 20 µm.
max = 20 µm
• PPA
• FC
• PBS

The calibration certificate of a photogrammetric camera contains :

· image coordinates of fiducial marks;

· image coordinates of FC, PPA and/or PBS;
· principal distance c;
· the mean radial distorsion curve;
· date of calibration;
· information on image sharpness.

In Italy the calibration certificate has to be renewed every 2-3 years
AA 2008 -2009 35
AA 2008 -2009 36
 Camera calibration
Radial distorsion can be modelled through a polynomial curve

Photogrammetric sw accept curves given in a discretized tables ∆ρ(ρ), or

in terms of polynomial coefficients : K0, K1…..(also for semi-metric
cameras)
3 5
∆ρ = K 0 ρ oss + K1 ρ oss + K 2 ρ oss + ....

( 2
∆ρ = A1 ρ oss ρ oss ) (
− R02 + A2 ρ oss ρ oss
4
)
− R04 + ....

The film magazine should be included in the calibration too

For high precision works it is advisable to take calibratiion

photographs under the same flying condition, before and/or
after normal photographic survey, over a test field with
signalized points.

AA 2008 -2009 37
 Camera calibration

Cameras for close range are

also often calibrated on a
spatial test field, computing all
parameters of inner and outer
orientation together.
(self-calibration)

Test field

AA 2008 -2009 38
AA 2008 -2009 39