GE312 Lecture 7:

Orientation of Stereopair
Engr. Mark Dave B. Plaza
Instructor, GE Dept., CEA, USTP-CDO
Outline

▪ Introduction to stereopair orientation

▪ Interior orientation
▪ Exterior orientation
▪ Stereopair Orientation

2
Intended Learning Outcomes
After the Lecture and Laboratory the students will be able to:

▪ Identify concepts pertaining to orientation of Stereopairs

▪ Explain the processes involved in photo interior orientation
▪ Explain the processes involved in exterior orientation
▪ Explain the processes involved in stereopair orientation

3
Introduction to Stereopair
Orientation
Lecture 6.1

4
Model Space and Coordinate System

Model Space – is the virtual

representation of space (3D) of
objects in ground. It is also the
space established by two images.
establishment of model coordinate
system can be establish through
series of orientation.

Source: *dji.com

5
Model Space and Coordinate System

Model Coordinate System – is the

reference system established to
determine the corresponding
position of objects on ground in the
model space.
establishment of model coordinate
system can be establish through
series of orientation.
Source: *dji.com

6
 Model Space and Coordinate System
y

Two Coordinate Systems Z

x
involved in the model space:
1. Image/Photo Coordinate
System Y

2. Object Space Coordinate

System

Source: *Lilesand and Kieffer (2015) 7

What is Orientation?
Alignment?
- being in one and same line?
Coincidence of direction?
- having one and same direction?
Determination of direction and position in reference with
a standard?
- being the direction and position known in relationship to a
certain reference standard.
8
What is Orientation?
Orientation is the establishment of
the relationship between the
objects position and identified
standard reference frame. 𝑦 = 𝑚𝑥 + 𝑏

In mathematics relationship can be

established using equations.

𝐿𝑖𝑛𝑒𝑎𝑟 𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛

9
Equation and Parameters
Standard Equations – are
established equations that
represents specific relationships.
Ex. Linear - 𝑦 = 𝑚𝑥 + 𝑏 𝑦 = 𝑚𝑥 + 𝑏

Quadratic - 𝑦 = 𝑎𝑥 2 + 𝑏𝑥 + 𝑐
Parameters – are the coefficients in
the equation
Ex. m and b in linear equation
a, b, and c in quadratic equation 𝐿𝑖𝑛𝑒𝑎𝑟 𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛

10
Review of Workflow of Photogrammetry
Interior Orientation – necessary only
for analog photos to be digitally
scanned.
-Images acquired digitally doesn’t need
an interior orientation

Exterior Orientation – necessary only

for images without corrected positioning
and/or angular rotation values.
-Images that has been acquired with
known positioning and angular rotation
values including those have been bound
with GNSS for Positioning and IMU for
rotation parameters.

11
Review of Workflow of Photogrammetry

Stereopair Orientation –
establishment of model
coordinate system before
determination of Model

12
Orientations involved in stereopair

1. Interior Orientation - is the establishment of relationship between

the camera internal positioning and digital image positioning in
computer’s monitor of a single photo.
2. Exterior Orientation - is the establishment of relationship between
the camera internal positioning or digital image positioning and
ground positioning of a single photo.
3. Stereopair or model Orientation - is the establishment of
relationship between the orientation of two or more photos on
ground. It is the development of object model space coordinate
system.

13
Interior Orientation
Lecture 6.2

14
Camera internal positioning y

1. Camera internal
positioning is
defined by camera-
internal co-ordinate
system and guided x
by Fiducial Marks

15
Digital Image positioning

2. Digital image positioning

is based on pixel
coordinate system
and wherein its origin
is in the top left corner
of the image.
Coordinates are always
positive.

+y 16
Affine transformation

Affine transformation is a linear mapping method that

preserves points, straight lines, and planes*.
Translation Rotation Scaling Mirroring

Source: *mathworks.com 17
Affine transformation

Affine transformation is a linear mapping method that

preserves points, straight lines, and planes*.
P
Similarity
Shearing

Homothety
Source: *mathworks.com 18
 Affine transformation

Affine transformation establishes the relationship between the

camera internal coordinate system (x,y) and pixel coordinate
system (x’,y’) or vice versa in 1st degree equation.

𝒙′ = 𝒂𝟎 + 𝒂𝟏 𝒙 + 𝒂𝟐 𝒚
𝒚′ = 𝒃𝟎 + 𝒃𝟏 𝒙 + 𝒃𝟐 𝒚

𝑎0 , 𝑎1 , 𝑎2 , 𝑏0 , 𝑏1 , 𝑎𝑛𝑑 𝑏2 are the transformation parameters

This requires at least 3 points (e.g. fiducial marks)
19
2nd Order Polynomial Warp

2nd Order Polynomial Warp establishes the relationship

between the camera internal coordinate system (x,y) and
pixel coordinate system (x’,y’) in 2nd degree equation.

𝑥 ′ = 𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑦 + 𝑎3 𝑥 2 + 𝑎4 𝑥𝑦 + 𝑎5 𝑦 2
𝑦 ′ = 𝑏0 + 𝑏1 𝑥 + 𝑏2 𝑦 + 𝑏3 𝑥 2 + 𝑏4 𝑥𝑦 + 𝑏5 𝑦 2

This requires at least 6 points (e.g. fiducial marks)

20
High Order Polynomial Warp

3rd Order Polynomial Warp

4th Order Polynomial Warp
.
.
.
.
Source: *esri.com

21
Exterior Orientation
Lecture 6.3

22
 Exterior Orientation

Exterior Orientation – in a single

photo it is the establishment of
relationship between the
-camera internal positioning or digital
image positioning
-ground positioning.

Source: *Lilesand and Kieffer (2015) 23

 Image/Photo and Ground Plane

Non-parallelism of Image and

Ground Plane.
Angular Rotation or Attitude is the
evidence of non parallelism of
image/photo and ground plane.
in every major axis:
-roll (x-axis) 𝜔
-pitch (y-axis) 𝜙
-yaw (z-axis) 𝜅

Source: *Lilesand and Kieffer (2015) 24

 Collinearity Principle

Collinearity Condition –
The exposure point (L),
point in the image (p), and
point on ground (P)

lies in straight line.

Source: *Lilesand and Kieffer (2015) 25

 Collinearity Principle

Consideration of Collinearity
1. Object Image/photo position (𝑥𝑝 , 𝑦𝑝 )
2. Object Ground position (𝑋𝑃 , 𝑌𝑃 , 𝑍𝑃 )
3. Exposure Point Ground Position
(𝑋𝐿 , 𝑌𝐿 , 𝑍𝐿 )
4. Angular Motion (𝜔, 𝜙, 𝜅)

Source: *Lilesand and Kieffer (2015) 26

 Collinearity Principle

Collinearity Equation – mathematical

expression of the collinearity condition.
Where,
𝑓 is the focal length
𝑥𝑝 , 𝑦𝑝 is the object image/photo
coordinates
𝑋𝑃 , 𝑌𝑃 , 𝑍𝑃 is the object ground
coordinates In a single photo, image/photo coordinates is known
𝑋𝐿 , 𝑌𝐿 , 𝑍𝐿 is the exposure point ground thus the unknowns are the following 9 coordinates:
coordinates 1. 𝑋𝑃 , 𝑌𝑃 , 𝑍𝑃
2. 𝑋𝐿 , 𝑌𝐿 , 𝑍𝐿
𝑚11 , … , 𝑚33 is the coefficient of 3 x 3 rotation 3. 𝜔, 𝜙, 𝜅
matrix

Source: *Lilesand and Kieffer (2015) 27

 Collinearity and Exterior orientation

The purpose of Collinearity Equation is to establish the relationship

between object’s image/photo coordinate system (𝒙𝒑 , 𝒚𝒑 ) and ground
coordinate system (𝑿𝑷 , 𝒀𝑷 , 𝒁𝑷 ).

Thus the only remaining unknowns to be determine are the ground

coordinates of L (𝑿𝑳 , 𝒀𝑳 , 𝒁𝑳 ) and the angular rotations (𝝎, 𝝓, 𝜿) these
are the exterior orientation parameters.
How can 𝑿𝑳 , 𝒀𝑳 , 𝒁𝑳 , 𝝎, 𝝓, 𝒂𝒏𝒅 𝜿 be determined?
It can be determine with the use of Ground Control Points or GCPs.

Source: *Lilesand and Kieffer (2015) 28

 Ground Control Points

Ground Control Points (GCPs)

- are special and specific point objects on
ground that are expected to be visible on
photos, which serves as a reference points
for model orientation.
-Ground coordinates can be determine
using conventional surveying techniques or
using GNSS.
-Image/photo coordinates may be extracted
from the image/photos during
photogrammetric process.

Source: *Lilesand and Kieffer (2015) 29

 Ground Control Points

Source: *Lilesand and Kieffer (2015) 30

 Resection in Space

Resection in Space PP
Z’

-the method of determination of Y’

2
3
the position of L and angular Z 1
X’
rotation of photo/image plane in
the ground space by intersecting
the collinear object points in
image and ground. Y

𝑋𝐿 , 𝑌𝐿 , 𝑍𝐿 , 𝜔, 𝜙, 𝜅 are the unknowns, which are 6.

Thus it requires at least 6 equation to be known. GCP 2
GCP 1 GCP 3
Each GCPs can create 2 equation thus it needs at
least 3 GCPs in a single image for exterior
orientation. X

Source: *Lilesand and Kieffer (2015) 31

 Application of 3-D Conformal Transformation
Computation

in application of 3-D Conformal transformation, the

corresponding parameters has been utilized as PP
Z’
shown in the table below. 2
Y’ 3
Z 1
3-D Conformal Parameters EO Parameters
X’
Output Coordinates (X,Y,Z) Image Coordinates (𝑥𝑝 , 𝑦𝑝 , 𝑓)
Input Coordinates (x,y,z) Diff. of GCP (P) and PP (L)
Coordinates (𝑋𝑃 − 𝑋𝐿 ,
𝑌𝑃 −𝑌𝐿 , 𝑍𝑃 − 𝑍𝐿 )
Y
Scale F over Distance PL
Rotation Parameters Rotation Parameters (𝜔, 𝜙, 𝜅)
(𝜃1 , 𝜃2 , 𝜃3 ) or r or m GCP 2
GCP 1 GCP 3
Translation Parameters Zero or (0,0,0)
(Tx,Ty,Tz) Origin is equivalent to PP
X

Source: *Lilesand and Kieffer (2015) 32

 Application of 3-D Conformal Transformation
Computation

Assignment: PP
Z’
2
Y’ 3
Derive collinearity equation using 3D- Z 1

Conformal Transformation. X’

1. Derivation (10 points)

2. Identification of m parameter values
(10 points) Y

GCP 2
GCP 1 GCP 3

Source: *Lilesand and Kieffer (2015) 33

Stereopair Orientation
Lecture 6.3

34
 Determination of Ground Coordinates in
Single Photo

After exterior orientation the exposure station ground coordinates

(𝑿𝑳 , 𝒀𝑳 , 𝒁𝑳 ) and attitude (𝜔, 𝜙, 𝜅) of single photo.
In a single photo you only have 2 image/photo coordinates that can be
extracted: 𝑥𝑝 , 𝑦𝑝 or 𝑥𝑝′ , 𝑦𝑝′ .
The objective of photogrammetry is to determine the ground
coordinates of the points of objects in the photo or the 𝑿𝒑 , 𝒀𝒑 , 𝒁𝒑 .
By using the collinearity equations and the given parameters, we can
only produce 2 equations, one for 𝑥𝑝 and one for 𝑦𝑝 .
With 3 unknowns and 2 equations we won’t be able to extract exact
values.
Source: *Lilesand and Kieffer (2015) 35
 Stereopair and Determination of Ground
Coordinates

Therefore we need another photo containing the same objects with

known exterior orientation parameters.

Together with the first photo they will create a stereopair or model.

With stereopair we can now have 4 equations for 3 unknowns. Which

is solvable using least squares method.

Source: *Lilesand and Kieffer (2015) 36

 Purpose of Stereopair Orientation

But before determination of points

ground coordinates, the established
exterior orientation parameters of the
two photos must be coincided in one
coordinate system.
This process is called the stereopair
orientation.
Which will produce the model
coordinate system.

model coordinate system

Source: *Schenk, T., 2005 37

 Creation of model space

After stereopair orientation it will create

a single space,

Which is now known as model space.

This is where the common points in the

two image will be established.

model space

Source: *Schenk, T., 2005 38

 Methods of Steropair Orientation

1. Dependent Relative Orientation

2. Independent Relative Orientation
3. Direct Orientation
4. Absolute Orientation

Source: *Schenk, T., 2005 39

 Dependent Relative Orientation

The position and the orientation is

identical to one of the two photo-
coordinate systems, say the
primed system.
Example: 1st photo
2nd photo
we fixed the ground position to the
orientation of the first photo and the
orientation of the second photo
relates to it.

Source: *Schenk, T., 2005 40

 Dependent Relative Orientation

The origin of the coordinate system is

based on the central perspective of the
first photo (c’) or the ground
coordinates of L and its orientation
parameters. 1st photo
2nd photo
The relationship with the second photo
is established by the base components
(𝒃𝒚 , 𝒃𝒛 ) and rotational angles
(𝝎",ϕ", 𝜿")
𝑏𝑥 has been extracted in the
parameters since it coincides with xm
Source: *Schenk, T., 2005 41
 Independent Relative Orientation

Only the origin is identical to one photo

coordinate system.
The orientation is chosen such that the
positive xm-axis passes through the
perspective center of the other photo- 1st photo
2nd photo
coordinate system.
This requires determining two rotation
angles in the primed photo-coordinate
system.

Source: *Schenk, T., 2005 42

 Independent Relative Orientation

Moreover, it eliminates the base

components 𝒃𝒚 , 𝒃𝒛 .
The rotation about the x-axis (𝝎) is set
to zero.
1st photo
Thus the remaining parameters will 2nd photo
be:
ϕ′, 𝜿′, 𝝎",ϕ", 𝜿"

Source: *Schenk, T., 2005 43

 Direct Orientation

the model coordinate system

becomes identical with the
ground system (ex. PPCS-PTM)
Points with known coordinates are
called control points. A point with
all three coordinates known is
called full control point.
If only X and Y is known then we
have a planimetric control point.

Source: *Schenk, T., 2005 44

 Absolute Orientation

Uses the Helmert 7-parameter

orientation, consists of:

3 Rotation parameters
3 Translation parameters
1 Scaling parameter

Source: *Schenk, T., 2005 45

 Absolute Orientation

Matrix Equation:
𝒑 = 𝑠𝑹𝒑𝒎 − 𝒕
Where,
𝒑 is the ground coordinates matrix of P
𝑠 is the scaling parameter
𝒑𝒎 is the point coordinates in the
model coordinate system
𝒕 is the translation parameter matrix

Source: *Schenk, T., 2005 46

 Absolute Orientation

When solving using Least Squares to

determine the parameters:

𝒓𝒊 = 𝑠𝑹𝒑𝒎 − 𝒕 − 𝒑𝒊
Where,
𝒑i is the ground coordinates matrix of
each P
𝒓𝒊 is the residual

Source: *Schenk, T., 2005 47

Seatwork

Prepare 1 whole sheet of yellow paper.

10 items Identification
2 items essay (5 points each)

48
References
1. Lillesand, Thomas M., 2015, Remote sensing and image
interpretation / Thomas M. Lillesand, Ralph W. Kiefer, Jonathan
W. Chipman.—Seventh edition.
2. Schenk, T., 2005, Introduction to Photogrammetry
3. Online resources: Youtube videos

49
THANK YOU

50