DDPS Users Guide

background image

1















(DDPS)


DIDACTIC AND DIGITAL

PHOTOGRAMMETRIC SOFTWARE

User’s Guide























background image

2




Table of contents

INTRODUCTION ........................................................................................................ 3

POSSIBILITIES AND FUNCTIONNALITIES OF THE SOFTWARE .......................... 4

DATA PROCESSING ................................................................................................. 8

STEP BY STEP EXAMPLE ...................................................................................... 25

ANNEXES ................................................................................................................ 39

background image

3

INTRODUCTION




This program was developed within the framework of a cooperative project between
the SURFACES Laboratory, Department of Geomatics, University of Liege and the
Institute of Geodesy and Cartography (IGiK), Department of Photogrammetry ,
Poland.

Our main objective is to produce a complete and integrated educational
photogrammetric package. The graphic interface was made with a didactical
objective, rapid as possible and easy to understand. The final result is a very user-
friendly interface. No special knowledge is needed (except in photogrammetry) for
using the software.


background image

4

POSSIBILITIES AND FUNCTIONNALITIES OF THE SOFTWARE


1. REQUIREMENTS AND POSSIBILITIES

1.1.Requirements

Digital Photogrammetric Software

processes digital images with 8 bits/pixel (grey

level). File format is BMP.

The images must be photographs taken with a metric camera for that characteristics
are known. So the user must have at one's disposal the calibration certificate of the
camera.

To enable the absolute orientation of the photos, a sufficient number of control points
discernible on the two photos must be known.


1.2.Possibilities

A complete photogrammetric treatment is possible with Digital Photogrammetric
Sofware. The following general workflow can be use to realise it:

background image

5



INNER

ORIENTATION

Pixel coordinates => Photo coordinates


Calibration Certificate

Left Image

Right Image

EXTERIOR ORIENTATION

RELATIVE ORIENTATION ABSOLUTE ORIENTATION

Photo coordinates => Model coordinates Model coordinates => Ground coordinates

Homologous points

Control points

EPIPOLAR

RESAMPLING

Determination of the bounding box

IMAGE MATCHING

GENERATION OF THE

DEM

ORTHORECTIFICATION

VRML MODEL GENERATION

background image

6


2. TO MOVE IN THE IMAGES

The views of the different steps of the photogrammetric orientation (inner, relative
and absolute orientation) are organised in three different views of the same image.




1

1

3

3

1





2

3

2

2




The view 1 is the over view that shows the entire image.
The view 2 corresponds to the original size of the image and is the part of the view 1
included in the white rectangle.
The view 3 is a detail view that shows the part of the image within the white rectangle
of the view 2.

When you want to track down a point in the image, place the cursor in the view 1
near by this point and click on the right mouse button. The white rectangle is then
centred on the selected point and the part of the image included inside appears in the
view 2. Next do the same thing with the view 2 to locate more accurately the point.

3. TO TRACK DOWN POINTS IN THE IMAGES

Before tracking down a point in the image, you must select the line of the current
point in the chart below the image. Then the cursor becomes a hair cross.


Points can

be selected in all the three views. To be more accurate, it’s better to

locate it in the view 3 (refer at 2: to move in the images)
When you have adjusted the view on the point, place the centre of the hair cross on
this. Then click two times on the left mouse button to select the point. The image
coordinates appear on the correspondent line of the chart below the image. . When
the hair cross isn’t just on the point, it can be moved by holding the left mouse button
and move the mouse.

INNER ORIENTATION

RELATIVE AND ABSOLUTE

ORIENTATION

background image

7

4. TO MOVE IN THE CHARTS

Some additional functions have been added to make easier the displacements in the
relative and absolute orientations charts.
After the selection of the current line (by clicking on this in the chart with the left
mouse button), you can move from line to line or laterally with the help of the
direction arrows of the keyboard.

5. TO MAKE A ZOOM

It’s possible to increase or decrease the size of the white rectangle in the views 1 and
2, modifying in this way respectively the views 2 and 3.
After the selection of the view within you want to make the zoom, click concurrently
on Alt and * to decrease the zoom and on Alt and / to increase it. Then the view of
the superior level will automatically change.

6. THE THREADS

Sometimes, by mistake or inexperience, the user can choose parameters that lead to
process the data for long time. In this case, it’s useful to know how many times the
process will take.
The thread is useful for estimating the duration of a specific task.


Another interesting functionality is that a process putted in a thread can be
interrupted at each time the user wants.

background image

8

DATA PROCESSING



1.INNER ORIENTATION

An inner orientation must be done for every image. This process allows establishing
a relation between the pixel and image coordinates. This relation is determined with
the help of the fiducial marks photo coordinates given in the calibration certificate
(see annex 1).

To change the pixel coordinates into photo coordinates, you must encode the values
of the photo coordinates of the fiducial marks. Then a parallel between these values
and the pixel coordinates will can be drawn.

Encode the calibration certificate

A. The camera has never been encoded.

In the project window, choose the option New CC. The following dialog box will pop
up:



The first thing to do is the choice of the units, that will be the units used in all the
project. For more facilities, the units can be these of the calibration certificate. With

background image

9

the help of the arrows on the left of the case called Unit, you can increase or
decrease the value of the units. When you have determined it, you must think that all
the values will be expressed in these units.

The following step is the encoding of the values that you have in the calibration
certificate:

- f focal length: the value must be written in the Focale case

- PP position of the principal point:

The value of

0

corresponds to the E

0

case

The value of

0

corresponds to the N

0

case

- position of the fiducial marks:

The superior part of the left window is devoted to receive the photo coordinates of
the fiducial marks. By clicking on the arrows below the chart can you determinate
the number of fiducial marks. Next, the ID, X Coord and Y Coord (corresponding
to the photo coordinates) of every fiducial marks can be encoded.


- lens distortion:

The superior part of the right window is devoted to receive the values of the lens
distortion. You can also determinate the number of radial distance and mean
radial distortion values by clicking on the arrows on the left of the case. Then the
chart can be completed with the radial distances in the left column and the mean
radial distortions in the right column.

All these informations can be saved in a “.cc” file with the option save CC or by
answering YES at the

question “Do you want to SAVE Calibration Certificate?” when

you quit.


B. The camera is the same that these that has already been encoded for precedent
photos.

Then the calibration certificate previously encoded can be got back. Click on Load
CC
in the CC File Path case and take the appropriate file.

Location of fiducial marks


Now that the photo coordinates of the fiducial marks are known, the parameters of
the transformation of the pixel coordinates will be computed. Those will depend of
two things: the pixel coordinates values of the fiducial marks and the orientation of
the X and Y axes.

In the project window, choose the option Left or Right IO to make the inner
orientation of every one. The following window will appear:

background image

10


There are three views of the same image. The right superior view is the over view
that shows the entire image. The left view corresponds to the original size of the
image. The right inferior view is a detail view that shows the part of the image within
the cursor of the left view.

The first step is the determination of the photo orientation. According to it, the
parameters of transformation will change. You have the choice between four cases.


This step will transform the start system in the ideal case (see the theory about this
part)

In the chart below the image, you find in the two first columns the values of the photo
coordinates given in the calibration certificate. The pixel coordinates are directly
registered in the XImg and YImg columns when you track down the fiducial marks on
every image. To move and locate points on the image, see the section
Functionnalities of the software.

background image

11

Parameters of transformation will appear as from three fiducial marks and residues
as from four fiducial marks.

Correction of the radial lens distortion


In the menu Inner orientation module, choose the Radial Distortion option. The
window in that the parameters of the polynome are computed will pop up:



In this window, the values of the radial distances and of the mean radial distortions
previously encoded will automatically appear. You have the possibility to choose the
polynomial order but it must be less than the number of associated radial distances
and mean radial distortions. The you can compute the polynomial, in the third
column appear the residues.

2. EXTERIOR ORIENTATION

2.1. Relative orientation by rotations

Relative orientation allows determining the relative positions of the two bundles of
rays by creating spatial stereomodel in an arbitrary coordinate system. Tracking
down homologous points on the left and right images is necessary to make this
orientation.

In the project window, click on the button RO by Rotations. The following window
box will appear:

background image

12


By clicking with the right mouse button on a grey case of the chart under the left
image, you can choose to add a point with the option Add Point. Then you must
track down the homologous point on the left and right images: The left and right
image coordinates of the point will appear in the current line of the chart

It is possible to delete one point by clicking on the ID with the right mouse button and
by choosing the option Delete This Point.

As from five homologous points, it is possible to have values of RO Final Parameters
(expressed in Radians) by clicking on Reset the first time and after Iterate. Then
values of the chart will change: the parallax will decrease, new RO Final Parameters
are computed and new model coordinates (X

i

Coord, Y

i

Coord) are computed for every

image. Stop the iterations when you think that the parallax values are small enough
and that RO Final parameters don’t significantly change.





background image

13

2.2. Absolute orientation

The purpose of absolute orientation is the transformation from model coordinates to
ground coordinates. To realise it, you must have points that ground coordinates are
known. These points are called the control points (CTRL) and will serve to the
computing of the parameters allowing the transformation from model coordinates to
ground coordinates.

Another type of points is the checkpoint (CHECK). It corresponds to points that
ground coordinates are known but that are not used for the computing of the absolute
orientation parameters. The residues on these points give an idea on the global
accuracy.

Last type is the tie point (TIE). It corresponds to points that ground coordinates are
unknown and for that ground coordinates are computed with the absolute orientation
parameters.

In the project window, click on AO & Restitution. Then appears the window of the
absolute orientation

As for the relative orientation, you can add or delete points by clicking on the right
mouse button when you are in the chart.. The following window will appear:


This window allows the encoding of the control, the check and the tie points. In the
point class, write ctrl for control points, check for checkpoints and tie for tie points.
When ground coordinates of the current point are known, they can be encoded in this
chart. Coordinates can also be encoded in the general chart below the image by
selecting the correspondent case.

When points are located, control points appear in blue, checkpoints in green and tie
points in red. As for three control points, absolute orientation parameters can be
computed by clicking on Reset and then Iterate.

Stop the iterations when the residues on the control and checkpoints are small
enough (in the window, the residues are expressed in the same units that the units of
the check and control points). The values of the absolute orientation parameters are
in the grey window.

background image

14



The value of the base has no importance; it will only change the value of the scale
factor.

The angle κ must sometimes be changed. Indeed, the first value of this angle is
calculated with the two first control points. In accordance with the relative position of
these points, you obtain the value of the κ-angle or of (π-κ). If the computed value is
(π-κ), residues will increase. Either you inverse the two first control points or just
after clicking Reset, change the given value by (π-κ) and then Iterate.

background image

15

3. RESTITUTION

3.1. Monoscopic Restitution

In the project window, click on AO & Restitution - menu Module option Application.

This module allows to track down additional points for that ground coordinates are
unknown and to compute their ground coordinates with the absolute orientation
parameters. For the location of the homologous points, proceed in the same way
that for the others modules. The ground coordinates of the new points will directly
appear. These points will receive the type resti points.




3.2. Geo

– Coding


Create a Configuration File (.cl) that has the following structure (ex: tess.cl):


BEGIN_CODES
point
line
spline
END_CODES
BEGIN_LAYERS
road
hedge
fence
END_LAYERS

Then you choose for every point its code and its layer. A file with the ID, the type, the
layer, the code and the coordinates of the points is created. This file can directly be
employed in topographic programs.


you put here all the layers that you want to use

you put here all the codes that you want to use

background image

16


3.3. Image Matching techniques

In the project window, click on AO & Restitution - menu Configuration option Auto
Find

– ON

This module allows the automatic detection on the right image of the homologous
point tracked down on the left image.

In this program, an area based matching is used to identify the homologous points.
This method works with pattern, target and search window. Points are associated if
they have similar grey values. Before starting, you must configure the parameters of
the matching. The parameters of the matching that you can change are:

The size of the pattern and target window: there are the windows surrounding the
matched pixel

The size of the search window: the search window contains the matched pixels for
which correlation coefficients are computed

The minimum correlation value: if this value isn’t reached, the window “not
correlation” will appear and the cursor will be on an approximate position of the
homologous point.

The maximum of mean difference


When you have chosen to activate the auto-correlation, the homologous point on the
right image of the left image current point will automatically appear in the right
window. Sometimes the program can’t find the homologous point and then sends the
message “No Correlation”. Sometimes the point is not exactly the homologous of this

background image

17

of the left image. In these two cases, you have always the possibility to displace the
point on the two images or to delete it. When the two points are tracked down, their
ground coordinates are computed in accordance with the absolute orientation
parameters.


4. EPIPOLAR RESAMPLING

The resolution of the created image can also be chosen. This parameter will
determinate its size in Ko.

Empty images are then created. Every point in these images undergoes inverse
transformation to catch in the original image its grey value.

Epipolar images have only horizontal parallaxes. These images offer the big
advantage to simplify the automatic correlation procedure, because it works in a
single direction.


4.1. Application

In the project window, click on Epipolar Images. The following dialog box will pop
up:



Load the necessary files.

Click on Compute Bounding Box to have the limits of this. These limits can be
changed. The X-limits depend on the overlay. The user can directly change the Y-
limits.

background image

18

The options are:

 you can choose to use the relative orientation parameters or put other values

for dω

2

, dφ

1

, dφ

2

, dκ

1

, and dκ

2

. Indeed, this module is completely independent

of the rest of the treatment and can be used with values found by another
program.

 the overlay that will change the x min value of the left image and the x max

value of the right image

 the resolution that will determinate the size of the created epipolar image (the

size appear above the y min and y max values).


When all these parameters are fixed, click on Process. You must directly give a
name and a File Path for the created image.

5. IMAGE MATCHING

5.1. Application

In the project window, click on Matching:


First choose the two images on which the matching is made. Indeed, image
matching can be made on the original images or on the epipolar images. In
accordance with the chosen images, the parameters of configuration can differ.
Click on Generate in the Matched Hmlg Points part. A dialog box allowing the
configuration of the matching will appear:

background image

19

When you click on Configure Correlation Process, the same window that in the
automatic detection of homologous points will appear (see the absolute orientation to
configure it).

The superior part of the window contains various parameters:

 Size of the cell: The chosen images are broken down into several cells in that

homologous points are searched.

 Step in the cell is the number of points that it increments when homologous points

are found

 Matched Points Nbr is the number of matched points that must be find in each

cell. As soon as this number is reached, points are computed in the following cell.

These three parameters must be chosen in accordance with the desired number of
matched points.

 Overlay









Original image (15*15)

Matched point

Non-matched point

Example :

Size of the cell X :5
Size of the cell Y :5
Step in the cell X :2
Step in the cell Y :3
Matched Points Nbr : 3

background image

20

For the predicted position of homologous points, there are two options:

 Loading a homolog file; the predicted position are calculated in accordance with

the relative positions of the previously tracked down points.


 Computing an average shift X and Y values in pixels. The shift values are given

by the respective differences between the X and Y image coordinates of two
homologous points. For one computed point in the left image, the homologous
point on the right image will be search in a zone displaced of the shift X and shift
Y values compared with the position on the left image.


When all these parameters are fixed, click on Process Matching. A window informs
you on the course of the.process:


The under graph shows the cells within homologous points are searched. If the cell
remains green, then a sufficient number of homologous points for that the correlation
coefficient is sufficient have been found. If the cell remains red, then a number
inferior to the number asked by the user have been found.

When the matching is finished, it is possible to visualize the matched points by
clicking Visualize. This module allows deleting eventual non-homologous points. To
delete these points, click on its ID and press Del.

N.B: It is also possible to visualize a file with already matched points by clicking on
Find and by choosing the adequate file.

background image

21


6. GENERATION OF THE DIGITAL ELEVATION MODEL (DEM)

6.1. Application

In the project window, click on Matching. The following window will appear:



For generating a DEM, you need a file containing ground coordinates. To have this
file, either load an already OAApply file by clicking on Find or generate a new file
with Generate. This new file is generated as from an hmlg file that must be load in
the Matched Hmlg Points case.

The hmlg file can be constituted of homologous points tracked down on the epipolar
images

– then mark the Epipolar Img case – or of homologous points tracked down

on the original images:

 For the epipolar images, you must load the cc and OAparam files.
 For the original images, you must load the Left and Right OIparam, cc,

ORparam and OAparam files.

When the OAApply file is loaded or generated, you can click on Generate in the DEM
part. The DEM config window will appear:

background image

22

The predicted values that are proposed correspond respectively to:

 degree of elevation is 1. So the distance is elevated at the degree 1. The fact of

increasing the degree of elevation allows to reduce the importance of the points
that are more far of the evaluated point.

 the number of points by dial is 4. This number depends of the already determined

points and their distribution.


 the bounding box is the framework built on the points the ground coordinates of

which are known. The user can change the area of the DEM if he wants to make
the DEM only within a part of the original image. He musts then enter the ground
coordinates of the X and Y minimum and maximum of the desired area. It is also
possible to make a DEM on a larger area but the points, which are out of the
framework, will be less accurate because they have

n’t known points in each

direction surrounding them.

the mean of the distance between each point known in the ground coordinates

and its nearest neighbour. This distance can be increased to accelerate the
procedure if few points are desired. It can be also reduced if the DEM must be
more accurate.


background image

23

7. ORTHORECTIFICATION

7.1. Application

In the project window, click on Ortho:




To construct the orthophotograph, dem file, epipolar images, cc and OAparam files
must be loaded.

The first step is the choice of the future orthophotograph size. The default size given
by the program is the size of the area covered by the DEM. This can be changed but
must have limits inferior of those of the DEM. The predicted step corresponds also to
the step of the DEM. This factor will determine the resolution of the orthophotograph.

For every point in the orthophotograph, the DEM case in that it is localised is
computed. The Z value of the current point is then interpolated with a bilinear
interpolation.

Every point in the photograph must also receive the grey value of the corresponding
pixel in the epipolar image. This operation is realised by a transformation of the

background image

24

ground coordinates into the epipolar coordinates. The current pixel receives the grey
value of the corresponding pixel of the epipolar image.


8. VRML MODEL GENERATION AND VISUALIZATION

The Virtual Reality Modeling Language allows to represent every 3D space with a
sufficient reality.

So it’s possible to have a view of the DEM draped with the orthophoto by activating
the option Generate VRML file in the Ortho module. The user chooses the Z scale
factor.


You can visualise the generated file with a VRML viewer available on Internet.

Here below, an example of VRML result realised as from Sart-Tilman photos and
visualized with COSMO Player (free plug-in for Internet Explorer):

background image

25

STEP BY STEP EXAMPLE






The following step by step example have been realised with aerial photographs of the
Sart-Tilman. All the documentation necessary for the treatment is in the annexes.

Annex1: Calibration certificate
Annex2: Control and check points

To have more details on the different modules, refer on the precedent sections

1. Create a project

Start

“Digital Photogrammetric Software“

On the File Menu, choose the option New Project or click


The project window will pop up:

background image

26

2. Load the images

The program processes digital images with 8 bits/pixel (grey level). File format is
BMP.

The corresponding images for the Sart-Tilman are

- 1_7159.bmp for the left image
- 1_1760.bmp for the right image.


In the project window, click on the browser and choose 1_7159.bmp for the left
image and 1_7160.bmp for the right image.


3. Inner orientation

An inner orientation must be done for every image. This process allows establishing
a relation between the pixel and image coordinates. This relation is determined with
the help of the fiducial marks photo coordinates given in the calibration certificate
(see annex 1).

Encode the calibration certificate


In the project window, choose the option New CC. The following dialog box will pop
up:

background image

27

y

x


The units of the calibration certificate are millimeters (10

-3

m). All the necessary data

are in the annex 1

When the entire certificate is encoded, save it. A CC file is created.

Location of fiducial marks

In the project window, choose the option Left and Right IO. The following window will
appear:

Below the images automatically appear the photo coordinates of the fiducial marks.

The first step is the determination of the photo orientation. According to it, the
parameters of transformation will change. In the example, the photo orientation is :






Then, you must track down the fiducial marks on the image.

background image

28


Parameters of transformation will appear as from three fiducial marks and residues
as from four fiducial marks.

Correction of the radial lens distortion


In the menu Inner Orientation Module, choose the option Radial Distortion:
Then you can compute the coefficients of the polynomial after having chosen the
polynomial order. The values of the residues appear in the third column.


Before leaving the inner orientation, you must save the parameters:

menu Inner Orientation Module option Save OIparam. An OIParam File is created.
This operation must be done for the left and for the right image.

4. Relative orientation

Relative orientation allows determining the relative positions of the two bundles of
rays by creating spatial stereomodel in an arbitrary coordinate system. Tracking
down homologous points on the left and right images is necessary to make this
orientation.

Before starting the relative orientation, you must load the OIparam files of the left and
right images in the project window.

Click on the button RO by Rotations. The following window box will appear:

background image

29


By clicking with the right mouse button on a grey case of the left inferior part of the
screen, you can choose to add a point with the option Add Point. Then you must
track down the homologous point on the left and right images:

The left and right image coordinates of the point will appear in the left inferior part of
the screen.

NB: It is possible to delete one point by clicking on the ID with the right mouse button
and by choosing the option Delete This Point.

As from five homologous points, RO Final Parameters (expressed in Radians) can be
computed by clicking on Reset the first time and after Iterate. Then values of the
chart will change: the parallax will decrease, new RO Final Parameters are computed
and new model coordinates (X

i

Coord, Y

i

Coord) are computed for every image. Stop

the iterations when you think that the parallax values are small enough and that RO
Final parameters don’t significantly change.

background image

30

Before leaving the RO orientation, save the parameters by choosing the Save
Relative Orientation
option in the Relative Orientation Module menu. An
ORparam File is created for every couple of images.

It is also possible to save the file with the image coordinates of the homologous
points by choosing the Save Hmlg File option in the Relative Orientation Module
menu. An hmlg File is created.

Another possibility is the loading of an already realised Hmlg File. Either you load
this file in the project dialog box before starting the RO by Rotations or you load the
file by choosing the option Load Hmlg File option in the Relative Orientation
Module
menu.

5. Absolute orientation

The purpose of absolute orientation is the transformation from model coordinates to
ground coordinates. Control and check points are in the annex 2.

In the project window, click on AO & Restitution. There are two steps:

Determination of the parameters (menu Module option Determination)

As for the inner orientation, you can add or delete points by clicking on the right
mouse button. The following window will appear:



This window allows the encoding of the control, the check and the tie points. In the
point class, write ctrl for control points, check for checkpoints and tie for tie points
(statues of the points can be changed in the following steps).

Then you can track down these points on the two images. On the over view, control
points will appear in blue, checkpoints in green and tie points in red. As for three
control points, absolute orientation parameters can be computed by clicking on Reset
and then Iterate.

Stop the iterations when the residues on the control and checkpoints are small
enough.

background image

31


To save the absolute orientation parameters:
menu Absolute Orientation Module option Save

– OA Params

A Oaparam file is created

To save the ground coordinates of the points:
menu Absolute Orientation Module option Save - Hmlg & Ground Points
An OA hmlg file is created, with the image coordinates of the homologous points
An OActrl file is created, with the ground coordinates of these points. Tie points will
not have ground coordinates. To have them, see the Restitution module

It is also possible to load files before realised:
menu Absolute Orientation Module option Load - Homolog Points to load the
image coordinates of the homologous points
menu Absolute Orientation Module option Load - Ground Points to load the
ground coordinates of the homologous points

Restitution (menu Module option Application)


This module allows to track down additional points for that ground coordinates are
unknown and to compute their ground coordinates with the absolute orientation
parameters. The ground coordinates of the new points will directly appear. These
points will receive the type resti points.

background image

32

To save the ground coordinates of these new points:
menu Absolute Orientation Module option Save

– OAApply

An OAApply File is created

To load already tracked down points:
menu Absolute Orientation Module option Load OAApply

To make the eventual future realisation of plans easier, you can use the
geocodification module. This module allows generating files that can directly be
employed in topographic programs.

Create a Configuration File (.cl) that has the following structure

BEGIN_CODES

point

line
spline
END_CODES
BEGIN_LAYERS

road
hedge
fence
END_LAYERS

So you choose for every point its code and its layer. A file with the ID, the type, the
layer, the code and the coordinates of the points is created. This file can directly be
employed in topographic programs.

you put here all the layers that you want to use

you put here all the codes that you want to use

background image

33

Automatic detection of homologous points (menu Configuration option AutoFind

– ON)


This module allows the automatic detection on the right image of the homologous
point tracked down on the left image. For the configuration of the correlation window,
refer to the points 3.5 and 3.6 of the data processing part.


6. Epipolar resampling

Epipolar images are images for that the

-parallaxes are removed.


The necessary inputs are:

- the original images (left and right)
- the calibration certificate
- the inner orientation parameters

In the project window, click on Epipolar Images. The following dialog box will pop up:

background image

34


Load the necessary files in the corresponding cases.

Click on Compute Bounding Box to have the limits of this. These limits can be
changed. The X-limits depend on the overlay. The user can directly change the Y-
limits. To have more information about the determination of the limits, refer to point 4
of the data processing part.

When all these parameters are fixed, click on Process. You must directly give a
name and a File Path for the created image.


7. Image matching/DEM generation

In the project window, click on Matching

Image Matching


Image matching allows the automatic search of homologous points on the
overlapping area of several images. First choose the two images on which the
matching is made.

background image

35


Click on Generate in the Matched Hmlg Points part. A dialog box allowing the
configuration of the matching will appear.

background image

36

When you click on Configure Correlation Process, the same window that in the
automatic detection of homologous points (see the absolute orientation) will appear.

The superior part of the window contains various parameters. These are described
at the point 5 of the data processing part

For the predicted position of homologous points, there are two options:

Loading a homolog file

Computing an average shift X and Y values in pixels

When all these parameters are fixed, click on Process Matching. A window showing
the progress of the treatment appear. If you have made an error of configuration or if
the treatment is too long you have the possibility to stop it by clicking on Stop.

When the matching is finished, it is possible to visualize the matched points by
clicking Visualize.

ex:
MatchEpi.hmlg

So you can delete eventual non-homologous points. It is also possible to load a file
with already matched points by clicking on Find and by choosing the adequate file.

DEM Generation


A DEM is a numeric representation of the heights of a surface. It is a regular grid of
points, localised with their planimetric coordinates, for that the heights were
interpolated as from known points.

For this step, you need a file containing ground coordinates. To have this file, either
load an already OAApply file by clicking on Find or generate a new file with
Generate. This new file is generated as from an hmlg file that must be load in the
Matched Hmlg Points case. The hmlg file can be constituted of homologous points
tracked down on the epipolar images

– then mark the Epipolar Img case – or of

homologous points tracked down on the original images.

For the epipolar images, you need the cc and OAparam files.

For the original images, you need the Left and Right OIparam, cc, ORparam and
OAparam files.

ex: ST.OAApply

When the OAApply file is loaded, you can click on Generate in the DEM part. The
DEM config window will appear:

background image

37





Different parameters are proposed, that can be changed by the user. These are
described at the point 6 of the data processing part.

When the DEM is finished, save the file. A dem file is created.

ex: ST-5m.dem

8. Orthorectification

Orthorectification is the process of removing geometric errors inherent within
photography.

In the project window, click on Ortho.

background image

38




To construct the orthophotograph, dem file, epipolar images, cc and OAparam files
must be loaded.

Predict parameters appear, that can be changed by the user. Click on Generate Left
Ortho
and Generate Right Ortho. Save the images before leaving.

ex: ST_orthoG-0.5m.bmp
ST_orthoD-0.5m.bmp


It’s possible to have a view of the DEM draped with an orthophoto by activating the
option Generate VRML file. You can visualise this file with a VRML viewer available
on Internet.

background image

39



















ANNEXES

background image

40

Annex 1






CALIBRATION

No.

AF/ZEISS JENA LMK 266632B/1

DATE

14. 3. 94











CAMERA CALIBRATION CERTIFICATE


LMK LENS CONE

No.

266632B

TYPE

ZEISS JENA LMK

background image

41


CALIBRATION

No.

AF/ZEISS JENA LMK 266632B/1

DATE OF CALIBRATION

14.3.94

LENS TYPE:

ZEISS JENA LMK

Serial No

: 2666632B

FILTER TYPE

: NONE FITTED

ORIGIN OF MEASUREMENTS o:

The point of Symmetry

SIGN CONVENTION:

Distortion is positive if away from origin

CALIBRATED AT A TEMPERATURE OF 20°C

CALIBRATION PERFORMED BY:

D T PHILPOT


MEASUREMENTS

CALIBRATED PRINCIPAL DISTANCE:

152.314 mm


COORDINATES OF POINT OF SYMMETRY

x = -0.021

y = +0.003


COORDINATES OF PRINCIPAL POINT OF AUTOCOLLIMATION

x = -0.017

y = -0.003




RADIAL DISTORTION IN MILLIMETRES





Back of camera

Radius (mm)

20

40

60

80

100

120

140

Semi diagonal (5)

0,003

0,004

0,003

0,001

0,000

-0,001

-0,005

Semi diagonal (6)

0,003

0,002

0,003

0,003

0,002

0,000

-0,002

Semi diagonal (7)

0,004

0,004

0,004

-0,001

0,000

-0,002

-0,006

Semi diagonal (8)

0,003

0,002

0,002

-0,001

0,002

0,004

-0,003

Mean

0,003

0,003

0,003

0,001

0,001

0,001

-0,004

background image

42




CALIBRATION RESULTS FOR ZEISS RMK



CAMERA

ZEISS LMK 266632B

DATE

15 MARCH 1994

CENTRE OF GRAVITY

x = 110.0101

y = 109.9923

FIDUCIAL COORDINATES (SIDES)


x1 = 111.9978

x2 = -112.012

x3 = -0.01112

x4 = 0.009875

y1 = 0.020375

y2 = -0.00562

y3 = 111.9883

y4 = -111.998

DISTANCES

1-2 = 224.0100

3-4 = 223.9870

FIDUCIAL COORDINATES (CORNERS)


x5 = 109.9978

x6 = -109.992

x7 = -110.001

x8 = 110.0108

y5 = 110.0083

y6 = -110.007

y7 = 109.9843

y8 = -109.989

DISTANCES (DIAG)

7-8 = 311.1170

5-6 = 311.1312

DISTANCES (SIDE)

7-5 = 219.999

5-8 = 219.998

8-6 = 220.003

6-7 = 219.992

background image

43

Annex 2



background image

44






ROV001

– CTRL

middle of the sewage

X

Y

Z

234944.589

142672.707

242.564





ROV002

– CTRL

middle of the circular slab

X

Y

Z

234936.900

142094.695

232.874




ROV003

– CTRL

middle of the sewage

X

Y

Z

234566.378

142849.604

233.904


2

























































background image

45

ROV004

– CTRL

extremity of the sewage










ROV005

– CTRL

middle of the sewage

X

Y

Z

234859.241

142401.975

246.506

ROV006

– CHECK

bottom of the parking banisters





X

Y

Z

234866.893

142323.762

244.944

























































X

Y

Z

234451.001

142059.345

245.169

background image

46





ROV008

– CHECK

middle of the sewage

X

Y

Z

234444.199

142440.695

241.045

ROV009

– CHECK

extremity of the stile









X

Y

Z

234422.382

142348.426

242.670






















































ROV007

– CHECK

middle of the sewage








X

Y

Z

234644.270

142036.037

241.687




Wyszukiwarka

Podobne podstrony:
M12 Oncore Users Guide Supplement
Mathcad Users Guide
Audio?ughter?rd Users Guide
Echo Link Users Guide
DFMProForNX Users Guide
MMConverter v2 0 Users Guide
PICkit 2 Users Guide
M12 Oncore Users Guide Supplement
Mathcad Users Guide
users guide PL
users guide
Faces 4 0 Users Guide
metasploit users guide
nikon d70 users guide
PBGrid Users Guide
PipBoxer V2 0 6 Users Guide

więcej podobnych podstron