3D Forest

User Guide
Release 0.51

A tool for processing of point clouds acquired by terrestrial laser scanning in forests
The Silva Tarouca Research Institute, Pub. Res. Inst.
Department of Forest Ecology.
Authors: Michal Petrov, Martin Krůček, Jan Trochta, and Kamil Král
©2020
Table of contents
3D Forest 1

User Guide 1

Home 4

1. About Application 5

2. Detectable Attributes 6

3. Installation 7

4. Brief Insight and Control of the Application 9

5. Basic Workflow 14

6. Setting a Project16

7. Terrain analysis 18

8. Vegetation 21

9. Trees 24

10. Crowns 29

11. QSM Models and Tree assortments 32

12. Data Export 34

13. Other Features 36

14 References 38
Home
Welcome to the 3D Forest wiki/user guide. All information about the installation, usage, and
methods implemented within the 3D Forest application can be found here. For more info see
the sidebar/table of contents.
1. About Application
The 3D Forest is an application created to process terrestrial laser scanning (TLS) data, point
clouds, and to gain detailed information about forest stands and individual trees.
The application is released under the terms of the GNU General Public License v3 as
published by the Free Software Foundation. For more info about the license, one can view the
LICENSE.txt file located in the program folder or on the web page:
http://www.gnu.org/licenses/.
The application is written in C++ and depends on the libraries: VTK, PCL, Eigen, Boost,
Flann, LibLAS, and Qt. For a successful installation, users have to build these libraries before
compiling the 3D Forest. The source code is provided for downloading and compiling the
application on any computer. The Windows installer, as well as trial data, are available on the
site: www.3dforest.eu.
2. Detectable Attributes
In the current version of the 3D Forest (0.51) the following characteristics can be calculated:
Tree Attributes
 Position: gives X, Y, Z coordinates of the tree base in the Cartesian coordinates
system. More information in section Trees.
 DBH: this attribute determines diameter at breast height, i.e. the tree stem diameter
calculated from a sub-set of points from 1.25 to1.35 m above the tree base. See also:
least squares regression and randomized Hough transformation.
 Height: vertical distance between the tree base and the highest point of the tree, (i.e.
the max. difference in Z coordinate in meters).
 Cloud length: gives the longest distance between two points in the cloud. It is suitable
for the length calculation of highly inclined or lying trees.
 Stem Curve: determines the stem centers and diameters calculated in various heights
above the tree base (0.65m, 1.3m, 2m, 3m, etc.).
 Convex planar projection of the tree: returns polygon with the shortest boundary
containing all points of the tree cloud orthogonally projected on the horizontal plane.
 Concave planar projection of the tree: returns polygon with the smallest area
containing all points of tree cloud orthogonally projected on the horizontal plane.
 The number of tree points: yields the number of points representing a single tree.
Crown Attributes
 Crown centroid: gives coordinates of the tree crown center position (X, Y, Z) in the
Cartesian coordinates system as computed from crown external points. More
information in section Crowns.
 Crown position deviation: returns the crown position deviation from the tree base
position. It is defined by distance (m) and direction (°).
 Crown bottom height: provides the vertical distance (i.e. the difference in Z
coordinates in meters) between the tree base position and height of the place where the
lowest living branch attaches the main stem.
 Crown height: computes the vertical distance (i.e. the difference in Z coordinates in
meters) between the crown bottom height and its highest point.
 Crown total height: gives vertical distance (i.e. the difference in Z coordinates in
meters) between the lowest and the highest point of the tree crown.
 Crown volume by voxels: returns crown volume computed from voxels of a given
size.
 Crown volume and surface area by concave polyhedron: computes crown volume and
surface area utilizing cross-sections of a given height and concave hull threshold
distance.
 Crown volume and surface area by 3D convex hull: returns volume and surface area of
crowns 3D convex hull.
 Crown intersections: gives volume and surface area of intersecting space between
convex hulls of two crowns.
3. Installation

WIN
The 3D Forest installer for Windows is located on https://github.com/VUKOZ-
OEL/3DForest/releases/download/v0.5/3DForest_05.exe. Once downloaded, the installation
wizard will guide you through. Furthermore, sample data to test the 3D Forest are placed on
https://github.com/VUKOZ-OEL/3dforest-data. For uninstalling run the uninstall.exe file
located in the installation folder.

WIN building from source code

There is an option to download and compile the source code for Windows too. In such a case,
additional libraries and compiler are needed as prerequisites: VKT, PCL, Qt, Eigen, Flann,
libLAS, and for example MinGW64 cross-compiler. At the moment, there is a modified
version of tpoint cloud library (PCL) for point and area picking in the selection mode.

For installation from the source, it is important to install dependencies first. It is also important
to have compilator, IDE and GIT installed.

Let's start with Visual Studio, and VCPKG package manager.

 first download vcpkg: git clone https://github.com/microsoft/vcpkg.git

 build vpckg again: .\bootstrap-vcpkg.bat
 find the right version of dependency - problem is with VTK and Qt those will be
installed separatelly. -Finally install first round of dependecies: vcpkg.exe install
eigen3 boost flann liblas qt5 --triplet x64-windows
 vcpkg.exe integrity install

FLANN

Even you installed Flann library in previous step, you will need to use different version in vcpkg
directory try to find all versions vcpkg x-history flann and try to use version 1.9.1 git
checkout c626675abb963d15f5d290a56005556d95b160bd -- ports/flann after this
remove flann vcpkg remove flann and install back vcpkg install flann

VTK

Since vcpkg has some isssues with vtk versions, vtk should be compiled separatelly from
source.

 get source https://www.vtk.org/files/release/8.2/VTK-8.2.0.zip using

CMAKE configure dont forget use with Qt and install using i.e. VS project. I had
problem with QpaintPath, so just include in into file.
Install PCL From source
 git clone https://github.com/janekT/pcl.git
 open in Visual studio configure,build and install
Install all from source
 QT: git clone https://code.qt.io/qt/qt5.git qt5
 BOOST
https://dl.bintray.com/boostorg/release/1.73.0/source/boost_1_73_0.zi
p
 eigen: git clone https://gitlab.com/libeigen/eigen.git
 flann: git clone https://github.com/mariusmuja/flann.git
 VTK: https://www.vtk.org/files/release/8.2/VTK-8.2.0.zip
 libLAS: git clone https://github.com/libLAS/libLAS.git
 PCL: git clone https://github.com/janekT/pcl.git
 3D FOREST: git clone https://github.com/janekT/3DForest.git

macOS, Linux
For the macOS and Linux users, there is only an option to download and compile the source
code.
At first, dependencies are installed using brew install boost vtk qt flann Eigen
LibLAS Cmake. The second step is to compile PCL from the source as 3D Forest uses
unapproved changes in code. It can be downloaded from
https://github.com/janekT/pcl/tree/pointpicking using git or as a zip file. Cmake creates a
project for Xcode. The PCL Library has to be compiled and installed then.
The 3D forest can be downloaded from the repository https://github.com/VUKOZ-
OEL/3DForest/tree/master. CMake finds all dependencies for the configuration of the
project. The 3D forest can be run via Xcode.

Linux installation is done by the package manager similarly to MacOS.

4. Brief Insight and Control of the Application
The control toolbars and menus are on the top of the application window. Toolbars are
divided into logical sections – project bar with opening, creating new, importing of projects,
etc. The next toolbar affects the view of the clouds in the visualization area. Six icons
represent six predefined views of displayed clouds. These are followed by two icons
representing the change of the view – orthogonal and perspective projection. The tree and
crown toolbar serves for displaying computed attributes. There is a list of all clouds imported
into the project on the left side of the application. The biggest part of the application
win\sdow is the visualization area where selected clouds and tree/crow variables are
visualized. There is a status bar where information about selected methods or clouds is
displayed. In the right bottom corner is displayed progress bar when some action takes place.

Tree and crown toolbar. If not coloured, values must be computed first
Main menus using the main menu.
Show/hide icons. They serves for refreshing visualized information
Project icons after recounting too.

Tree widget

Visualisation
toolbar

Information line; progress bar is shown in this area. Visualization area

Figure 1: The application window and description of its parts.

Basic control
Mouse button configuration in the visualization area:
Left-click: draw to rotate the view
Middle-click: draw to zoom the view
Middle rotate: zoom the view
Right-click: draw to move the view
Shift + left click: information about trees at the selected point
Ctrl + left click: clockwise or counter-clockwise rotation

Context menu on cloud list (right-click on cloud name):

Delete: delete cloud from the project and if needed from the disk
Color: change the color of a cloud
Color by field: set the color of points in the cloud by selected field (axis x, y, z, and intensity)
Point size: change the size of points in the cloud
All Clouds ON: all clouds in the project are visible
All clouds OFF: all clouds in the project are invisible

Keyboard
p, P: switch to a point-based representation
w, W: switch to a wireframe-based representation (where available)
s, S: switch to a surface-based representation (where available)
+/ -: increase/decrease overall point size
g, G: display scale grid (on/off)
u, U: display lookup table (on/off)
r, R: reset the camera
f, F: focus on the point
x, X: selection (in selection mode)
o, O: perspective (on/off) Alt +: show and allow quick access keys in the menu (underlined
character)
Tips: Switching between wireframe and point-based representation does not affect the point
clouds. The correct values of the scale grid are shown only when the perspective is off. The
display lookup table shows values extent and color range if Color by field is used.
Cloud type description
All data imported into the 3D Forest are stored in the PCD file format (binary compressed
PointCloudData). PCD file contains coordinates of all points with intensity value. These four
variables of the point representation are being used for the moment. Additional information
such as color or path to the transformation matrix file can be found in the MyProject.3df file.
There are different point cloud representations during 3D Forest data processing:
Base cloud:
Represents raw data imported into the 3D Forest. The data should be preprocessed in some
other software (usually provided with the scanner), where cloud fitting and registering are
made. The base cloud is not differentiated. It embodies points of all objects such as terrain,
vegetation, buildings, etc.

Figure 2: Example of the base cloud before any segmentation – all points contain just X, Y,
Z, and intensity values; no objects or surface types are differentiated.

Terrain cloud:
Using Terrain analysis the base cloud can be divided into vegetation and terrain – two main
parts of the forest ecosystem. The terrain cloud represents ground surface and can be used for
better calculation of tree position or exported into GIS software for detailed
(micro)topography analysis.
Figure 3: Visualization of Terrain cloud (brown) and Vegetation cloud (green) after
ground/vegetation separation.

Vegetation cloud:
The part of the base cloud which is not terrain is labeled as vegetation. From this point cloud,
the individual trees can be segmented automatically or manually selected so to create another
cloud type – tree cloud.
Tree cloud:
This cloud represents a single tree after manual segmentation. Only for tree clouds, it's
possible to compute basic tree variables like DBH, height, position, etc.
Figure 4: Visualization of Tree clouds after tree segmentation – individual trees are displayed
in different colors.

Other:
Clouds that represent unclassifiable points or points which do not belong to any other cloud
type. These clouds can be displayed or used as vegetation for tree selection/segmentation.
5. Basic Workflow
Once the application is installed, the workflow is to set up a project, import data, and segment
data to terrain, vegetation, and trees. Only then the tree and crown features can be evaluated.

Start a Project (Project Menu)

The first step when using the 3D Forest is to set up a project. This can be done via the Menu -
> Project -> New Project. In this step, a directory of the whole project can be selected. This
directory then becomes a place for data storage. The transformation matrix, i.e. the data
extension, can be set at this point.

Import Data (Project Menu)

Once the project is set, the raw data can be imported. It is essential to import them as the base
cloud if they consist of all points of the forest (terrain, trees, buildings, etc.). In case the data
was already separated into terrain and vegetation, each point cloud has to be imported
separately according to its type (terrain or vegetation). Likewise, the clouds of individual trees
should be imported as tree clouds.

Visual Check
It is important to check the cloud quality after the import. Mainly if there are any doubts
during setting-up a project. If the cloud looks like parallel lines, the transformation matrix is
wrong. Then the project has to be recreated with a different (appropriate) transformation
matrix.

Separate Ground and Vegetation (Terrain Menu)

Once the base cloud is imported, it needs to be divided into two basic parts: terrain and
vegetation. Two alternative methods are provided to fit this request. Terrain from octree,
which returns ground points of cloud. And terrain from voxel, which gives a voxelized cloud
represented by centroids of original points in each voxel. All points that do not belong to the
terrain cloud can be manually selected and removed during a visual check. You can also use
some of the implemented filtering methods.

Analyze Terrain (Terrain Menu)

There are functions to clean terrain point cloud using the following filters: statistical outliers
filter and radius outlier filter provided by PCL. Besides, there are functions for interpolation
of missing terrain using inverse distance weighting (IDW) if gaps are present, and functions
for evaluation of slope, aspect, curvature, hillshade, and a function that helps to identify
features within the terrain point cloud.
Segment Vegetation (Vegetation Menu)
For a tree analysis, it's essential to work with clouds that represent single trees. This is
achievable by automatic or manual segmentation of the vegetation. During manual
segmentation, all the points that do not belong to the tree have to be removed before saving a
single tree cloud. On the contrary, automatic segmentation treats the whole vegetation cloud
according to the input parameters. A more detailed description of the automatic segmentation
method is presented in section Vegetation.

Edit a Tree Cloud (Trees Menu)

If needed, the tree cloud edit helps to manually remove points that do not belong to the
particular tree. It also enables to merge a tree separated into two clouds.

Analyze Trees (Trees Menu)

The 3D Forest is designed for individual tree analysis. Thus the first attribute to be evaluated
is the tree position (defined as the tree base position). This is required for the estimation of all
other tree variables like tree DBH, height, length, stem curve, or tree planar projection.

Analyze Crowns (Crowns Menu)

TLS is a great tool for scanning objects in remote distances, e. g. a tree crown. However, the
analysis of tree crown requires the position and tree height to be calculated first. Unless this is
done, the crown analysis menu is inactive in the Crowns Menu (greyed out). Besides, the
crown analysis has to be computed using the cloud that consists of points that belong to the
tree crown only, i.e. not to the stem. There are both manual and automatic segmentation
procedures to do that. Once the crown is segmented from the tree, crown features (crown
height, crown length, crown width, crown centroid, crown displacement, crown voxel volume,
crown convex 3D hull, crown concave hull, or intersection of crowns) can be evaluated.

Quantitative Structure Model (QSM Menu)

The last group of attributes is listed under the tree quantitative structure model (QSM). It
gives information about stem and branches (length, connection). As well as for the crown
analysis, the position and tree height has to be evaluated first.
6. Setting a Project
New project set up
Setting up a new project is done via the Menu -> Project → New Project. Also, there is the
New Project icon in the project context menu. In the project manager window, set the
name of your project (without spaces) and select the path where to save it. In this location, a
new folder with the project name will be created and all files of the project will be stored
there, i.e. project base file (MyProject.3df) and all imported clouds. The next step is to set a
transformation matrix. There is only one transformation matrix for each project to be
employed. The transformation matrix serves for reducing the number of digits in coordinate
values for faster data management (RAM and cache-friendly administration). To create a new
transformation matrix it's necessary to know the exact extent of your data coordinates. It is
also possible to select from pre-prepared matrices or create a new one. For creating a new
matrix set name of matrix, and values of transformation. The goal is to set the transformation
matrix as a number as big as possible so that subtracting this number from your coordinates
do not change the data extent that carries information. There is an example in table 1. There is
also the NO_MATRIX option if doubts about data extension. Data are transformed into the
new coordinate system automatically during import, only for .pcd files there is a possibility to
decide if transformation gets involved. Correspondingly, exported data from 3D Forest are
transformed back into the original coordinate values.
Table 1: An example of the data before and after transformation.

Original data Transformation Data extent after

extent matrix transformation

max. x 18857748,76 648,755

-18857100
min. x 18857104,96 4,96
max. y -987657123,6 -323,55
987656800
min. y -987656816,6 -16,55
max. z 425,33 425,33
0
min. z 358,13 358,13

Project Opening
A project can be opened via Project→ Open Project, or click on the Open Project icon in
the project context menu and select your project file (file with extension .3df).

Project Import
Since projects save their path, simple copying into another location (disc or directory) would
break the project. If such a change is needed, it has to be done as an import of an old project
into a new location. Go to the Project → Import Project. Choose the path to the old project,
set the name of the new project and location where the new project will be created. The new
folder with the project name will be created and all data files will be moved there. It is also
possible to select the option for removing the old project from disk after the import.

Data Import
The 3D Forest enables to import data into a project via Project → Import. It is possible to
import data in following formats: .txt, .xyz, .las, .pts, .ptx. If the user's data already have the
3D Forest native format (PCD), it is important to choose an appropriate cloud type (see Cloud
type description) and to set the transformation matrix. Once the data are imported, they are
listed at the main window and in the list of project clouds.
7. Terrain analysis
There are two automatic methods for terrain processing. None of these methods are perfect, so
it is advised to use both of them to combine the advantages of their outputs. There is also a
possibility of manual adjustment of the results.

Terrain from octree

Go to Terrain → Terrain from octree. In the dialog window select the base cloud that should
be processed. Enter the name of the new terrain cloud, the name of the cloud where the rest of
the data (usually vegetation) will be stored, and set the resolution (length of cube edge in cm).
The input cloud is divided into cubes. Cubes containing points and having the lowest z-value
are considered to be the "ground cubes". The terrain is then defined by the points in these
ground cubes. This means that all the points in ground cubes are considered to be terrain and
contrary to voxelization there is no reduction of points and no generalization. This method is
more laborious for manual post-processing. However, the results are more detailed and built
by original points.
HINT: If there are points that belong to the terrain within the vegetation cloud, i.e. points
connecting trees after the segmentation, it's recommended to use a higher resolution.

Terrain from voxels

Go to Terrain → Terrain from voxels. In the dialog window select the base cloud to extract
the terrain. Enter the name of the new terrain cloud, the name of the cloud containing the rest
of the data, and set the resolution of voxelization (length of voxel edge in cm). Appropriate
resolution to gain relevant results is below 50 cm. The input point cloud is reduced into voxels
and centroids of voxels (i.e. points with averaged coordinates from all original points of
individual voxels). Voxels with the lowest z-values define the terrain. Increasing voxel size
also rise the overestimation of terrain height (z-value of centroids is affected by more points
which lie above the terrain). However, as the resulting cloud contains fewer points than the
cloud produced by Terrain from the octree also manual post-processing is usually faster.

Statistical outlier removal filter

Both terrain cloud from voxels or form octree may be filtered using statistical outlier removal.
Go to Terrain → Statistical outlier removal. In the dialogue window select terrain cloud set.
Name the cloud with removed points, and set the number of neighbors for computing mean
distance and standard deviation. Points with a mean distance longer than standard deviation
are removed. For more details see the PCL documentation.

Radius outlier removal filter

Applying radius outlier removal filter to terrain cloud is possible via Terrain → Radius
outlier removal. In the dialogue window select terrain cloud set. Name the cloud with
removed points, set the radius, and the number of neighbors. Points with fewer neighbors in a
given radius are removed. For more details see the PCL documentation. Tips: If too many or
too few points were removed by filters use the cloud merge feature and repeat the procedure.
For the best result combination of both filters is recommended. Note that different voxel sizes
and various initial cloud density (if Terrain from octree is used) need different filters settings.
Both filtering methods are better to use with cloud created by Terrain from voxels due to its
regularity.

Manual adjustment
The above-described terrain extraction can be also adjusted by manual editing. Go to Terrain
→ Manual adjustment and select terrain cloud for editing. Set a name for a new cloud where
points removed from the terrain cloud will be saved. Pressing the "x" button activates the
selection box. The left mouse button makes the selection. The designated points will be
removed from the terrain file and stored in the separate file defined above. For a step back
press the undo icon → in the top panel . There is also a split function that creates user-
defined strips of terrain for easier point selection. Arrows switches among the strips. The strip
selection is finished by clicking at the split function again. Stop EDIT icon saves all
changes in the original (input) terrain cloud.

IDW interpolation
Areas with missing terrain points (typically shaded by thick trees or located right under the
scanner during scanning) may be filled in by Inverse Distance Weighted interpolation (IDW)
from the surrounding terrain points. Go to Terrain → IDW to select an input terrain cloud for
interpolation. Set a resolution of interpolation in cm, a number of the closest points (n) to be
included in the interpolation. Name the new interpolated terrain cloud that will be free of
empty areas. The interpolation uses n-closest points of the original terrain to estimate Z value
of the new terrain point.

Slope
The slope analysis is a method that evaluates the terrain slope. This function is reachable via
Terrain → Slope. The terrain slope, in general, is defined as the first derivative of the terrain.
The slope is thus calculated from the height difference based on a user-defined sector. Users
can choose between the sector defined by the nearest neighbors or the radius. Each point of
the selected terrain cloud is then compared to the others within the defined sector and the
mean slope value is computed. The result is stored as a field of intensity in degrees or in
percent, where 45 deg represents 100%.

Aspect
The aspect is a method for evaluation of the slope orientation for a given terrain layer. There
is defined surrounding for each point of the terrain layer where the average orientation is
calculated. The result is given in degrees, i.e. -180 to -135 = West, -135 to -45 = South -45 to
45 = East 45 to 135 = North 135 to 180 = West. The result is stored as a field of intensity.
Hillshade
This function returns shaded relief of a given layer of terrain. The values of the resulting
cloud range from 0-255 according to the azimuth and orientation.

Terrain Features
This is a method that looks up identical places in the field based on their parameters. Identical
places are defined by their value in the field of intensity, the number of points, the size of the
major axes, the axes ratio, the size of the maximal area (the concave), the convex, and the
convex/concave ratio. The result is not saved to disk nor project as there is the Export
Features function to do it.

Feature table
The feature table displays a table of attributes for each terrain feature.

Export Features
This function exports the terrain feature information. The user selects the directory to save the
file, delimiters of fields, and files to export. There is an option to choose between an attribute
file and Convex/concave polygon. The attribute file is a file with the position of the centroid
of selected voxels and its attributes. Convex/concave polygons are polygons delimiting voxels
of given features. These are saved in the VKT format.
8. Vegetation

After the terrain extraction, there are two types of clouds in the project: the terrain and the
vegetation clouds. However, the vegetation cloud needs to be further segmented into single
trees to get the desired information about tree attributes. Right now there are two methods
implemented in the 3D Forest, automatic and manual segmentation.
Automatic segmentation - version 0.5 and higher
The automated segmentation is called via Vegetation → Automatic tree segmentation. There
are ten inputs needed to start. The vegetation and terrain clouds of interests are obvious. The
voxel size, descriptor type, descriptor threshold value (%), amount of iterations, number of
voxels in element, and the distance from terrain are explained in detail in the next paragraphs,
as well as the method itself. The cloud prefix and non-segmented points output name define
the output.
The algorithm used in version 0.5 and higher is based on searching neighboring voxels
according to the chosen descriptor. At first, voxels of a given size (voxel size in cm) are made
through the whole vegetation cloud. Each voxel is then evaluated by user-defined descriptor:
Principal Component Analysis (PCA), slope, intensity, and PCA-slope multiplication.
PCA computes Principal Component Analysis of x,y,z coordinates of all points inside each
voxel. The ratio of different PCA axes is then used as the descriptor. Based on our
experiences, the PCA descriptor usually gives the best segmentation results for TLS data.
The threshold value is defined as a percentile of the descriptor value. The range is 0-100%
and represents the lower limit of the used voxels for segmentation. For example, if the input
threshold value is 70 then only the voxels whose descriptor value is at 70% of the actual value
range or higher are used for the tree extraction. The higher the threshold is, the more similar
and compact surfaces are represented. The ideal values of the input parameters are higly
dependent on the point density and structure of the forest stand (see Tips below this
paragraph). Resulting representations are the surfaces of a stem or big branches.
All voxels above the descriptor threshold value and within the voxel size are then grouped.
This means that these groups of voxels contain points of similar topological characteristics
and the voxels are neighbors (connected without interruption). Such groups of voxels are
called elements if they satisfy the condition of a minimal number of voxels in an element.
Within these elements, the tree bases are classified by complying with the given distance
above terrain points.
As branches or stems may be interrupted by occlusions bigger than the voxel size and/or small
sprigs do not always fill the descriptor threshold value, it is needed to add more voxels to
make the tree complete. The established tree bases are thus gradually connected with voxels
forming the rest of the tree parts during the iterative process. This allows us to find all free
voxels that are onward from the voxel size.
The iterative method is automatic and has two parts. It starts by omitting the descriptor
condition. Thus all voxels that are up to the voxel size are considered as belonging to the tree.
The surrounding of every added voxel is instantly searched for the neighbors. Once all
neighbors are added, the voxel size is increased by its value, i.e. plus the voxel size, and the
new surrounding of established tree bases are searched again. If the free voxels or elements
within this space have the descriptor value up to the threshold they are added to the tree.
Afterward, all free voxels laying within the doubled voxel size are added to the enlarged tree
as well as their neighbors. These two steps are repeated iteratively according to the number of
iterations. The appropriate number of iterations thus depends on your data quality and forest
characteristics. The data with minor occlusions can be segmented well with few iterations. On
the other hand, too few iterations can lead to omission errors (missing treetops), too many
iterations can lead to commission errors, especially in very dense forests.
When the iteration is finished, segmented trees are saved into the project folder as separate
clouds with a given prefix and number of the tree. All points that are not selected as a tree are
saved in vegetation as a “rest cloud”.
Tips:
 Voxels should have a size to incorporate at least 3 points from the point cloud. If the
point cloud is dense enough, a smaller voxel size can be used and vice versa.
 For dense forest stands smaller voxel size is advised to maximize the number of
segmented trees. On the other hand, with small voxel size trees might not be complete
after segmentation.
 For sparse/even forest stands larger voxels can be used.
 If neighboring trees are merged after segmentation, try to increase the threshold of the
descriptor value or/and use smaller voxels.
 If you want to avoid segmentation of small trees increase the distance from terrain and
voxel size.
 It is better to use fewer iterations for deciduous trees scanned in the leaf-off state
(lower occlusion is anticipated). On the contrary, for coniferous (evergreen) trees
usually, more iterations are required due to occlusions in crowns.
 In some cases, two iterations of segmentation can be useful, first run the segmentation
with finer parameters. Then segment the rest-of-vegetation point cloud with coarser
parameters.

Manual segmentation, segmentation of alternative vegetation

Manual tree segmentation or segmentation of additional vegetation is done via Vegetation →
Manual tree selection. The source cloud with vegetation can be selected in the dialog window
as well as the name of the cloud where the rest of unsegmented points will be saved. The
name of the cloud with a segmented tree (i.e. the tree cloud) is set after actual segmentation.
The segmentation itself is done via the "x" button that activates the selection mode. The
selection box is drawn using the left mouse button. Selected points are removed from the
view. There is an “undo” key if needed. The goal is to remove all points that do not belong to
the target tree. After editing, i.e. the target tree is shown only, “Stop EDIT” icon ends
editing mode. The filename of the new tree cloud has to be set. Once the segmented tree is
saved, another segmentation process can be started (all the removed but unsegmented points
will reappear) or terminated. Some points of vegetation belonging to trees, e.g. isolated
branches, may remain in the unsegmented vegetation file after segmentation. These points
may be attributed to any of the segmented trees by the “Cloud merge” function.
Tips:
 Larger areas are recommended to split into smaller tiles for improved visibility and
easier handling. This may be done by employing additional software, e.g. LAStools.
 The segmented trees are removed from the edited vegetation cloud. This leads to
gradual simplification. Thus, the easy shaped and/or stand-alone trees are suggested
to be segmented first and the complex parts afterward.
 If the tree position and DBH are the only required variables, it is not necessary to
segment the whole trees. For these variables, it's enough to separate the bottom part of
the stem only (approximately 2 meters above the terrain).
 The neighboring tree clouds, and/or the vegetation cloud can be displayed by
checking the box in the tree widget. This may help to decide on the controversial
points.
9. Trees
The trees is a part of the 3D forest menu that serves for computing attributes of individual
trees or tree cloud editing.

Tree Cloud Edit

To remove points from an existing tree cloud, go to the Trees → Tree cloud edit. The dialog
window gives options to choose the tree cloud to be edited and to set a name for a new cloud
with all removed points. Points are removed by the same scheme as in the manual
segmentation, see section Vegetation.
DBH Cloud Edit
If the results of DBH (Diameter at breast height) estimation are suspicious, there is an option
to display and edit the calculation source point cloud. The path Trees → DBH cloud edit
enables the selection of the tree and the respective DBH cloud to be edited. The editing
scheme is as in the manual segmentation again. All selected points will be excluded from the
DBH computation. However, they will be removed from the tree cloud temporarily, i.e. for
the 3D Forest session. The DBH cloud edit is voided by closing the project.
Tree Base Position
Tree base position is a key variable providing a baseline for computation of other tree
variables such as DBH, tree height, stem curve, and the visualization of convex/concave tree
projection. Therefore, these functions are not available before the tree position is calculated.
The 3D forest offers two methods for tree base position calculation. However, their results
may slightly vary (Fig. 5). The selected method thus influences all dependent tree variables.
For that reason, these variables are automatically modified according to the actual tree
position. Still, it is important to keep in mind, which method of the tree position calculation is
being used for the particular trees. Both methods may be calculated with or without the terrain
cloud. This is used for adjustment of the tree base Z coordinate. It is strongly recommended to
use terrain adjustment if the terrain cloud is available. In such a case, the influence of the tree
cloud lowest outliers on the resulted tree base position is lowered (Fig. 5). This is particularly
important on steeper slopes. The possibility to calculate tree position without terrain is thus
set aside for cases when terrain cloud is missing.
Tree Base Position by Lowest Point
The tree base position computed by the lowest point method can be used via Trees →
Position lowest point. The dialog window enables the selection of trees whose base positions
should be calculated. Besides, it is necessary to set the vertical distance [cm] from the lowest
point of the trees to set the tree points included in the calculation. The terrain cloud, if
available, should be selected to adjust the Z coordinate to the terrain level. Correspondingly,
the number of the closest terrain points (N) employed in this adjustment has to be set. The XY
position is then computed as a median coordinate of the points of the tree that are lying
between the lowest point of the tree cloud and the user-defined distance. The Z coordinate is
defined as the median Z value of N closest points of the terrain at the XY position. If there is
no terrain cloud available, the Z value is defined by the lowest tree point. The tree base
position is displayed as a sphere with the center at the position and radius of 5 cm.
Tree Base Position by RHT
The tree base position computed by randomized Hough transformation (RHT) is accessed via
Trees → Position RHT. The dialog window enables users to select the tree(s) for tree base
position calculation and set the number of RHT iterations. It is recommended to use at least
200 iteration steps. The terrain cloud, if available, should be specified to adjust the Z
coordinate to terrain level as well as the number of the closest terrain points (N) used for that
adjustment should be set. The iteration takes the lowest point of the tree cloud as an initial
position. The tree base position by RHT then computes the tree base position using centers of
two circles fitted by RHT to the stem at 1.3 and 0.65 m above the initial or previously
computed position. The RHT position is then defined as the intersection of the vector aiming
from the upper circle center (at 1.3 m) to the lower circle center (at 0.65 m) and the horizontal
plane defined by the median Z value of the N closest points of the terrain cloud.

Diameter at Breast Height (DBH)

The diameter at breast height (DBH) is computed from the subset of points of the tree cloud
(so-called DBH cloud). This cloud lies between 1.25 m and 1.35 m above the tree base
position. It is highly recommended to build the DBH cloud on the tree base position adjusted
to terrain level as mentioned in section 7.3, otherwise, the DBH might be significantly
misrepresented. There are two methods of DBH calculation implemented in the 3D Forest: i)
based on the randomized Hough transformation (RHT) and ii) based on the least square
regression (LSR). Both methods use the same DBH cloud as input, however, their results may
differ. The RHT is a robust method if there are enough points of the segmented stem. This
method is however computationally more demanding. On the other hand, the LSR function
works well even with a small number of points of the segmented stem. However, it frequently
overestimates the diameter value as it is very sensitive to the presence of outlying points. A
significant difference between results given by these two methods points to the low quality of
the tree segmentation or the overhanging branches. This may be fixed by the Tree cloud edit
function (8.1) or by the DBH cloud edit function (8.2), where only the points of the tree
subset from 1.25 to 1.35 m above the ground are adjusted.

Figure 5: Differences between the DBH computed according to the tree base position
estimated from the tree’s lowest points (left) and the DBH recomputed according to the tree
base position adjusted to the terrain cloud (right). The tree base position is represented by the
red dot.
DBH by Randomized Hough Transformation
Computing DBH using randomized Hough transformation is done via Trees -> DBH RHT.
The dialog window gives the option to select trees of interest and set the number of iterations.
As always, the amount of iterations is a trade-off between computation time and accuracy. It
is recommended to use at least 200 iterations for fast computation. A higher amount of
iterations costs more computational time with rising accuracy. The use of 2000 iterations
already provides fairly consistent results. Resulting DBH is displayed as a 10 cm high
cylinder with the diameter of the best-fitted circle.
The method itself is based on the parametric description of objects within the polar coordinate
system. The DBH subset of the tree cloud (i.e. from 1.25 to 1.35 m) is projected to a
horizontal plane. The Z coordinates are transformed to 1.3m. Then, the scheme searches every
possible center of the circle for each point of the subset. The most frequent circle center is
selected as a result (Xu and Oja, 1993).

DBH by Least Squares Regression

Computing DBH using Least squares regression is done via Trees → DBH LSR. The dialog
window gives the option to select trees of interest. Resulting DBH is again displayed as a 10
cm high cylinder with the center and the diameter of the best-fitted circle. As already
mentioned, the LSR method is very sensitive to outlying points that do not represent the tree
stem (Fig.6). Therefore, this method is suitable for tree clouds with clearly visible stems.
The method starts similarly as DBH RHT, i.e. DBH subset of points lying between 1.25 and
1.35 m is projected to a horizontal plane and the Z coordinates are transformed to 1.3m. Then
the circle is fitted to these points by the Least-squares regression. The LSR method is based
on minimizing the mean square distance between the fitted circle and data points. The circle
fitting is done employing the Gauss-Newton method (Chernov and Lesort, 2003).

Figure 6: Differences between the RHT (left) and the LSR (middle and right) method; the
wrong value given by LSR is caused by an overhanging branch, which was included in an
automatically defined DBH subset of the tree cloud.
Tree Height
To obtain the tree height go to Trees → Height. There is an option to select trees of interest in
the dialog window. The result is displayed in meters on the top of the tree. This tool also
displays a vertical line from the tree base position to the highest point.
Tree Length
The tree or cloud length calculation is done via Trees → Length. The dialog window gives an
option to select tree clouds of interest. The tool calculates Euclidean distance between the
most distant points in the tree cloud and displays their connection. The cloud length is
displayed in meters at the bottom of the tree. This tool is suitable for the calculation of the real
length of inclined or lying trees.

Convex Planar Projection

The tree convex planar projection is evaluated via Trees → Convex planar projection. The
dialog window gives an option to select tree clouds of interest. The convex planar projection
tool calculates and displays the area of the convex hull of the tree cloud orthogonally
projected to the horizontal plane at the tree base height. The result can be exported/imported
to the ArcGIS polygon shapefile (see Exporting tree planar projection in section Data Export.
The calculation is based on the Gift wrapping algorithm (Rosén et al., 2014). As the first step,
the point A with the lowest y coordinate is found. This point (A) is the first and the last point
of the polygon. Next, two vectors are defined. They are the vector defined by points A and (-
1,-1), and the vector defined by point (A) and another point of the tree cloud (X). The angle
between these vectors (α) is then calculated for each point (X) of the point cloud. The point
(X) that leads to the largest angle (α) is then selected as the second point in the polygon. Other
points of the polygon are added similarly, only the vector defined by points A and (-1,-1) is
replaced by a vector given by the last two points of the polygon. This continues until the last
point is equal to the first point A.

Concave Planar Projection

The tree concave planar projection can be computed via Trees → Concave Planar Projection.
The tree cloud of interest is selected in the dialog window as well as the maximum edge
length, i.e. the threshold distance TD in cm. The tool calculates and displays the area of
orthogonal concave tree projection to the horizontal plane in the tree base height and the result
can be exported/imported to the ArcGIS polygon shapefile (see 9.2).
The calculation is based on the Gift wrapping algorithm and the modified Divide and Conquer
algorithm (Rosén et al., 2014). In the first step, the convex hull is computed similarly as in the
previous section. Then edges longer than the initial threshold/searching distance (TD) are
found. These edges are split in the next step and the algorithm looking for points inside the
polygon is as follows: A, B are adjacent points of the convex hull and X is a point tested to be
added in between them. If the distance AX is shorter then TD and BX < AB, or BX < TD and
AX < AB, then X is considered as a possible candidate. The edges created by split must be
shorter than the original one. If more than one point complies with these rules, the one with
the largest α AXB is selected. These conditions avoid too sharp angles in the new polygon.
Once the new point X was selected, the algorithm continues with edge AX or BX if one of
them is longer than TD. In the case that edges may not be split, because there are no points to
satisfy the split conditions, the algorithm increases TD by 10 cm and repeats the split
procedure. This continues until all edges are shorter than TD.

Figure 7: Planar projection of the same tree, convex (red), concave (blue).

Stem Curve
The stem diameters within 1 m intervals along the tree length can be received via Trees →
Stem Curvature. The stem diameters are computed as circles by Randomized Hough
transformation (see DBH by randomized Hough transformation) from 7 cm high slices of the
tree cloud. They are displayed as 7 cm high cylinders defined by the RHT fitted circles. The
number of RHT iterations is set by the user. The algorithm starts with the stem diameter at
0.65 m above the ground, then at 1.3 m and 2 m above the ground. Then it continues
computing diameters with 1 m spacing until the new diameter is two times wider than the
previous two diameters, i.e. the calculation is terminated when branches forming the crown
got involved in the calculation. Because numerous circles are fitted to each tree, the higher
number of RHT iterations significantly increases the computational time. Though, for higher
accuracy more iterations are recommended. The output of the stem curve calculation is x,y,z
coordinates of the centers of fitted circles and the diameters itself. See also Export Stem
Curve in the chapter Data Export.
 10. Crowns
The crown attributes (Fig. 8) can be evaluated if the tree position and tree height are
computed. Crown height, crown base height, crown width, crown centroid, and crown
position deviation are then computed automatically. These variables are recomputed
automatically if the tree position is changed.

Figure 8: A - The description of the crown attributes: crown height (CH), crown base height
(CBH), crown center (CC), crown total height (CTH), crown length (CL), crown width (CW).
B – The example of calculated parameters and the deviation of the projection of the crown
center (green) from the base of the trunk (blue). C – The concave planar projections and
volume/surface of the crown as calculated according to the concave polyhedron. D – The
convex planar projections and crown volume/surface calculated using a convex envelope. E –
The example of the visual output of the 3D Forest when calculating the shared space of crowns
(yellow) using 3D convex covers.

Crown Cloud Adjustment

The crown cloud can be adjusted manually via Crowns → Set manual or automatically via
Crowns → Set automatic. The dialog window gives an option to select the tree of interest.
The manual mode then enables to remove the stem points.
The automatic mode creates cross-sections of the tree along the z-axis with the height 0.5 m at
first. The width of each section is computed by averaging the x and y-axis values of the
respective section. The widths are successively compared from the lowest to the highest
section. If three consecutive sections are wider more than 25 % than the last thin section, the
last thin section is marked as a starting place for detailed search. The detailed search begins
by setting a subset of points with a height of 10 cm at the beginning of the thin section. Using
LSR the stem diameters at the start and the end of the subset of points are computed. These
diameters enable to predict the center of the next diameter, located another 10 cm along the
section. The diameter itself is then computed using only the points lying within the range of
the doubled value of the preceding diameter. This is to avoid the influence of overhanging
branches. Predicting the new diameter center and diameter evaluation continues until the
difference between the last two diameters is less than 25%. This point is then considered as
the base of the crown.

Crown Volume by Voxels

The crown volume by voxels is estimated via Crowns → Volume by voxels. The dialogue
window enables to select the tree and to set voxel size in cm. The volume is shown in the
project attribute table. Voxels are not visualized due to the computational demands.

Crown Volume and Surface by Concave Polyhedron

The crown volume and surface area by cross-sections are computed via Crowns → Volume
and surface by concave polyhedron. The dialogue window enables to select the tree of
interest, set the section height, and set the threshold distance for section concave hull. In case
that other than implicit values are set, the crown position and position deviation are
recomputed with external points extracted by these new parameters. The crown position is
calculated as an average coordinate from external points, i.e. the 2D concave hulls of
horizontal cross-sections (slices) with the threshold distance 1 m and section height 1 m; see
Concave Planar Projection. The external points can be visualized by the icon . Crown
volume is then computed as the total amount of sections volume, where each section volume
is computed by its planar projection area (created by concave hull) times section height. The
surface area is computed by strip triangulation. The algorithm is reliable, although there are
cases when intersecting triangles may occur. To show/hide resulting objects use icon . The
triangulation algorithm connects vertices of concave hulls of adjacent sections by the shortest
possible sum of edges.

Volume and surface by the 3D convex hull

The crown volume and surface area by 3D convex hull are accessible via Crowns → Compute
3D ConvexHull. The computation (3D Delaunay triangulation and mesh parameter
algorithms) is done using the VTK libraries. The dialogue window allows selecting the tree of
interest and offers the checkbox for computation using all crown points. The external points
and points belonging to the lowest and highest sections are used by default. This provides a
significant decrease in computational time with comparable results as if all crown points are
used. To show/hide resulting shapes use icon .

Crown intersections
The crown intersections evaluation is available via Crowns → Intersections. All existing
crowns are tested for intersections. The existing intersection between two crowns is computed
using VTK as Boolean AND in 3D space. The convex shapes are only possible for the
evaluation thus to create the 3D objects volume and surface by 3D convex hull have to be
computed first. The following variables of intersecting crowns are computed: horizontal angle
(azimuth), vertical angle (from horizontal plane), and distance from crown position to
intersection center of gravity, intersection volume, and surface area. These attributes are listed
in the intersections table, called by icon . Intersecting parts can be shown/hidden by icon
, the lines connecting crowns positions with intersection center of gravity, and angles are
also visible.
11. QSM Models and Tree assortments
The Quantitative Structure Models (QSM) menu offers methods to obtain information about
branches and stem (volume, length, branching order) and methods for tree analysis (other
quantitative parameters of trees). The QSM models may be employed if a tree position and
height are known.

Tree Reconstruction
Tree reconstruction is the first step of the QSM analysis.
The input parameters are the tree of interest, voxel size, and multiplicator that defines the
close neighbor, i.e. maximal distance of used voxels. The reconstruction starts by dividing the
whole tree into voxels. The points that are inside the voxel, as well as points and voxels that
are close neighbor (voxel size * multiplicator), are recognized for each voxel. The stem
reconstruction starts by neighbor search at the lowest voxel. All neighbors connected in one
piece into the homogeneous part are considered as segments. If neighbors create two or more
groups of non-connected voxels they are classified assible branches. The segment is closed at
that place and the analysis continues up to the tree length. Finally, the information about
voxels, ordering, i.e. parent, and child succession of segments, is known for each segment.
Once the whole tree is divided into the segments, the tree reconstruction itself follows. Firstly
all segments without children are said to be the ending segments. Parents of each of the
ending segments are followed to the lowest voxel and their lengths are computed. The stem is
the longest sequence of segments. These segments are merged into a single segment and all its
children are assigned. The reconstruction of branches of the first order follows with similar
semantics until it reaches the stem segment. After all first-order branches are reconstructed,
branches of second, third, etc. order follow until all segments are assigned to the tree.

A B C D

Figure 9: A – unprocessed tree point cloud, B – results of tree reconstruction: stem (red) and
branches (green to blue according to branching order), C – cylindrical model of a tree, D –
wood assortments.
QSM model
After the tree reconstruction, the parametric model of the tree can be evaluated. First of all the
tree of interest has to be specified. Since the cylinder estimation is done by randomized hough
transformation (as for DBH) the number of iterations has to be set then. The size of the
estimated cylinders defines the height of the cylinders used for reconstruction. Reconstruction
can be further restricted to the branches up to given order, given length, or given diameter.
The last option is the stem profile evaluation checkbox. This result can differ from those
given by the algorithms offered in the “tree menu”. The final results of the QSM
reconstruction appear as connected cylinders for each branch with an attribute table
containing information about volume, length, and order if each fitted cylinder.

Tree assortments
The final step is to estimate possible wood assortments (as the base for timber value
estimates) according to Czech industry standards for wood assortments. The only input is the
trees of interest. The evaluated attributes of the tree are convergence, skewness, diameter,
length, and number of connected branches. The algorithm classifies the whole tree
represented by cylinders into quality classes, higher quality classes are prioritized.

Table 2: Example of tree criterium for the classification of wood quality.

Class Convergence Skewness Diameter size Length Branches

1. up to 1 cm 1,5 cm/m 45 cm 3m absent

2. up to 1 cm 2 cm/m 28 cm 3m up to 3cm

3. 3 cm/m 20 cm 2,5 m up to 4 cm

4. 6 cm/m 7 cm 2m

5. 10 cm/m 7 cm 1m

6.
12. Data Export
All data exported from the 3D Forest are re-transformed into their original coordinates
system.

Clouds Export
The cloud data are exported via Project → Export. There is an option to choose an output
format (.txt, .ply or .pcd). Then the cloud for export, the name, and the location of the new file
are selected.

Tree/Crown Parameters Export

All calculated tree/crown parameters as a tabular output can be exported via Trees → Export
tree parameters or in case of the crown via Crown → Export crown parameters. The name,
the location, and the separator within the exported .txt file have to be specified. The
parameters that were not evaluated are exported as -1 value.

Tree Planar Projection Export

The planar projections of trees can be exported via Trees → Export convex planar projection
or Trees → Export concave planar projection. The name and the location of the new file have
to be selected. The planar projection is exported in one file for all trees. The projections that
were ́t computed previously during the ongoing 3D forest session are not exported. Both
functions create .txt file with coordinates of vertices of the polygon in the following format:
Tree ID;Xstart;Ystart;Xend;Yend;Z
Tips: the data import into ArcGIS and polygon or polyline shapefiles creation
Add an exported file to your project and use function XY TO LINE in ARCTOOLBOX →
DATA MANAGEMENT TOOLS → FEATURES. This creates a new layer containing
geodetic line features constructed based on input files. To create a closed polyline from this
layer use function UNSPLIT LINE in ARCTOOLBOX → DATA MANAGEMENT TOOLS
→ FEATURES. Afterward, it is possible to create polygons that are not overlapping to use
function FEATURE TO POLYGON in ARCTOOLBOX → DATA MANAGEMENT
TOOLS → FEATURES. To create overlapping polygons it is possible to use function
CONVERT POLYLINES TO POLYGONS from „Tools for Graphics and Shapes“ that is
freely available at http://www.jennessent.com/.
Figure 10: Polygons of the concave tree planar projections as imported in the ArcMap.

Stem Curve Export

The stem curve can be exported via Trees → Export Stem Curve. The dialogue window
allows specifying the tree of interest, the file name, and the destination folder. The exported
file is in the .txt format with the following structure:
Tree name diameter: 70.6 67.6 64.6 ..... -1 -1 -1
Tree name x: 182.474 182.5 182.455 ..... -1 -1 -1
Tree name y: -1594.55 -1594.56 -1594.57 ..... -1 -1 -1
Tree name z: 1017.6 1018.25 1018.95 ..... -1 -1 -1
Diameters and coordinates of centers of fitted circles are in ascending order of tree height, i.e.
from 0.65 to 1.3, 2, etc. The value -1 indicates data that were not evaluated.

Intersections Parameters Export

The parameters of intersections can be exported via Crowns → Export intersections
parameters. The dialogue window facilitates to set the name of the new file and the
destination folder. Parameters are exported as .txt file with the next structure: the tree name,
the name of the intersecting tree, the intersection volume (m^3), the surface area of the
intersection (m^2), the horizontal angle (°), the vertical angle (°), and the distance (m). The
semicolon is used as a separator.
13. Other Features
This menu comprises functions that are generally applicable to all cloud types.
Cloud Merge
This function helps to join two point clouds into a new one. It can be called via Other features
→ Cloud merge. In the dialog window select the clouds to be merged, set the name of the new
cloud, select its type (e.g. Tree, Base, etc.), and check the box if you want to delete the input
clouds.

Cloud Subtraction
The cloud subtraction is done via Other features → Cloud subtraction. The dialog window
offers the option to select the point clouds to be subtracted and set the name of the output
cloud. The tool subtracts the smaller cloud from the bigger cloud. Identical points of both
clouds are removed from the bigger cloud and the result is saved as a new cloud.

Voxelize Cloud
The reduction of the point cloud density is possible via Other features → Voxelize cloud. The
dialog window gives the option to select the cloud of interest, set the name of the voxelized
point cloud, and the size of the voxels in cm. The tool generates a voxelized point cloud from
the input by centroids of voxels that included at least one point of the original point cloud.

Create Convex or Concave 2D Hull

The convex/concave hull of any cloud can be created via Other features → Create convex
hull or Create concave hull. These functions evaluate the convex/concave hull of the
orthogonal projection of the cloud into the horizontal plane at the height of the lowest point
and save it as a new cloud (type Others). The same algorithms as described in the chapters
Convex planar projection and Concave planar projection are used. The tool thus creates a
layer with the same extent and boundaries as the area in 3D Forest. This may be further used
to transfer areas of interest from the 3D Forest to a GIS. The hull cloud should be exported as
.txt if imported into ArcMap for polyline or polygon creation, for more details see chapter
Data Export.

Remove duplicate points

The redundant points may be removed via Other features → Remove duplicate points. The
dialog window enables to select the cloud for filtering. The point cloud is divided into smaller
point clouds using octree search (octree resolution is 20 cm) and each cloud is separately
voxelized with 1 mm voxel size. After that, the clouds are merged back.
Save into tiff
The current view from the visualization area can be saved via Other features → Save into tiff.
Then set the name for the .tiff file and the destination folder.

Change Background-color
The background color of the viewer can be changed via Other features → Change
background color. Then select a new color.
14 References
Chernov, N. and C. Lesort (2003). Least squares fitting of circles and lines. arXiv preprint cs/0301001.
McDonald, J. (2014). The Hough Transform-Explained and Extended. [online] cited 17.7.2014. Available
at: www.cis.rit.edu/class/simg782.old/talkHough/HoughLecCircles.html.
Rosén, E., E Jansson and M. Brundin (2014). Implementation of a fast and efficient concave hull
algorithm. Available at: http://www.it.uu.se/edu/course/homepage/projektTDB/ht13/project10/
Project-10-report.pdf. Project Report. Uppsala University.
Rusu, R. B., S. Cousins and Ieee (2011). 3D is here: Point Cloud Library (PCL). 2011 Ieee International
Conference on Robotics and Automation (Icra).
Xu, L. and E. Oja (1993). Randomized Hough transform (RHT): basic mechanisms, algorithms, and
computational complexities. CVGIP: Image understanding 57(2): 131-154.
Statisticall outlier removal filter:
https://pcl.readthedocs.io/projects/tutorials/en/latest/statistical_outlier.html#statistical-outlier-
removal
Radius outlier removal filter:
https://pcl.readthedocs.io/projects/tutorials/en/latest/remove_outliers.html#remove-outliers