IMPORT TRAJECTORIES

Now we
will import trajectories. Browse to Manage Trajectories icon. . This can be
accessed either through Applications > Terr Scan > General or by c icking and dragging the first
icon in the TerraScan toolbox.

Figure 8: Opening the Trajectories window.

Goto File > Set directory to set directory for storing trajectory files.
 Figure 9: Setting the directory for Trajectories

Browse to a folder of your choice. To import trajectory files click File > Import files for importing trajectory files.

Figure 10: Importing the trajectory files

Select sbet_01.out and click Add. Click Done.

Figure 11: Importing trajectory files

An Import Trajectories window opens. Browse to UTM-17N (81W) in WGS84 field. Check the
Thin positions check box.
 Figure 12: Setting the import parameters for trajectories

Click OK. Now select the trajectory file named 403542_409436.trj in Trajectory window and click on Show
location to view the trajectories.

Figure 13: Displaying the location of trajectories

TRAJECTORIES
In the Trajectories window, go to Tools > Draw into design

Figure 14: Drawing trajectories

Click OK. This will display the trajectories.

 Figure 15: Drawn trajectories

Place a rectangle box so that we can trim the trajectory information. This rectangle tool is available in the
Microstation toolbar.

Figure 16: Creating a rectangle

Select the polygon/rectangle. In Trajectories window, goto Tools > Delete outside polygon

Figure 17: Deleting features outside the polygon

Click OK.
 Figure 18: Parameters for deletion

Click OK in the information window. Now click on the trajectories and press delete.

Figure 19: Deletion process

Select the outside polygon and press delete button.

Figure 20: Deletion process outside the polygon

Goto Tools > Draw into design. Click

OK in Draw trajectories window to view the trimmed trajectories.

Figure 21: Trimmed trajectories

READING DATA
Now we will import data into TerraScan. Click File > Read points in TerraScan.

Figure 23: Initiating the process for reading data points

Browse to the folder containing the raw laser data and select all LAS files. Click Add and then
OK.

Figure 24: Selecting data point files

In the Load Points window, do not apply transformation in WGS84 field. Also load every 10th point in the data and
uncheck the remaining fields.1

1 We load every 10th point from the given data so as to conserve memory for basic processing. Further, the total number of points
from all the LAS files is too large.
 Figure 25: Setting parameters for reading data points

Click OK.

CREATING A PROJECT
Now we will create a project. Click Define Project icon. Goto File > New Project

Figure 26: Defining a new project

In Description, enter Niagra. In Storage field, set the option to LAS binary. Check the Load class list automatically
and select niagara.ptc from the given data.
Check the load trajectories automatically and browse to the folder containing the trajectory files. Click OK.

Figure 27: Setting project parameters and related information

For saving project, goto File>Save Project As.

Figure 28: Saving the project

Save the project to a desired folder where the file name is Niagara. Click Save. Now we will add block to the project.
For this, select the boundary by clicking on it.

Figure 29: Selecting the boundary for importing data

Goto Block > Add by boundaries in Project window.

Figure 30: Adding the boundary to the project

In File prefix field, enter Niagara_

 Figure 31: Adding information blocks by boundaries

Click OK. Click OK in the Information box. In the Project window, go to File > Import points into project for
importing data points into project.

Figure 32: Importing data points into the project

Browse to the folder containing all the raw laser LAS files and select all the LAS files. Click
Add and then click Done.

Figure 33: Importing data points – Selection of LAS files

Set flight numbering to increase by file. This is done to display data corresponding to flightlines.
 Figure 34: Parameters for importing the point data into the project

Click OK. A report is displayed.

Figure 37: Report of importing data points into project

Close the report. Click File > Open block in TerraScan to view the data points in the project.

Figure 38: Opening block

Click inside the block to see the point cloud.

DATA DENSITY MEASUREMENT
Now we will measure point density of data. Click the Measure Point Density icon.

Figure 35: Density measurement icon

In Measure Point Density window, enter 1000 in Width field and select Rectangle in Sample field.

Figure 36: Setting density measurement parameters

Now just by mouse click in the data, you can see the point density at the bottom pane of
Microstation.

Figure 37: Measurement of point density

Right click the mouse to close the Point Density measurement window.

DATA VISUALIZATION
Now we will practice various modes of displaying data. Click View > Display mode. In Color
field, select by Elevation. Click coloring scheme button.

Figure 38: Coloring the point data by elevation

Click Auto fit button. Click OK. The point cloud is displayed with color scheme corresponding to elevation.

Figure 39: Points displayed and colored by elevation

Now display with color scheme corresponding to intensity

 Figure 40: Data points displayed colored by intensity

Now display with color scheme corresponding to flight line.

Figure 41: Points colored by flight line

We will now practice the display of various section functions. Goto Draw section . Draw a line and keep small
width of section for display.

Figure 42: Drawing a section

 Figure 43: Section plotted

Goto move section

button . Draw a section and then select viewer for display. Press
left/right click button of mouse to move the section forward or backward.

Figure 44: Moving a section

Goto cut section

. Place a section with a certain width and then click in a viewer to see the
cut section.
Figure 45: Cutting a section
 CLASSIFICATION

CLASSIFY POINTS TO BELOW GROUND CLASS

We will classify points to low points in below ground classification.

Low point by Low Point Routine

Classification of low points can also be done using low points routine. Select Classify>Routine>Low points>From
default To Class Low points. In search Groups of points, Type 3 in Max count, Type 1.00 in More than and 10.00 in
within>click Ok

Figure 22: Classification by low point routine (group of point search)

Low point routine is required to be run more than once. Classify>Routine>Low points>From default To Class Low
points, In search Single point , Type 0.5 in More than, Type 1.00 in More than and 7.00 in within>click Ok

Figure 23: Classification by low point routine (single point search)

Check the performance of Low point classification: Display mode>view 1, color by class, switch off all class
except low point (you may edit color, weight of low class points as well as speed for better display (using
colors>select Low point class>edit its color and weight).

Figure 24: Check the performance of low point routine; only low point class is displayed

Figure 25: Cross sectional view. Low points are grouped to a separate class, marked in red

CLASSIFY POINTS IN GROUND CLASS

Here we will try to classify points approximately in ground class from default class
Approximate ground class by Ground Routine
We will carry out ground classification by routine. The ground classification routine needs parameter for largest
building. We will determine this parameter from building having largest width.
Set view 1 to elevation coloring. Color auto fit, Ok>Search and zoom into largest building width >Use distance
measurement tool to measure building width(it ensures at least one of the lasers strikes the ground correspond to
every building block area).

Figure 20: Determination of maximum building size

As the largest distance found to be about 71 m , we will take the parameter to be 75 m and classify ground
Classify>Routine>Ground

Figure 21 Starting ground classification routine

Type say 75 m in max building size, Terrain angle 88deg, Iteration angle 6 deg., Iteration distance 1.40m, Check
Reduce iteration angle when Edge length<5.m >Ok

Figure 22: Feeding parameters to ground classification routine

Check the performance by viewing the ground class (Select color by class in view 1using display mode
window)>Switch off all the classes except ground class>Apply
 Figure 23: Checking performance of ground classification

Figure 30: Cross section of approximate ground class (in view 2)

Ground classification using above technique will not result pure ground class points. As classified ground points are
mixed with some below ground points as well.

CLASSIFY POINTS IN BELOW GROUND CLASS

Low point by Below Surface Routine
The below surface points mixed in ground class will be separated here. At the beginning a temporary class will be
created. Low points from this group will be classified in to temporary class using different trial parameters. Finally
low points will be separated from approximate ground class using best parameter. And temporary class points are
transferred to low class. (one can also classify low points directly from ground points).
Creation of additional class (Temporary), which is not listed
Display mode>Colors>Add>Insert the description (name of class), select the color and weight for
representation>Ok
Now we will classify low points of ground class to temporary class using below surface routine
Classify>Routine>Below Surface>From class Ground To class Temporary, Type 10 in limit>Ok
 Figure 31: Feeding below surface classification parameter

To view results: Switch to View>Display mode>Switch off all other Classes except Temporary;(To see all the
point-Select , Normal-All point in Speed>Apply)>Switch to class coloring>Draw sections to evaluate results.

Figure 32: Displaying below surface classified point from ground class

Figure 33: Checking the performance of below surface point classification in cross section

We will have to test above classification with other parameter values as well, to select best parameter. In order to do
that we need to undo below surface classification and then classify it with different parameter.
Select point>Undo below surface
(Note: Above technique of undoing is very useful to revert back to previous stages at the time of classification)
Classify>Routine>Below Surface>From Class Ground To class Temporary, Type 6 in limit-
>Ok
To view results: Switch to View>Display mode>Switch off all other Classes except Temporary; (To see all the
point-Select , Normal-All point in Speed>Apply)>Switch to class coloring>Draw sections to evaluate results.
Similarly we will test with few other limiting values before finally deciding the suitable parameter. Let us say 4
gives us the best results. Set it to 4 and classify low points from ground class using above parameter. After
classification of points to temporary class we need to Transfer Temporary class points in to Low point class
Classify>Routine>from class Temporary To class Low point>Ok (Entire class will be shifted to low point class.

CLASSIFY POINTS IN BELOW GROUND AND GROUND CLASS USING MACRO

Select Tools>Macro>Type macro detail (say) Abs_low_ground_below surface in description>add>Classify points
(action), Routine (By class)-By Absolute Elevation>Ok;(set range of elevation); >Ok.
Add>Classify points(action),Routine (By class)>Low points>Ok; From class Default To class low point; Search
Groups of points; Max count 3; Type 1(More than), 10 (within)>Ok.
Add>Classify points (action), Routine (By class)>Low points>Ok; From class Default To class low point; Search
single point; Type 0.5(More than), 7 (within)>Ok.
Add>Classify points (action), Routine (By class)>Ground>Ok; From class Default To class Ground; select settings
i.e. Building size size(say 115),Terrain angle, Iteration angle, Iteration distance>Ok
Add>Classify points(action), Routine (By class)>Ground>Ok; From class Ground To class low point; Set
settings>ok.
Save as with a name i.e Abs_low_grnd_blosurface.mac>Ok>Close macro window
We can directly run the macro from Tool>Macro>File-Browse the macro you have created >Select it>Run it on
loaded points (In case we have block(s) and you want to run it on specific block we will do the following. Start
define project tool>Select rows of blocks >Select Tools>Run macro (run macros on blocks)>Brows and select
Abs_low_grnd_blosurface.mac>Type (say) 70m neighbours>Click save points at the end>Ok
(We can check that the number of points classified. We can Save the report as .txt file>Close report>Close project
window).

CORRECTION TO APPROXIMATE GROUND CLASS POINTS

Ground class points may not still be very accurate. Many non ground points may be shown as ground point and vice
versa. This finer classification needs interactive classification using ortho photo and shaded surface.

Rectify the error points by classification

Tear out edit laser tool box

Figure 43: Edit laser tool box

There are different ways to carry out interactive classification using above tool box such as Assign point class,
Classify using brush, Classify fence etc. One can use many of these tools to classify error points.
Assign point class (2nd icon from left in figure 43)
Assign Point class classification>From any visible point, To class say low point/building/water etc; select closest
within 2 m
Classify using brush (3rd icon from left in figure 43)> From class Ground To class low point
/building/water etc; > apply it at different places as per your judgment.

Figure 44: Classify using brush parameter setting

Classify fence( 4th icon from left in figure 43)

Classify Fence classify points inside a polygon we draw> From class Ground To class low point /building/water
etc;> apply it at different places as per your judgment. You may need to draw a few sections to achieve this.

Figure 26: Classify fence parameter setting

 Figure 27: Classify fence; fence is drawn using polygon and then classified

Classify below line (6th icon from left in figure 43)

Classify below line, classifies points below a line in a section view (one may use these tools to classify point inside
building> From class Ground To class low point/water etc.
Repeat different classification scheme till we are happy with classification.
Classify inside fence
Classification can also be restricted inside any drawn fence. To do that Place polygons along the approximate line
you desire>Select resulting polygons using selection tool>select classify
>Inside fence (multiple selected can be classified at a time)>From class-any class to class say water/low point
etc>using display option we can check the performance of this classification.
Interactive Classification using Terra Model
We will create triangulated ground model and Visualize editable ground models as shaded surface. This surface
model will be used to improve ground points further.
Start create editable model tool>select ground class only

Figure 28: Create editable tool (first left)

Figure 29: Ground class point to be used for modeling

Start display shaded surface tool (6th icon from left in display surface tools) >Type Surface information Name say
ClassLiDAR>make sure only view 1 is on>Ok>Switch off all classes in view 1 (we are also generating BEM of the
area in this procedure)
 Figure 30: Display surface tools

Figure 50: Writing surface setting

Figure 51: Display shaded surface parameter setting

Figure 52: Shaded surface is displayed in view1 and ground class with ortho photo is displayed in view 2
 Figure 53: Shaded surface is effectively the BEM of the area

Now we will use shaded surface to rectify our ground class further. In BEM (or shaded surface model) strange
looking places are easily visible. This might involve erroneous point. We need to zoom inside such places and
rectify such points.

Figure 54: Strange appearing shades in BEM generated as non ground point is classified as ground

Search strange looking places in model>classify it if required (use different viewing windows and sections with
interactive tools i.e. Classify Point, Classify by Brush for classification)>We need to continue classification until we
are happy> when complete>Select> file save points.

Figure 31: Error points in ground is classified to water class using assign point class classification tool

CLASSIFY ABOVE SURFACE POINTS

Now we will try to classify above surface points.
Classify points in Building class using Building Routine
We will classify points to building class using Building Routine
Classify>Routine>From class Default class To class Building; Provide the minimum building size say 25 sqm >ok

Figure 56: Classify buildings routine

Buildings routine classifies building points which form some kind of a planar surface. This routine requires that you
have classified ground points before. After building classification we can check the performance of building
classification by displaying only buildings class points over ortho photo in view 2 (by switching off all the class
except buildings using display mode options) and/or using a sectional view of building class.

Figure 32: Classified building points displayed over ortho photo

Figure 33: Classified building points in sectional view

Classifying Low, Medium and High Vegetation using Height from Ground Routine in
Macro
We will classify different vegetation points using Macro constructed with height from ground routine.
Select tool from Terrascan main window>macro>Type vegetation in description

Figure 34: Defining a macro

Select add>Inside Macro step select action-classify points, Routine –By class>ok>classify by class, From class
Default—To class Low vegetation >Ok>Add>action classify point, Routine By height from ground>Ok>From
class Low vegetation To class Medium vegetation, min height 0.25 , max height 999 Ok>Add>action classify
point, Routine By height from ground>Ok>From class Medium vegetation To class High vegetation, min ht 2 ,
max height 999

Figure 60: Defining macro steps

Figure 61: Classify entire default class to low vegetation using classify by class

Figure 62: Classify low vegetation points to vegetations of different heights using by height from ground routine
 Figure 63: Setting parameters in classify by height from ground for low to medium vegetation

Figure 64: Setting parameters in classify by height from ground for medium to high vegetation

Save the file as vegetation. Mac macro.

Figure 65: Saving macro as vegetation.mac

Now we will run the macro vegetation.mac from tool. (it can also be run from macro window itself if it is not been
closed)
Tool>run macro>browse vegetation.mac> run it on loaded points >ok
 Figure 66: Runing a macro on loaded points

After running the macro it will generate a report. The report should be consistent with the data. So we can Check
report and then close it.

Figure 67: Macro execution report

We can now check the performance of vegetation point classification. A section can be drawn across the terrain for
this purpose.

Figure 68: Sectional view of classified vegetation points i.e. low (pink),medium(sky blue),high vegetation(in green color)

Majority of over below ground, ground and over ground points have already been classified. We can check
performance of these classifications using a section.
 Figure 69: Classified low points in a section

Figure 70: Classified ground points in a section

Figure 71: Classified low vegetation points in a section

Figure 72: Classified medium vegetation points in a section

Figure 73: Classified high vegetation points in a section

 Figure 74: Classified building points in a section

Figure 355: Classified low point, ground, low vegetation, medium vegetation, high vegetation and building points in a
section

Vegetation class data points may not be giving us the pure vegetation class it would have data points from utilities
such as tower, overhead wire etc along with few building points which could not be classified ( particularly, gable
roofs) properly during building classification. Hence we can classify further the high vegetation to isolate possible
power or transmission line points.

VIEWING PROJECT STATISTICS

We can now evaluate statistically the performance of entire classification using tool as below. Terra scan main
window> Tool >show statistics
 Figure 78: Statistics of classification

Saving classified data points

We need to save all the classified points to use it for future use such as Building modeling (to be covered in
Laboratory 4)
Terra Scan main window>File>Save points As.. lidar_school_classified
After saving your data points you can close data points, displayed surface, ortho photo etc.

FURTHER READING
Ground
Ground routine classifies ground points by iteratively building a triangulated surface model. The routine starts by
selecting some local low points as sure hits on the ground. You control initial point selection with the Max building
size parameter. If maximum building size is 60.0 m, the application can assume that any 60 by 60 m area will have
at least one hit on the ground (provided there are points around different parts of the area) and that the lowest point
is a ground hit.
The routine builds an initial model from selected low points. Triangles in this initial model are mostly below the
ground with only the vertices touching ground. The routine then starts molding the model upwards by iteratively
adding new laser points to it. Each added point makes the model follow ground surface more closely.
Iteration parameters determine how close a point must be to a triangle plane so that the point can be accepted to the
model. Iteration angle is the maximum angle between point, its projection on triangle plane and closest triangle
vertex. Iteration distance parameter makes sure that the iteration does not make big jumps upwards when triangles
are large. This helps to keep low buildings out of the model.

The smaller the Iteration angle, the less eager the routine is to follow changes in the point cloud (small undulations
in terrain or hits on low vegetation). Use a small angle (close to 4.0) in flat terrain and a bigger angle (close to 10.0)
in hilly terrain.

Setting: Effect:
From class Source class from which to classify points.
To class Target class to put points into.
Select Selection of initial ground points. Use Current ground
points when you want to continue iteration in a
fenced,
previously classified area.
Max building size Size of largest buildings.
Terrain angle Steepest allowed slope in ground terrain.
Iteration angle Maximum angle between point, its projection on
 triangle
plane and closest triangle vertex. Normally between 4.0
and 10.0 degrees.
Iteration distance Maximum distance from point to triangle plane during
iteration. Normally between 0.5 and 1.5 m.
Reduce iteration angle If on, reduce eagerness to add new points to ground
when inside
a triangle when every edge of triangle is shorter than
Edge
length. Helps to avoid adding unnecessary point
density
into ground model and reduces memory requirement.
Stop triangulation when If on, quit processing a triangle when every edge in
triangle is shorter than Edge length. Helps to avoid
adding unnecessary point density into ground model and
reduces memory requirement.
Building Classification
Parameters
Setting: Effect:
Ground class Class into which ground points have been classified
before.
From class Source class from which to search building points.
To class Target class to put points into.
Minimum size Smallest building plane size as square meters.
Maximum angle Steepest roof angle.
Elevation tolerance Elevation accuracy of laser points.
Maximum gap Maximum empty gap between points belonging to same
roof. Should be smaller than smallest gap between
adjacent buildings.

Model keypoints
Model keypoints routine classifies laser points which are needed to create a triangulated surface model of a given
accuracy. This routine is normally used to create a thinned data set from points classified as ground hits.
You control the accuracy with elevation tolerance settings Above model and Below model. These settings
determine the maximum allowed elevation differences from a ground laser point to a triangulated model. Above
model determines how much laser points can be above the model. Below model determines how much laser points
can be below the model. The application will try to find a relatively small set of points (= keypoints) which would
create a triangulated model of given accuracy. These keypoints will be classified into another class.
All other ground points are within the given elevation differences from a model that the keypoints would produce
when triangulated. Some of the ground points are above the model, some ground points are below. This
classification is an iterative process similar to ground classification. The process starts by searching for initial points
inside rectangular regions of a given size. The lowest and the highest source point inside each rectangle is classified
as keypoint and those are used to create an initial triangulated model. During each iteration loop the routine searches
for source points which are too far above or below the current model. If such points are found, the furthest points are
classified and added to the model. The Use points every setting provides a method for ensuring a minimum point
density in the final model even in flat places. For example, if you want to have at least a point every 10 meters, you
should set the Use points every setting to 10.0.

Setting: Effect:
From class Source class from which to search keypoints.
To class Target class to put keypoints into.
Use points every Rectangle size for searching initial points. The highest and
the lowest point inside each rectangle will be classified.
Above model Maximum allowed elevation difference to the triangulated
model from a source point above the model.
Below model Maximum allowed elevation difference to the triangulated
model from a source point below the model.

Intensity Routine Classification

By intensity routine classifies points which have an intensity (strength of return) value within a given range. The
type of surface material hit by the laser beam affects the intensity. Metal, white paint and green leaves typically give
high intensity returns. This can be used to classify points which are possible hits on railroad rails; the metal surface
gives a high intensity return and the surroundings often give a low intensity return.
Adding points to ground
It is generally advisable to run Ground classification using an Iteration angle which is too small rather than too
big. It is easier to add points to ground than to remove points from it. Ground terrain may have some small
formations which you want to reprocess locally. One way
to do this is using Add point to round menu command. It lets you identify an unclassified point which
belongs to ground, adds that to ground model and reprocesses the surroundings.

Setting: Effect:
Select Which point to choose around mouse click:
• Closest – Point closest to mouse click.
• Highest – Highest point within search circle. •Lowest
– Lowest point within search circle.
Within Search radius around mouse click.
From class Source class from which to take a point.
To class Ground class to add point into.
Reprocess Size of the area to reprocess around the added point.

Isolated points
Isolated points routine classifies points which do not have very many other points within a 3D search radius. This
routine is useful for finding isolated points up in the air or below the ground. When possibly classifying one point,
this routine will find how many neighbouring points there are within a gived 3D search radius. It will classify the
point if it does not have enough neighbours.

Setting: Effect:
From class Source class or classes from which to classify points.
To class Target class to put points into.
If fewer than Number of other points that must be found within a search
radius for the point not to be classified. Normally 1 - 5.
Within 3D search radius. Normally 2.0 - 10.0 m.