LIDAR

1
LIDAR Images

• Arttu Soininen
• Terrasolid Ltd

2
LIDAR Images
3
Setup

• Same flight for lasercamera or separate flights?
• Vertical or oblique facing camera?
• Vertical facing laser scanner or tilted?
• One camera or multiple cameras?

4
Why Fly Separate?

• We fly images and lidar separately because
• Aircraft mount has room for only one sensor.
• Camera has 60 and laser 28 field of view and
  we want to get the same coverage.
• Laser and camera have different requirements on
  conditions. We can fly laser during the night.

5
Advantages of Combined Flight

• Lower cost
• Images and LIDAR capture the same world
• Images help in processing decisions
• Images and LIDAR have the same positioning

6
Vertical View

• Vertical facing camera is best for orthophoto
  production
• Orthophoto is the world as seen from directly
  above
• Vertical facing camera is best for capturing xy
  locations of objects

7
Oblique View

• Forward / backward facing camera is good for
  seeing vertical objects
• Towers, trees, walls, ...

8
Vertical View
9
Oblique View
10
Typical Setup

• Single flight using
• Laser scanner 20 000 100 000 Hz
• 4K 4K digital camera
• Both vertical facing with -20 .. 20 field of
  view
• Data volume example for a 50 km² project
• 50 kHz scanner 900 million points 21 GB
• 2422 images (4K4K) 112 GB

11
Positioning of Images

• Raw data gives accurate xyz and orientation for
  each image
• Positioning is based on the same information as
  positioning of laser points
• Each airborne image has a time stamp
• GPS gives us time stamped xyz
• IMU gives us time stamped heading, roll and pitch
• Misalignment between IMU and camera is known
• Laser data gives a ground surface to rectify
  against
• We can orthorectify without any known points

12
Altitude Accuracy

• Higher altitude
• Elevation accuracy becomes a little weaker
• Point density decreases
• Spacing across track decreases linearly
• Spacing along track stays the same if same flight
  speed
• Image resolution decreases linearly

13
Aerial vs Corridor Projects

• Aerial sites are easier to handle
• Overlapping images on all sides
• ? Good internal control
• Short distance from one end to another
• ? More uniform geoid correction
• ? Less issues in matching to local control
  points
• Corridors are difficult to handle
• Limited overlap in images
• Long distance

14
Aerial Site Positional Inaccuracies

• Positional mismatches can be very easily seen in
  images in built-up environment

15
Aerial Site Positional Adjustments

• Collect tie points in images and run aerial
  triangulation

16
Corridor Heading

• Heading inaccuracies cause clear mismatches in
  consecutive images

17
Corridor Heading

• Heading inaccuracies cause clear mismatches in
  consecutive images

18
Corridor Roll

• One flight pass gives no roll control
• Images match each other if they are off in same
  direction

19
Corridor Pitch

• One flight pass gives no pitch control
• Images match each other if they are off in same
  direction

20
Ways to Use Images with Laser Data

• Support classification decisions
• Recognize vectorize objects on orthophoto
• Draw breaklines on ortho, drape to laser
• Vectorize 3D objects in perspective views
• Extract color values for laser points
• Visualization
• Final xy position correction

21
Support Classification Decisions

• Recognize objects in images when classifying
• View at the same time
• Orthophoto
• Laser points
• Laser cross section
• Shaded ground model

22
Recognize Vectorize Objects on Ortho

• Recognize a single point object in the orthophoto
• Place xy manually
• Software gives elevation from laser ground
  surface model

23
Recognize Vectorize Objects on Orthophoto
24
Recognize Vectorize Objects on Orthophoto
25
Draw on Ortho Drape on Laser

• Draw linear features in 2D on orthophoto
• Drape to follow laser surface

26
Perspective Views for Vectorization

• In a perspective view
• Viewer eye is located at the position of the
  camera when it recorded one image
• Viewer is looking in the same direction as the
  camera
• One raw image is overlaid with 3D vectors and/or
  laser points
• Can compare positioning of image and vectors even
  if the vectors are above the ground

27
Perspective viewsBuilding vectorization
28
Perspective viewsMonoscopic measurement

• Roof plane equation is known
• Camera orientation is known
• We can measure points on the plane using one raw
  image

29
Perspective View - Vertical
30
Perspective View - Oblique
31
Extract Color Values for Points

• Derive an RGB color for each laser point from
  orthophotos or from raw images
• Computes average color inside laser point
  footprint
• Use color values for classification or
  visualization

32
Visualization
33
(No Transcript)
34
Final Xy Position Correction

• Remove systematic bias in xy
• Images captured during same flight will have same
  positioning and can be used for xy matching at a
  finer level

35
Workflow Order
Laser automatic

• Run automatic ground classification
• Produce quick ortho images to support laser
  processing
• Edit ground manually with the help of ortho
  images
• Produce final ground model
• (Optional) Vectorize buildings for true ortho
• Produce final orthos

Quick ortho
Laser manual
Final ortho
36
Questions ???