Lidar and the Oregon Lidar Consortium

1
Lidar and the Oregon Lidar Consortium

• Portland State Office Building
• photo and lidar point cloud

Beaverton photo and lidar highest hit model
Eagle Creek landslides, abandoned railroad photo
and bare earth model
2
What is lidar (light detection and ranging)?-
Simply making lots of accurate distance
measurements with a laser rangefinder.

• Accurate laser rangefinders are commonly used as
  surveying instruments, measuring tapes, rifle
  scopes, even golf aids! Distance is calculated
  by measuring the time that a laser pulse takes
  to travel to and from an object.

3
Millions of very precise laser range measurements
are made from a precisely located aircraft,
producing an accurate and detailed 3-D map of the
earths surface, as a point cloud.
On-ground RTK-GPS base stations broadcast
corrections to airborne GPS unit, locating the
aircraft with an accuracy of a few centimeters.
Aircraft attitude is precisely measured by
Inertial Motion Unit, so that the exact position
and orientation of the laser rangefinder is
always known.
The rangefinder scans across the surface at
100,000 to 200,000 pulses per second, collecting
millions or billions of precise distance
measurements, which are converted to 3-D
coordinates.
4
Point cloud data define the 3-D shape and
location of the land, vegetation, and structures.
The complete collection of measured points for an
area is called the point cloud, which is the
fundamental form of lidar data. It provides a
very detailed and accurate 3-D map of ground
surface, vegetation, and structures.
(above) Animated point cloud image of the
Portland LDS Temple points are colored by their
relative height red highest, blue
lowest. (right) Photo of the same building note
statue on left hand spire, visible in both images.
5
Each laser pulse can produce multiple consecutive
measurements from reflections off several
surfaces in its path.
This provides detailed images of vegetation
structure and density, and returns data from the
ground under tree cover.
Image on the left is a point cloud lidar view of
the tree in the photo on the right. Each point
is colored by which return it was from a
particular pulse

• red 1st return
• yellow 2nd return
• green 3rd return

6
OLC data are collected at very high pulse
density, producing very detailed images.
Point cloud image on left compared to orthophoto
on right shows actual point density of lidar data
acquired over school bus lot. Each bus has been
measured by 180-200 lidar points!
7
The lidar point cloud can even image livestock
in the field!
Red and yellow clusters of points above ground
are cattle standing or lying in pasture.
8
For comparison, the best previously available
ground model is shown on the left. The 10-m USGS
Digital Elevation Model shows only a crude
representation of the real surface.
The lidar vendor uses a variety of software
filters to choose the points out of the point
cloud that measure the ground surface. In the
image on the right, vegetation points are green
and ground points yellow. Even in thick forest
there are numerous ground points.
Even if only one point in one hundred is a ground
point, the huge number of points means that a
smooth seamless ground model can be made. The
image on the left is a bare earth digital
elevation model, with 3 ft pixels, and reveals
incredible detail of the terrain beneath the
trees, including a hidden landslide.

• Very high point density means that even in
  heavily forested areas, it is still possible to
  get a large number of measurements of the ground.
  Left image is orthophoto of the Tualatin River,
  right image is lidar point cloud with red points
  high, blue points low.

9
Bare earth lidar can show features that you
cannot even see on the ground.

• Perspective view of lidar (Dec. 2007) on left
  matches photograph
• (July 2008) on right. The lidar was flown before
  clear cut logging of the reddish-brown slope, yet
  clearly shows an old logging road that is barely
  visible in the photograph.

Arrows connect matching locations.
10
Additional standard lidar products include a
highest hit or first returns model, which
shows the tops of trees and buildings, and an
intensity image, which is a form of infrared
photograph.
True color orthophoto with 0.5 ft pixels
Transmission lines
Nursery stock
Residence
Quarry
Auto
11
Lidar data allow a wide variety of information
about forests to be measured with unprecedented
accuracy and completeness.
Tallest tree at 252 ft!

• Locate and measure individual trees in forest
• Estimate fuel loads, carbon content, timber
  volume
• Tell conifer from deciduous
• Identify damaged forest

150-250 ft forest
60-100 ft forest
Image at right shows a simple analysis,
subtracting the bare earth surface model from a
first return (highest hit) surface model to
produce a canopy height model. Low canopy is
violet, high is red. The shapes of individual
trees are apparent, and the tallest tree in the
forest can be easily found and measured.
135-190 ft forest
100-125 ft forest
Brush and grass
40-60 ft forest
50-80 ft forest
12
Comparing the highest hit or surface and bare
earth surface provides a detailed and accurate
model of building area and height
13
The highly detailed bare earth model allows for
accurate location of roads and provides easy
access to unprecedented levels of detail about
slopes and shapes Yellow lines are best current
digital road map.
14
Stream channels are readily apparent on lidar
bare earth images
Comparing the lidar-derived streams with the
current digital stream map shows that the current
data are often wildly inaccurate Dark blue lines
are best current digital stream map, light blue
are lidar-derived.
GIS software can automatically find stream
channels from lidar data Blue lines are streams
generated by ArcGIS
15
In addition to accurately locating streams, lidar
easily produces accurate and detailed profiles
and sectionsLight blue line is lidar derived
stream location, dark blue are section lines.
16
What can you do with lidar?

• You can quickly, cheaply, and accurately.
• Find landslides, old cuts and grades
• Measure and estimate fills and cuts
• Find stream channels, measure gradients
• Measure the size and height of buildings, bridges
• Locate and measure every tree in the forest
• Characterize land cover
• Model floods, fire behavior
• Locate power lines and powerpoles
• Find archeological sites
• Map wetlands and impervious surfaces
• Define watersheds and viewsheds
• Model insolation and shading
• Map road center and sidelines
• Find law enforcement targets
• Map landforms and soils
• Assess property remotely
• Inventory carbon
• Monitor quarries, find abandoned mines

17
The Portland Lidar Consortium was the first large
scale effort to collect lidar in Oregon.
Hood to Coast survey
In 2006, the USGS provided DOGAMI with 100k to
complete the City of Portland. DOGAMI formed the
Portland Lidar Consortium to develop funding
partnerships to increase the area.
In 2003, the USGS funded DOGAMI for a pilot lidar
survey to look for earthquake faults.
With the USGS funds to anchor the survey and
ensure a large enough area for the lowest
possible rate, Federal, State and local
government agencies added on their areas of
interest until the entire project had grown to
over 2300 square miles and 1.1 M, with over 20
funding partners.
This was followed in 2005 by another USGS-DOGAMI
flight in the Portland Hills, a USGS survey of
the Columbia River Floodplain, and a survey by
Oregon City of its urban growth boundary.
18
The Oregon Lidar Consortium (OLC) originated in
2007 with a request by DOGAMI to the 2007
legislature for funds to acquire lidar over the
inhabited parts of Western Oregon.

• The legislature provided 1.5 M of the 4.5 M
  request and encouraged DOGAMI to seek funding
  partners to increase coverage
• The relatively small amount of funding requires
  prioritization to areas with significant local
  contributions

Blue hatch at left shows the original 4.5 M
target based on the inhabited area of Western
Oregon. Red hatch shows the area that could be
covered by 1.5 M, magenta shows existing data.
19
DOGAMI Business Plan for the OLC

• Collection areas should be large and contiguous
• Collection areas initially anchored by
  significant contribution from local funding
  partner
• OLC builds on anchor funding by finding
  additional partners
• State funds used to knit together partner areas
• State funds are spent on the inhabited areas of
  the state
• Collection areas completely outside the inhabited
  areas are fine if fully partner funded
• Data in public domain
• Where possible, collection areas should include
  entire 6th field watersheds

20
DOGAMI selected a vendor to provide lidar to the
consortium.

• A nationwide RFP led to the selection for
  Watershed Sciences Inc. of Corvallis, Oregon, as
  the lidar vendor for the consortium under Oregon
  Price Agreement 8865.

Watershed Sciences
21
OLC lidar prices are a function of area. DOGAMI
adds 10 to the vendor price for quality control
and management.
22
Data Specifications

• Laser spot size on ground 15-40 cm

1m

• Aggregate pulse density gt 8/m2

• Absolute accuracy of each point
• 20 cm horizontal and vertical

• 50 sidelap for complete double coverage

1m

• Swath to swath consistency 15 cm (same point
  measured by adjacent swaths must have similar
  value)

Point cloud image of field and building red
points are from one swath, blue from another.
23
DOGAMI provides three-way independent quality
control for OLC data.
OLC lidar image showing DOGAMI quality control
points (red triangles) collected by RTK-GPS
survey.

1. Compare accurately surveyed control points to the
   final lidar product to test absolute accuracy
   (/- 20 cm).

Colors indicate data from different swaths
Bird anomalies produce spikes in bare earth
model

1. Compare adjacent points from overlapping swaths
   to test consistency (/- 15 cm)

GPS elevation 50.40 m Lidar elevation 50.38
m Error 2 cm

1. Inspect bare earth models for artifacts,
   processing errors

Lidar production software is used to
automatically compare locations for huge numbers
of points from overlapping swaths.
Swath to swath differences, measured on hundreds
of thousands of points per swath, average about 3
cm in this example
24
OLC Data Products
3 ft pixel first return DEM ESRI format (quad
tiles)
3 ft pixel bare earth DEM ESRI format (quad
tiles)
Point cloud, LAS format 1/100 quad tiles
1 ft pixel intensity images (1/4 quad tiles)
Ground points in LAS format (1/100 quad tiles)
Aircraft trajectories

• Report and metadata !!

25
Data Distribution Options
Funding partners
Copies are provided on external hard drives as
soon as DOGAMI completes QC.
Public

• NOAA LDART website (point cloud)
• USGS CLICK website (point cloud)
• USGS NED website (DEM)
• GEO spatial data library website (DEM)
• METRO (Portland area only)
• PSLC (Hood to Coast area only)
• DOGAMI website (planned)
• DOGAMI publications on disk or drive (planned)

26
As of September 2008, 25 partners have added 2.7
million to the Oregon Legislatures 1.5 million.
The City of Turner
http//www.lincolncity.org/Portals/29/logo2.gif
27
As of September 2008, the OLC has been successful
in building partnerships for several lidar
collections around the state.Current status at
http//www.oregongeology.com/sub/projects/olc/defa
ult.htm
28
Future Plans

• DOGAMI is seeking funds in future biennia to
  extend coverage to other parts of the state, as
  illustrated in the conceptual draft below.