GEOTECH 2004 Workshop 3 LIDAR Bob Ryan, CP, PLS

1
GEOTECH 2004Workshop 3 LIDARBob Ryan, CP, PLS
2
(No Transcript)
3
Presentation Overview

• Lidar Theory
• Lidar Systems
• Data Applications

4
What LIDAR Is

• LIght Detection And Ranging
• Active Sensing System
• Uses its own energy source, not reflected natural
  or naturally emitted radiation.
• Day or Night operation.
• Ranging of the reflecting object based on time
  difference between emission and reflection.
• Direct acquisition of terrain information,
  whereas photogrammetry is inferential.

5
What LIDAR Is Not

• NOT Light/Laser Assisted RADAR
• RADAR uses electro-magnetic (EM) energy in the
  radio frequency range LIDAR does not.
• NOT all-weather
• The target MUST be visible. Some haze is
  manageable, but fog is not.
• NOT able to see through trees
• LIDAR sees around trees, not through them. Fully
  closed canopies (rain forests) cannot be
  penetrated.
• NOT a Substitute for Photography
• For MOST users, LIDAR intensity images are NOT
  viable replacements for conventional or digital
  imagery.

6
LIDAR Characteristics

• Vertical accuracy for commercial applications at
  15 cm on discrete points
• Capable of collecting millions of elevation
  points per hour much faster than traditional
  methods
• Produces datasets with much greater density than
  traditional mapping
• Systems capable of capturing single, or multiple
  returns per pulse and/or intensity images
• Supported by rigorous QA/QC similar to
  traditional surveying principals

7
Viable Through Combined Technologies

• Several enabling technological advances have made
  LIDAR possible
• Airborne GPS
• Inertial Measurement
• Availability of affordable lasers and other
  specialized materials and sensors
• Declassified military technology
• Advances in computer technology (speed,
  performance, size, and of course, price)

8
LIDAR Instrumentation

• Laser Source and Detector
• Scanning mechanism controller
• Airborne GPS (location, speed, direction)
• Inertial Measurement Unit (roll, pitch, and yaw)
• High-accuracy, high-resolution clock for timing
  laser emissions, reflections, GPS, IMU and scan
  angle measurements
• High Performance Computers
• High Capacity Data Recorders

9
LIDAR System
Supporting electronics
Laser Detector
Head Unit
IMU Controller
Pulsing Laser
Removable Hard Drives
IMU
Scan Mirror Controller
Scan Motor
Operator controls
LIDAR System Controller
Scan Mirror (internal)
Laser Power Supply
Power inverter
10
LIDAR Operational Theory

• A pulse of laser light is emitted and the precise
  time is recorded.
• The reflection of that pulse from the surface is
  detected and the precise time is recorded.
• Using the constant speed of light, the time
  difference between the emission and the
  reflection can be converted into a slant range
  distance.
• Knowing the position and orientation of the
  sensor, the XYZ coordinate of the reflective
  surface can be calculated.

11
Electro-Magnetic Energy

• EM energy can be conceptualized in two ways
• As waves of electrical and magnetic energy as
  the frequency of the crests increases, the
  wavelength decreases
• As mass-less particles called photons which
  travel with energy and momentum
• EM energy is continuously emitted from anything
  whose temperature is above absolute zero
• All EM energy travels at the same constant speed
  186,282 miles per second, in a vacuum.
• EM energy can be absorbed, scattered, reflected,
  or transmitted

12
The Electro-Magnetic Spectrum
13
The Electro-Magnetic Spectrum
Passive Microwave
Film
Active RADAR
Electro-optical Sensors
Thermal IR
Typical Terrestrial LIDAR Laser
Grayed sections indicate significant bands of
water or atmospheric absorption
14
Lasers

• Device which generates a stream of high energy
  particles (photons), usually within an extremely
  narrow range of radiated wavelengths
• Produces a coherent light source
• Wide diversity of power and wavelengths
• CD players and Pointers
• LIDAR units
• Industrial cutting tools
• Weaponry

15
NIR Reflectivity Examples

• White Paper up to 100
• Snow 80-90
• Beer Foam 88
• Toilet Paper 60
• Deciduous Trees 60
• Coniferous Trees 30
• Dry Sand 57
• Wet Sand 41
• Asphalt with Pebbles 17
• Black Neoprene 5
• Clear Water lt 5

16
GPS The Driver for Precision Elevation Data
17
Airborne GPS

• Satellite-based radio-navigation system put in
  place by the U.S. Department of Defense
• 21 satellites 3 spares
• 6 orbital planes
• altitude of 20,200 km
• Uses triangulation from multiple satellites to
  provide accurate 3D positioning
• Satellites transmit position and time information
• Receiver uses information to compute range
• Measurements from 4 satellites allows computation
  of 3D position.

18
Differential GPS

• Improves the accuracy of the airborne GPS by
  providing additional GPS measurements from a
  known local reference point The Base Station.
• Variations in the measured GPS location at the
  Base Station and its surveyed location are
  recorded and timed during the mission.
• These differences are applied to the airborne GPS
  data during LIDAR Pre-processing.
• Accuracies of 3 to 4 cm in X, Y and Z are
  obtainable.

19
Orientation Information

• Accurate determination of the reflected point
  location requires precise information on aircraft
  attitude
• Need the rotation around 3 axes of the aircraft
  roll, pitch, and yaw
• Two techniques
• Multiple GPS Receivers
• Less accurate
• Less common
• Helicopter installations
• Inertial Measurement Units
• More accurate
• More common

20
Inertial Measurement Unit

• 3 Accelerometers, 3 Gyros, and signal processing
  electronics
• Digital output high-accuracy acceleration and
  angular rate (accuracies of 18-25 arc seconds)
• Update speeds up to 200 Hz
• All measurements logged using a common system
  clock
• Hard mounted to the sensor head unit
• Sometimes integrated with the GPS system

21
LIDAR Accuracy Components

• Instrument Error Budgets
• ABGPS precision
• IMU precision
• System noise floor
• Timing resolution
• Mechanical tolerances (temperature and pressure
  variations)
• Atmospheric distortions

22
Terrain Slope and Accuracy

• As slope increases, horizontal uncertainty has a
  greater impact on vertical accuracy
• LIDAR pulse footprint becomes distorted
• Laser footprint spread across a greater range of
  elevations

23
Slope Effects
24
LIDAR in Extreme Terrain

• Because LIDAR is and ACTIVE sensor, sunlight
  shadows (or nighttime) are not concerns
• This allows surface collection where normal
  techniques fail
• And provides more flexibility in planning
  missions
• However, LIDAR shadowing can occur if the project
  is not designed with due regard for the terrain

25
LIDAR Shadowing
In extreme terrain and urban environments

• Mission planning considerations
• Use smaller FOV angle keep the beam more
  perpendicular
• Flight line configuration for best LIDAR
  line-of-sight
• Keep critical areas closer to nadir
• Post-processing considerations
• Data combining from adjacent flightlines in
  overlap areas

26
Deep Shadows - Grand Canyon
Traditional mapping in this area would not be
possible
27
Deep Shadows - Grand Canyon
Traditional mapping in this area would not be
possible
but a properly planned LIDAR collection is
unimpeded
28
Right from the Start

• The quality of the final product cannot be better
  than the worst of its components.
• Garbage In Garbage Out
• Missions must be planned carefully.
• Surveyed Ground Control will be required.
• Establish Redundant Base Stations on Valid Survey
  Points.
• Avoid LIDAR Acquisition Greater than 50 Km from
  the Base Station.

29
Typical LIDAR Mission Flight Plan
30
Verification of Coverage

• Inspect for data gaps caused by
• shadowing
• flight plan deviation
• turbulence
• instrument failure
• computer failure
• Quickly pre-process to produce geo-referenced
  data sets decimated at 10
• Convert vector points to raster for speed
• Display with GIS/CAD software
• If necessary, re-fly BEFORE the flight crew
  leaves the project site!

31
On-site Verification of Coverage
32
9 out of 10 squirrels do not like Chicago beer!
33
LIDAR Pre-Processing
Positioning the Laser Point

• Straight-Forward Geometry Problem
• Calculate the 3D angle and XYZ origin of each
  pulse-reflection vector from the IMU, GPS, and
  Scan Angle data.
• Calculate the length of the vector from the
  reflection timing and the speed of light.
• Calculate the target XYZ location.
• Repeat until all 87 gazillion points have been
  calculated

34
Positional Data Pre-Processing

• Forward Time Processing
• Estimate Data Prior or at the Time-of-Validity(Tof
  V)
• Differential Correction of GPS Data, Rigorous
  Strapdown Computations of IMU Data, Blend the GPS
  and IMU Data with a Kalman Filter
• Backward Time Processing
• Incorporates GPS/IMU Measurements that occur
  after the TofV
• Smoothing improves the estimated values
• Convert Data to Local Coordinate System
• Output Orientation Angles (roll, pitch, heading -
  ?, ?, ?)

35
Horizontal Coordinate Systems

• ORIGINAL LIDAR/GPS measurements are usually in
  Lat-Long coordinates
• The positional information for each point is
  translated into the Cartesian (X-Y) coordinate
  system required by the client
• Typically this is a defined State Plane
  coordinate system or UTM, referenced to NAD27,
  NAD83, or HARN
• Units are typically U.S. Survey Feet,
  International Feet, or Meters

36
Vertical Coordinate Systems

• Elevation
• Collected in GPS using Ellipsoid elevation,
  typically in Meters
• Translated into feet or meters as required by the
  client
• Re-referenced to the more accurate geoid
  elevation
• NGVD29 or NAVD88 are most common

37
Boresight Validation

• Whats a boresight?
• GPS antenna, IMU center of mass, and the LiDAR
  detector are all at physically different
  locations
• We do not really care what the position of the
  antenna or attitude of the IMU is rather, we
  use these to calculate a precise offset to
  determine the exact location and orientation of
  the LiDAR detector
• Shock of landing and other environmental changes
  can shift the position of the instruments enough
  to upset data.

38
Boresight Validation

• Performed for each flight mission
• Removes systematic errors
• Flight-to-flight variations on instrument
  mounting
• Environmentally induced changes
• Performing a kinematic GPS survey on a roadway
  within the collection area is a reliable check
• Validate the vertical for a feature with first
  returns on an unobstructed surface
• Apply any vertical correction as necessary (a
  z-bump)

39
LIDAR Boresight Calibration Survey
40
The Importance of Boresight Calibration

• Control points on a building and other regularly
  shaped surfaces are acquired
• Results of calibration flights compared to
  control points to correct for
• edge curl
• pitch
• timing issues (GPS, IMU)
• other system biases
• If the instrument is not dialed-in for each
  mission, data from successive lifts will not
  match!!

41
Quality Assurance

• Validate the Boresite
• Validate the Accuracy with Spot Checks on Known
  Points within the Project
• May require ground survey
• Prepare a Survey Report for IMU Data
• Similar to a Traditional AT Report
• Express Design Accuracy and Results Achieved

42
Elevation Data Standards

• NMAS
• 90 of points must be within ½ of the contour
  interval
• ASPRS
• RMSE (68.27) of points must be within 1/3 of the
  contour interval
• NSSDA The NSSDA was developed to report accuracy
  of digital geospatial data that is not
  constrained by scale.
• Data accuracy is reported at the 95 confidence
  interval
• Not tied to CI or map scale

43
LIDAR Accuracy Standards
ASPRS Guidelines Vertical Accuracy Reporting for
Lidar Data Version 1.0 Drafted May 15,
2004 Released May 24, 2004 Ownership ASPRS Lidar
Committee (PAD) Editor Martin Flood Reference
Standards These ASPRS guidelines are harmonized
with the relevant sections of the Guidelines for
Digital Elevation Data (Version 1.0) released by
the National Digital Elevation Program (NDEP).
The sections on vertical accuracy testing and
reporting from the NDEP guidelines have been
submitted to the Federal Geographic Data
Committee (FGDC) for inclusion as approved
revisions to the National Standard for Spatial
Data Accuracy (NSSDA). The NDEP guidelines are
available online at www.ndep.gov.
44
NDEP Vertical Accuracy Standards
45
NDEP Horizontal Accuracy Standards
46
Field Survey QC Points

• A minimum of 20 checkpoints shall be tested,
  distributed to reflect the geographic area of
  interest and the distribution of error in the
  dataset. When 20 points are tested, the 95
  percent confidence level allows one point to fail
  the threshold given in product specifications.
• Open terrain (sand, rock, dirt, plowed fields,
  lawns, golf courses).
• Tall weeds and crops.
• Brush lands and low trees.
• Forested areas fully covered by trees.
• Urban areas with dense man-made structures.

47
MultipleReturn LIDAR
48
FIRST PULSE Return

• First-pulse
• Measures the range to the first object
  encountered - in this illustration, the tree
  foliage.

49
LAST PULSE Return

• Last-pulse
• Measures the range to the last object - in this
  case, the ground.
• By acquiring first- and last-pulse data
  simultaneously, it is possible to measure both
  tree-heights and the topography of the ground
  beneath in a single pass.

50
Discrete Multiple Return LIDAR
51
Discrete Multiple Return LIDAR
52
LIDARCanopyPenetration
53
LIDAR vs. Stereo Compilation1100 Scale Mapping

• Compiled Mass Points are more widely spaced 60
  feet vs. 12 feet
• Compiled DTMs use breaklines LIDAR usually does
  not
• Compiler can place points LIDAR is semi-random
• Compiler must be able to SEE THE GROUND LIDAR is
  self-illuminating

54
Ground Points from Single Returns
55
Ground Points - Single LAST Returns
56
Wooded Area Example
57
Detail with LIDAR ground points
58
Final TIN surface
59
Canopy Penetration Bottom Line

• Multiple-return LIDAR systems will provide more
  points on the ground than either single-return
  systems OR conventional photogrammetric
  collection.
• Surface models for wooded areas are more reliable
  and detailed.
• Areas with extremely dense canopy which do not
  have adequate LIDAR penetration can still be
  delineated as Obscured Areas, just as in
  conventional mapping.

60
Theres more than one way to skin a cat
A quick look at Three different LIDAR systems
61
Sinusoidal (sawtooth) Scanner
Oscillating Mirror Scan Pattern
62
Rotating (circular) Scanner
Rotating Mirror Scan Pattern
63
Glass Fiber Scanner
Bundled Fiber Scan Pattern
64
LIDAR Technical Considerations

• Scan Angle (FOV)
• Fixed or Variable
• FOV appropriate for intended operational altitude
  and pulse rate
• Laser Pulse Rate
• Fixed or Variable
• Will impact operational envelope relative to
  desired post spacing
• Can impact maximum operational altitude

65
LIDAR Technical Considerations(cont.)

• Return Per Pulse Detection
• Single
• 1st Generation units
• Limited vegetation penetration
• First-Last
• Superior vegetation penetration ground
  detection
• Discrete Multiple
• Allows advanced analysis of vegetation structure
• Widest range of applications

66
LIDAR Intensity Imagery

• Return Intensity Capture
• Provides data for correction of systematic errors
• Can provide useful image information
• Careful !! --- This may be promoted as Black
  White photography --- IT ISNT !!

67
(No Transcript)
68
REDNECK HORSESHOES
69
Spacing of LIDAR Points

• Sometimes referred to as the average or nominal
  post spacing
• For scanning mirror systems
• Along the flight line track spacing
• Along the scan line cross track spacing
• Sometimes referred to as the point density per
  square unit
• e.g. points per square mile
• QuestionIs more better?

70
Post Spacing No. of Points
Points Per Tile are based on full tiles
measuring 20,000 x 20,000 (14.4 sq. mi.)
71
Optech ALTM System

• Up to 50,000 pulses per second
• First Last Returns
• Intensity measurement
• Many models to choose from

72
Leica ALS40 System

• 50,000 kHz pulse rate
• 0-45 degree FOV
• 33 returns per pulse, each with intensity

73
DATIS II System

• 50,000 pulses per second
• Five discrete returns
• Intensity measurement capability - digital
  imagery option
• Designed for portability

74
TopoSys Falcon

• 83,000 kHz pulse rate
• 14.3 degree FOV
• Discreet multiple returns

75
LIDAR Systems

• Approx. 60 Systems in Operation Worldwide
• 34 systems operated from North America
• Major Manufacturers
• Optech, TopEye, Leica (78)
• Several Offer Lease Options
• Several Custom/Proprietary Systems
• Calibration Methods and Software Variable

76
Value-Added Processing

• Large number of companies process LIDAR
• Many more than LIDAR data collectors
• Produce full range of services
• Surface generation, other sensor integration
• Proprietary software common
• Often embedded in an existing GIS/CAD package
• Some off-the-shelf software is available
• Software to compliment photogrammetric systems

77
LIDAR Consulting Services

• Some Firms Offer Consulting Only
• System Design, Development Construction
• Project Planning
• Application Development
• Sensor Integration
• Business Development Strategic Planning

78
Business Models

• Consortiums are Common
• Joint Ventures, Strategic Partnerships
• Typically Share the Same LIDAR System and
  Processing Capabilities
• LIDAR Services as Part of a Larger Corporations
  Business
• Some Firms Simply Request Services from Vendors
  As Needed
• Some Firms Lease Systems

79
Higher Performance LIDAR Systems

• Faster Pulsing and More Reliable
• Improved IMU
• Recording of Intensity Data for Multiple Returns
• Recording Waveform Data
• Increased Computing Power
• Improved Post-Processing Software
• More Intelligent Software Algorithms
• Fusion with Photogrammetry for QC

80
Integration With Other Sensors

• Aerial Cameras, Both Traditional Digital
• Multi-spectral, Hyper-spectral, Thermal Sensors
  Satellite Imagery
• Water-Penetrating LIDAR systems
• Various Types of RADAR

81
Increased Capability andIndustry Competition

• More Systems Reduced Mobilization Costs
• Firms Developing Niche Markets
• Off-the-Shelf LIDAR Data Sets
• Definition of the Types of LIDAR Service
  Available from Specific Vendors
• Similar to Traditional Mapping Today

82
Increased Market Demand

• A rapidly maturing technology and increased
  number of available systems has led to increased
  demand
• LIDAR has taken its place with other established
  geospatial mapping technologies
• New applications and solutions will continue to
  push demand for services

83
Lidar Techniques
84
Data Visualization

• LIDAR data is visually meaningless and impossible
  to understand in its raw XYZ form.
• More readily interpretable visualizations are
  needed
• TINs
• GRIDs
• 3D Perspective Renderings
• !! BUT !! Visualizations can be misleading and
  must be managed carefully

85
LIDAR Points
86
Simple Elevation Raster
Rasters offer speed and flexible display but
often lose detail
87
LIDAR TIN
TINs provide much more detail but they can be
slow to render
88
Hillshaded GRID
Hillshaded rasters are often the most
useful compromise.
89
Exaggerated GRID
Vertical Exaggeration can help in low
relief areas, but must be used cautiously!
90
LIDARDataManagement
91
Data Handling - then
92
Some Perspective
93
Data Handling - now
94
LIDAR Data Volumes

• Hardware/Software upgrades are often required
  (weve done it twice!)
• End users need to be aware of the size of the
  data
• The number of points can be reduced by
• Trimming overlap from between adjacent
  flightlines
• Clipping data to the limits of the site or
  project area
• Tiling the data into manageable size files
• Point Thinning (special algorithms)
• Rasterization of the point data (loses precision)

95
This didnt happen last yearand it aint
happening this year!
96
Vegetation RemovalFeature Extraction
97
Classification

• A LIDAR return unfortunately does not know what
  surface or material it was reflected from.
• Intensity cannot provide this answer IR
  reflects well from both healthy vegetation and
  snow, and it is absorbed by both fresh asphalt
  and water.
• For an elevation surface to be meaningful, it
  must represent a homogeneous set of features
• FIRST Reflective surface
• Bare-Earth only
• Bare-Earth and man-made structures

98
Classification

• Classification is the process by which LIDAR
  points are identified and grouped together so
  that meaningful datasets can be extracted.
• Vegetation Removal and Feature Extraction are
  types of classification
• Using the database structure of the GIS,
  individual LIDAR points can be tagged with their
  class for later and extraction
• Open Ground
• Tree Canopy
• Brush
• Buildings
• Water
• Noise

99
Classification
100
Identifying and Isolating the Ground Surface

• Use of Multiple Returns
• Automatic Classification
• Manual Editing

101
Raw FIRST Return LIDAR Data
102
Raw LAST Return LIDAR Data
103
Automatic Classification

• Clearly the volumes of data involved in a LIDAR
  project demand that this task be in large part
  automated.
• Manual editing is VERY labor-intensive there is
  tremendous pressure to do more and more
  automatically
• Three factors have broad impact on this process
• LIDAR Point Density (Nominal Post Spacing)
• Terrain Type Flat, Rolling, Rugged
• Land Cover Type Wooded, Urban, Marsh
• Both raster- and vector-based algorithms are
  employed
• The use of raster-based algorithms does NOT
  require that the LIDAR data be in raster format.

104
Automatic Classification (cont.)

• Most LIDAR software (both proprietary and OTS)
  incorporate some means of addressing these
  factors, though the labeling may not be
  apparent.
• Best results are obtained when processing
  homogeneous datasets (but that isnt the real
  world)
• More adaptive processing software is evolving
  and will result in significant increases in
  accuracy while reducing the processing time.
• Greatest advantage to the user community that
  has evolved over the past 3-4 yearsconcensus
  that there is no fully automated system for
  producing a bare earth lidar surface

105
Find The Ground
106
So how does it work?

• Morphology A study of structure or form.
  -Websters dictionary
• Morphological Analysis Using a localized
  window on the dataset, the program determines
  what general structures exist and which points
  are part of which form.
• The structure most commonly sought is the ground
  buildings and vegetation are the most common
  interferences.
• These are all readily identified by the human eye
  and mind, but the computer is a tougher client
• Trend-Surface Analysis

107
Technically Speaking...

• Raster-World
• Create a raster of the data based on the lowest
  point within each pixel
• Run a low-pass filter over the raster (smooth it
  out some)
• Filter the LIDAR points against the raster
  Points lower than the surface are ground those
  above it are not.
• Vector-World
• Build a TIN from the LIDAR points, BUT
• Ignore those points which would make any facet
  too steep
• Those points which are part of the TIN are
  ground those that were omitted are not.

108
Issues

• Parameter setting is tricky
• Settings too aggressive can remove subtle ground
  features like berms, hilltops, small ditches
• Settings too conservative leaves too much work
  for manual editing ---
• Need more, or more effective, adaptive algorithms
• Large features like buildings can be
  misclassified as groundor trees
• Very processor-intensive

109
Automatic Vegetation Removal

• Automatic programs begin the noise and vegetation
  removal process
• These remove approximately 80 - 90 of
  vegetation (depending on the land cover and
  terrain characteristics)
• This typically uses about 20 of the vegetation
  removal time budget

110
Before
111
...after
112
LiDAR Software

• TerraScan
• Add-on to MicroStation
• LiDAR data all held in memory NOT subject to
  MicroSataion file size limitations, only limited
  by amount of RAM
• Robust data management and automatic
  classification algorithms
• Flexible display options Plan, profile,
  perspective
• ESRI
• No dedicated toolset yet
• Data volume will likely require SDE
• Improved surface rendering is in the works

113
DEMs and Derived Products
114
Contour Generation

• Contours are a cartographic construct used to
  visualize topography. They have ALWAYS
  sacrificed accuracy in favor of appearance.
• Conventional DTMs are a combination of masspoints
  and breaklines.
• LIDAR does not have breaklines Contours will
  behave differently at feature edges.
• Users need to recognize the respective strengths
  and weaknesses of the two products. They are not
  the same!

115
Contour Generation

• LIDAR points are much denser (typically 21 times
  as many mass points as conventional DTMs) more
  detail is rendered, particularly in low-relief
  areas.
• Contours produced directly from LIDAR are NOT
  aesthetically pleasing.
• Either the LIDAR data or the resulting contours
  must be generalized before an acceptable product
  can be made.
• Point/TIN Thinning
• DEM Interpolation
• Contour Smoothing

116
DEM Resolution Comparison
117
Contours generated from the DSM
118
Contours generated from the DEM
119
Lets Discuss DEMs

• A Digital Elevation Model, or DEM, is a dataset
  consisting of XYZ points representing the surface
  heights on a regularly spaced grid.
• A DEM therefore can be (but isnt always) stored
  as a raster
• DEMs are INTERPOLATED, not measured.
• There are MANY ways to interpolate a grid
• Inverse Distance Weighted (IDW)
• Spline
• Trend
• ANUDEM (TOPOGRID)
• Each has its own set of parameters, and
• THEY WILL ALL PRODUCE DIFFERENT RESULTS FROM
  THE SAME INPUT DATA!

120
A contour by any other name
121
A contour by any other name
122
A contour by any other name
123
A contour by any other name
124
Lets Discuss DEMssome more!

• Different software products use different
  algorithms to generate DEMs from points data
• In many cases, the user cannot change the
  parameters for DEM generation and has to take
  whatever comes out as gospel (but which gospel
  would that be?)
• There are also different algorithms for
  generating contours, so different software can
  produce different contours even if they are input
  the identical DEM!

125
So which one should I use?

• Within certain reasonable limits, it doesnt
  matter.
• The interpolated DEM is the means to the end of
  generating contours, and contours are only an
  approximate visualization tool to being with.
• The LIDAR TIN provides good precision and detail,
  for reasonable costs, and with excellent
  production time.
• LIDAR will give you neither the aesthetic look of
  pulled contours, nor the precision of 50-scale
  mapping

126
Just another tool in the box

• LIDAR is a nothing more than a different tool
  with which we can measure the Earth.
• It has strengths and weaknesses, just like all
  other tools.
• It is perfectly suited for some applications.
• It is completely inappropriate for others.
• It is NOT the be-all and end-all of everything
  topographic!

127
Two for the price of one!
128
Lidar Bare Earth Filtering

• Filtering of the lidar bare earth points results
  in a greately reduced file size
• Utilizes slope and distance algorithms to remove
  unnecessary points
• The remaining points are actual bare earth
  XYZs
• Unlike TINs, ther is NO interpolation of the
  ground surface
• Provides useful file size for GIS platforms

129
Original Bare Earth Points 2,557,705 2' Contours
130
Filtered Bare Earth Points 301,736 2' Contours
131
Original Bare Earth Points 2,557,705 5' Contours
132
Filtered Bare Earth Points 1040,352 5' Contours
133
Ortho DEM Generation

• Ortho DEMs can be created at a much wider post
  spacing than the contour DEM
• Orthos down to 100-scale can be made without
  breaklines
• Ortho DEMs do not require 100 bare-earth
  inclusion of bridges and some vegetation actually
  improves quality and reduces PhotoShop editing!
• Ortho DEMs can be produced before Manual LIDAR
  Editing this significantly shortens the
  delivery schedule for final orthophoto products

134
Ortho created from LIDAR Points ONLY
135
Breakline Synthesis
136
Breakline Synthesis

• Lack of breaklines regarded as a major drawback
  to using LIDAR in some applications
• Breaklines can be merged from traditional
  photogrammetry, but this defeats the cost/time
  benefit of using LIDAR
• Critical breaklines are being synthesized using a
  variety of techniques
• Processes will improve with continued research

137
Hydrologic Breakline Synthesis
138
Hydrologic Breakline Synthesis (cont)
139
Digitize 2D Linework
140
3D-er-ize Stream Centerline
141
3D-er-ize Stream Edges
142
TIN From Points Breaklines
143
Hydrologic Breakline Synthesis
144
Huh!
145
LIDARApplicationsOverview
146
LIDAR Markets

• DTM Generation
• Effective Method for Large-Area Projects
• Faster at Less Cost Often More Accurate
• Mining Assessments/Volume Calculations
• Forestry
• Derive the Canopy, the Canopy Structure, and a
  Bare Earth DTM under the Trees
• Watershed Analysis for Management Plans

147
LIDAR Markets (cont.)

• Coastal Engineering
• Beaches and Coastal Zones
• Monitoring Dunes and Dikes
• Corridor Mapping
• Map Utility Right-of-Ways Effectively
• Measure Catenary Curves and Sag of Utility Lines
• Evaluate Clearance and Map Tower Locations
• Transportation Applications
• Highway Analysis, Infra-structure Identification

148
LIDAR Markets (cont.)

• Flood Plain Mapping
• Topographic Data for Flood Risk Maps Modeling
• FEMA Performing Many Assessments in the US
• Urban Modeling
• First Surface Models for Telecommunications,
  Line-of-Sight Determination, Law Enforcement, and
  Disaster Planning
• Bare Earth Surface for Flood Modeling

149
LIDAR Markets (cont.)

• Disaster Response/Damage Assessment
• Post-fire Hurricane Assessments
• Wetlands Restricted Areas
• Environmentally Sensitive Areas
• Areas Difficult to Map
• Wildlife Habitat Monitoring
• Mapping the Live Crown in Forestland Supporting
  Endangered Species

150
New Applications

• Huge Amount of On-Going
• Research and Application Development
• Forestry Biomass Assessment
• Riparian Vegetation Mapping
• Wildland Fuels
• Advanced Forestry Applications
• Arboreal Habitat Models
• Landslide Hazard Mapping
• Land Use/Land Cover Classification
• Hydrographic Mapping
• Hydrologic Modeling
• Breakline Generation
• Change Detection and Feature Extraction
• Urban Models for Buildings and Vegetation
• Obstruction Detection
• Feature Extraction
• Viewshed Analysis
• Advanced Surface Generation Rendering

151
Biomass Volume Analysis
152
Timber Volume Analysis
153
Classification by LIDAR Return

• Discrete Multiple-Return LIDAR data contains
  information beyond simple X-Y-Z values
• Number of returns from a pulse
• Vertical distribution of elevations
• Type of reflection First Only, Last of
  Many, etc.
• These can be used to delineate different types of
  land cover
• Last of Many points are only found in vegetated
  areas
• Difference between highest/lowest points in an
  small area can differentiate woods from shrubs
• Vertical distribution can indicate the type or
  maturity of a forest area

154
(No Transcript)
155
(No Transcript)
156
(No Transcript)
157
(No Transcript)
158
Viewshed Analysis
159
Glideslope Penetration Analysis
160
3D Data Visualization
161
3D Urban Visualizations
162
3D Urban Modeling with LIDAR
163
3D Urban Modeling with LIDAR
164
3D Urban Modeling with LIDAR
165
Disaster Response

• September 11, World Trade Center
• LIDAR elevation data collected every other day
  beginning September 15
• New TINs and DEMs were produced with daily
  turn-around
• Volumetric analyses of the site were critical
  for response
• Digital photography and Thermal imagery also
  collected

166
Disaster Response
BW Digital and Thermal Imagery over LIDAR TIN
167
Disaster Response
BW Digital and Thermal Imagery over LIDAR TIN
168
Highway Corridor Mapping
169
Pennsylvania Turnpike Corridor

• LIDAR Acquired in 9 Flight Hours
• Complete Project of 458 Miles for LIDAR and
  Digital Orthophotos
• Processed and Delivered Quickly
• Existing Mapping Available for Quality Assessment
• Typical of Freeway with Steep Side Slopes

170
Pennsylvania Turnpike Corridor
Raw LIDAR Surface
Bare-Earth LIDAR Surface
171
Cross Sections
172
Telecommunications Design

• Acquire and Process LIDAR to Derive a First
  Surface Terrain Model
• Telecomm Firms Insert Accurate Surface Model into
  Custom Software
• Customer Terminals and Hubs are Designed for
  Urban Telecom Applications
• Shadows of No-Service are Identified

173
30-Meter DEM of Sample Site
174
LIDAR Surface Model of Site
175
CTs Hubs Identified
176
Analysis of the Links
177
Identification of Shadows
178
Power Line Mapping/Inspection
Imagery Courtesy of Optech Incorporated Toronto,
Canada
179
Power Line Mapping/Inspection
Imagery Courtesy of Optech Incorporated Toronto,
Canada
180
Bathymetric Applications
Imagery Courtesy of JALBTCX, USACE Mobile,
Alabama
181
Bathymetric Applications
Imagery Courtesy of JALBTCX, USACE Mobile,
Alabama
182
Bathymetric Applications
Imagery Courtesy of JALBTCX, USACE Mobile,
Alabama
183
Bathymetric Applications
Imagery Courtesy of JALBTCX, USACE Mobile,
Alabama
184
Remote or Difficult Areas

• Remote areas which are difficult to establish
  traditional survey control
• Steep surfaces and canyons, mudflats
• Hazardous areas
• Firing or bombing ranges, waste sites

185
Orchard Training AreaIdaho National Guard
Boise
OTA

• South of Boise Idaho
• Approximately 760 square km
• Nominal 5m post spacing
• 37 million points in final data set
• Flown at 3048 meters AMT
• Swath width of 3520 meters

186
OTA LIDAR Terrain Surface

• Use of Data
• Combat Simulation
• Line-of-sight analysis
• Terrain analysis for endangered species
  habitat

187
BW Ortho of Project Area
188
Area With 1-Foot Contours
189
LIDAR Data in Hillshade Display
190
Cant we all just get along
191
North Carolina Floodplain Mapping Program

• Thanks to
• David F. Maune, Ph.D., C.P. - Dewberry Davis
  LLC
• and
• Gary Thompson - North Carolina Geodetic Survey
• for much of the following material

192
September, 1999Hurricane Floyd
193
Hurricane Floyd

• 10,000 square miles inundated
• 67,000 homes flooded
• 2.8 million livestock animals drowned
• 4117 uninsured/underinsured homes destroyed
• Revealed severe limitations in NCs FIRMs
• Majority were compiled in 1970s
• Many used approximate HH modeling techniques
• Many were based on USGS 30-meter DEMs
• Over 50 counties needed to be remapped digitally
  ASAP

194
Flood Insurance Rate Maps (FIRMs)

• FEMAs FIRMs used for floodplain management,
  regulation of new construction, determination of
  flood insurance requirements.
• FEMAs funds average one updated countywide flood
  insurance study per state per year
• North Carolina became FEMAs first Cooperating
  Technical State (CTS), whereby the state took
  primary ownership of, and responsibility for, the
  NFIP maps for all NC communities
• This includes resurveying the state, conducting
  flood hazard analyses, and producing updated
  DFIRMs.

195
LIDAR Deliverable Product Overview

• Raw Discrete Multiple Return LIDAR
• Longer-term information mining
• Bare-Earth LIDAR DSM
• Higher precision terrain applications
• Interpolated Bare-Earth LIDAR DEMs
• Large-area modeling and orthophoto production

196
Why High-Resolution LIDAR?
197
Why High-Resolution LIDAR?
198
NCFPM Phases
Phase ICoastalbluePhase IIPiedmontgoldPhase
IIIMountaingreen
199
Phase 1 Team Assignments
200
Starting Up

• Flight Planning and Ground Control Surveying
  began in late 2000.
• 493 flightlines
• 100s of Ground control points set/recovered
• 26 GPS Base Stations established
• ArcView project site established to monitor
  progress and status
• Base Station operation
• Priority collection areas
• LIDAR processing
• Collection progress
• Flight operations began on January 2, 2001 and
  were completed on April 8.

201
Flight Planning Ground Control
The primary inland areas were flown at 12,000
AMT for a 25 cm accuracy. Because of unique
coastal modeling requirements, the Outer Banks
were flown at 8,000 AMT for a 20 cm accuracy.
Special flights were also made to collect
crossties, shorelines, and the airport
calibration sites.
Three types of GPS base stations were used
throughout the project CORS stations maintained
by the NCGS, base stations surveyed, set, and run
for the duration of the project, and pre-surveyed
base stations set-up at each airport during
flight operations.
202
Processing

• Data processing began within days of the first
  flight, and were completed on December 21.
• gt 4.5 Trillion Points are processed into 1227
  bare-earth tiles (20,000 x 20,000)
• Custom software written to facilitate processing
  of the large (unthinkable!) quantity of data.
• EarthData team develops specialized new processes
  and products
• Synthetic hydrologic breaklines NO STEREO
  COMPILATION
• LIDAR-derived Land Cover classification

203
P R O D U C T I O N F L O W
Peer
Corrections
Review
204
Phase1 Tiling Scheme
205
Automated vs. Manual Editing
206
Automated vs. Manual Editing
207
Approximate Extent of Breaklines
208
LIDAR-Derived Land Cover
Neuse River Basin
209
Ingredients of a DFIRM

210
Representative Cross Sections

• Cross sections for hydraulic models are selected
  to represent reaches that are as long as possible
  while still having the same basic geometry and
  characteristics.
• Either photogrammetry or high-density LIDAR data
  may be suitable for cutting cross sections in the
  overbank areas, but the banks and channels
  typically must be surveyed.

211
LIDAR vs. Field Cross-Sections
LIDAR Elev.s from TIN vs. Field Survey Elev.s
Section Upstream of Northwest Bridge Rd in
Onslow County
212
Profile Data Fusion

• Surveyed data of the the inner banks and channel
  are merged with the LIDAR over bank profile

213
Checkpoint Land Cover Classes

• FEMA requires TINs to be tested separately for
  major land cover classes that predominate within
  the floodplain being studied, with 20 or more
  checkpoints per class. The NCFPM Project used
• Bare-Earth Low Grass
• Weeds/Crops
• Scrub/Shrub
• Wooded/Forest
• Urban/Built-up

214
(No Transcript)
215
Objective of NC QC Surveys

• Establish quality control (QC) checkpoints to
  evaluate the vertical accuracy of the LIDAR TIN
• 20-cm vertical RMSE in coastal counties
• 25-cm vertical RMSE in inland counties
• North Carolina established 20 checkpoints per
  county for four land cover classes, with 40 per
  county in forest class 120 total
• 5 of checkpoints allowed to be discarded to
  account for un-cleaned artifacts

216
Uncleaned Artifacts
Manual cleaning is regrettably very costly. How
clean is clean? is a very subjective
issue. Removing ALL artifacts is an elusive
goal. Research is being conducted to develop
techniques to objectively measure and
consistently quantify How clean is clean?
Artifacts such as these have minimal impact on
hydrologic modeling, but could impact other DEM
uses.
217
County QC Spreadsheet
Franklin County QC Points
218
QC Assessment Report

• RMSE summary using
• 100 of the checkpoints
• 95 of the checkpoints
• Checkpoints categorized by land cover type

The LIDAR data for Halifax County met the
specification as per the RMSE criteria of 25 cm.
All figures represented the data with the 95
data set. The data was of good quality with the
smaller data limited to the Tar-Pamlico basin
only. The land class type of built-up was
slightly high, but was well within specifications.
Halifax County, Tar-Pamlico Portion
219
QC Assessment Report
RMSE by Land Class
Grass
Weeds Crops
Scrub
Forest
Built-Up
Halifax County, Tar-Pamlico Portion
220
QC Assessment Report
Overall Statistics for the Checkpoint Data
Halifax County, Tar-Pamlico Portion
221
Quality Control Results
Specifications 25 cm inland, 20 cm on Outer Banks
222
Quality Control Results
223
LIDAR Helped To Produce
224
Progress Update
North Carolina FloodPlain Mapping Phase 2 2003
LiDAR Mapping Area
16,883 Square Miles 40 Counties 1331 Tiles _at_
20,000 x 20,000 6.5 Billion Points Completed
December 2003
225
Phase IIIperhaps in 2005
226
(No Transcript)
227
LIDAR TechnologyforForestry Applications
228
Forest Measurement Applications

• Analysis of Multiple Return LIDAR for Forest
  Characteristics
• Canopy Height
• Canopy Closure
• Dominant Stem Volume
• Basal Area
• Other Forest-related Data
• Wildlife Habitat
• Fuels
• Watershed Characteristics

229
Maximum Vegetation Height
Difference Between Highest and Lowest Returns
230
Percentage of Canopy Closure
231
Vegetation Height
232
Cover Percent _at_15-Meters
233
Lidar Potential

• Allows stand characteristics to be calculated in
  a few weeks for a 100,000 ha forest.
• Wall to wall coverage identifies problem areas
  low stocking, poor growth
• Additionally provides DTM for road layout,
  logging systems design, hydrographic data,
  wildlife habitat and riparian management

234
Ozone Research Using LIDAR

• differential absorption lidar (DIAL)
• ETL operates two ozone DIAL systems, a
  ground-based ozone DIAL and an airborne ozone
  DIAL, which have been deployed to many field
  projects in the United States to study ozone
  pollution.

235
LIDAR Used To Measure Wind

• Giant Aperture Lidar Experiment at Starfire
  Optical Range, NM
• The experiment was designed to measure high
  resolution wind and temperature

This time lapse photograph shows the 5 Na
Wind/Temperature lidar beams used during GALE
conducted by the University of Illinois,
Lockheed, Aerospace, and Phillips Laboratory at
Starfire Optical Range
236
LIDAR Used To Study Aerosols

• LITE is a three-wavelength backscatter lidar
  developed by NASA Langley Research Center to fly
  on the Space Shuttle
• LITE flew on Discovery in September 1994 as part
  of the STS-64 mission.
• the goals of the mission
• validate key lidar technologies for spaceborne
  applications
• explore the applications of space lidar
• gain operational experience to benefit the
  development of future systems on free-flying
  satellite platforms

237
LITE provided sensitive observations of the
distribution of desert dust, smoke, and other
aerosols
LITE provided views of multilayer cloud
structures.
LITE provided the first global observations of
planetary boundary layer height
238
LIDAR Used For Weaponary
YAL-1A ATTACK LASER
Boeing 747-400F

• The Airborne Laser (ABL), a bulbous-nosed
  aircraft unlike anything else in any military air
  fleet in the world
• A specialty project from its beginning in 1992,
  ABL has one primary mission destroy enemy
  ballistic missiles during their vulnerable boost
  phase.

239
The End
240
Special Thanks To

• Spencer B. Gross
• www.sbgmaps.com
• Optech Incorporated
• www.optech.com

• Geoserv
• www.geoserv.net
• JALBTCX (SHOALS)
• www.shoals.sam.usace.army.mil

241
Thank you for your patience
Contact Information Bob Ryan EarthData
International, LLC bryan_at_earthdata.com