GPSAided Inertial Technology

1
Testing Airborne GPS Digital Mapping Systems
Navigation-based Photogrammetry for Antarctic
Deployment
RICHARD D. SANCHEZ KENNETH W. HUDNUT LARRY D.
Hothem U.S. Geological Survey
U.S. Department of the Interior U.S. Geological
Survey
2
Objectives

• Investigate the positional accuracy of Applanix
  Digital Sensor System (DSS)
• Validate the instrumentation and procedure for
  applying direct geopositioning data in the
  navigational photogrammetric process
• Assess the potential of direct geopositioning
  for near real-time mapping

3
Background
4
Background Conventional vs. digital GPS/INS
survey mapping scenarios
Potential Benefits of Digital Airborne GPS/INS
Survey Mapping Cost Effective minimizes the
use of GCPs to QA/AC no aerotriangulation
performance of all measurements on PC Fast
emergency situations easy integration with
conventional software input/output in near
real-time
Illustration adapted from June 1997 GIS World
article on AIMS
Digital airborne integrated GPS/INS has the
potential of eliminating 4 processing steps
required in conventional surveying mapping,
saving from 40 - 50 of the cost of a large
scale aerial mapping project
5
Background Direct geopositioning as the
enabling technology

• Film Digital Cameras

• DG system records the GPS and IMU data and time
  of mid-exposure
• DG post-processing software computes the
  perspective center positions and camera
  orientations at the mid-exposure times
• Exterior orientation parameters are computed of
  each image in a format suitable for entry into a
  stereo-plotter

6
Background Airborne Cameras

• Film and Digital Single Shot Camera

• RC30 and Applanix DSS are examples of single shot
  BW and color cameras that can be processed
  within existing photogrammetric software

7
Resources
8
Background Digital Sensor System
DSS Camera Contax Lens Zeiss Focal
Length 55.0000 mm Frame/format Size 36.83 mm x
36.70 mm Solid State CCD, full frame 4,092 x
4,077 pixels Pixels per frame 16,683,084 Pixel
spacing 9 microns (center to center) Pixel size 9
x 9 microns GSD Optimum 10 cm (4 in)
POS-AV Platform
IMU
Courtesy of Applanix
9
Digital Sensor System aboard the Cessna 182
aircraft
10
Close-up view of the Digital Sensor System
11
(No Transcript)
12
DSS
13
(No Transcript)
14
BORESIGHT ALIGNMENT Applanix POS/DG test flight
trajectory and processed results
Aircraft trajectory
Data from POS/EO and AT solution are used to
calibrate the misalignment angles between the
IMU and camera
15
Methodology photogrammetric processing
Digital Aerial Photo Import
POS/DG COMPUTED DATA AT CAMERA PERSPECTIVE CENTER
Exterior Orientation (removal of elev.
distortions)
GPS/IMU derived EO (X, Y, Z, w, j, k) output
provided by Applanix
DIGITAL CAMERA REPT
Interior Orientation (removal of optical
distortions)
Camera Calibration Data (focal length,
lens distortions, fiducial coordinates-corner
pixels, etc.)
Model Setup
BY-PRODUCTS
Digital Elevation Models, Orthorectified Image
Mosaics, Surface Elevation Contours
Stereo Model Generation
Comparison of X,Y,Z, coordinates of known check
points with orthoimage values
Mapping Accuracy Assessment
Image processed with SOCET Set digital
photogrammetric software in UTM, State Plane, or
Geographic coordinates system, WGS84 Datum,
ellipsoid heights, and meter units
Documentation
16
Location
Satellite view of California Wildfires of 2003
Project Area
17
DSS Flight Corridor
Lone Pine
Devore
San Andreas fault
Cucamonga fault
Demen
Seven Oaks
San Bernardino
18
(No Transcript)
19
Courtesy of Applanix
DSS Flight Corridor
20
Courtesy of Applanix
21
Courtesy of Applanix
22
Courtesy of Applanix
23
Devore
Frames 1446 Format TIF Size 36.83 mm x
36.70 mm Solid State CCD, full frame 4,092 x
4,077 pixels Pixels per frame 16,683,084 Pixel
spacing 9 microns (center to center) Pixel size
9 x 9 microns GSD varies from 8- to 35-cm
DSS frame number and distribution
24
HVC 09
Validation of Panel Control Coordinates at Seven
Oaks (GSD 30 cm)
25
Aerial Panel HVC 06 in Devore (GSD 8 cm)
GCP 214
26
Ground level view of burnt hills and debris flow
area near panel HVC 06 in Devore
27
Debris flow damage
28
Findings

• Stereo overlap in the highest resolution dataset
  (8 cm GSD) was not attained due to strong cross
  winds excessive roll of the aircraft
• Horizontal positional accuracy was achieved in
  the 18 cm GSD dataset, with minor corrections the
  vertical positional accuracy can be improved to
  meet large scale mapping standards

• Factors that influence data quality and need
  further research

• Camera focal length, flight altitude, speed, and
  inclement weather such as strong cross winds
• Camera offset discrepancy in antenna relation
  to camera by drift compensation and gyro
  stabilizing camera mount
• Geodetic complexities - low accuracy gravity
  model
• Cycle slips loss of SV signal causing different
  ambiguity errors
• Poor stereo overlap due to excessive roll of
  aircraft
• Lack of precise GCP coordinates in post-processing

29
Preliminary Assessment of Ground Positioning
Accuracy of the DSS
DIFFERENCE (Meters)
Delta_vertical
Delta_northing
Delta_easting
1
5
6
7
2
3
4
8
REFERENCE POINT
Table 1. Statistical rundown of the difference
(Delta) between the San Bernardino County
ground-surveyed reference points and
corresponding aerial panels measured on the
digital photogrammetric workstation
30
Comparison of Ground Positioning Accuracy of San
Andreas data with other tests
Glen Canyon test using GPS integrated INS with
RC30 Camera
Glen Canyon test using GPS integrated INS with
Digital Sensor System
31
Final Assessment of Ground Positioning Accuracy
of the DSS
Table 2. The final positional accuracy offset
averages, in meters, of the ground-surveyed
referenced points
32
(No Transcript)
33
A DEM can be developed from a stereo model of
two overlapping frames, but 60 80 end lap and
30 side lap is required for mosaicking a DEM
project area
18 cm GSD
DEM mosaic coverage derived from DSS data flown
at an altitude of 1,067 m (3,500 ft)
34
GSD 8 cm
coverage gaps
DEM mosaic coverage derived from DSS data flown
at an altitude of 457 m (1,500 ft)
35
Conclusions

• The Digital Sensor System has the potential of
  providing near real-time mapping of the San
  Andreas and other fault systems

• More research is recommended to further improve
  vertical positional accuracy and refine the
  navigational photogrammetric process

• Where vertical positional accuracy is crucial for
  now a combination of airborne GPS/INS imaging and
  Lidar is recommended

36
Devore
Samples of DSS image derived products
37
ArcGIS 3-D Perspective
Devore
Note By default ArcGIS adjusts resolution to
reduce data volume and improve speed of display
38
(No Transcript)
39
(No Transcript)
40
(No Transcript)
41
(No Transcript)
42
(No Transcript)
43
(No Transcript)
44
(No Transcript)
45
(No Transcript)
46
(No Transcript)
47
(No Transcript)
48
(No Transcript)
49
Point of Contact

• For more information about this study contact

Kenneth W. Hudnut Geophysicist U.S. Geological
Survey 525 South Wilson Ave. Pasadena, CA
91106 hudnut_at_usgs.gov 626 583-7232
Richard D. Sanchez Physical Scientist U.S.
Geological Survey 521 National Center Reston, VA
22092 rsanchez_at_usgs.gov 703 648-5121
Larry D. Hothem Geodesist U.S. Geological Survey
521 National Center Reston, VA 22092 lhothem_at_usgs.
gov 703 648-4663
50
Thank you for your attention!