MULTI

1
Fusing Absolute and Relative Methods for
Enhanced Geolocation Accuracy Verification Chuck
OHara, Spatial Information Solutions,
Starkville, MS / Mississippi State University,
Geosystems Research Institute Greg Stensaas, U.S.
Geological Survey, Sioux Falls, SD
Image-to-Image (I2) Tools accuracy analyst TM
ASPRS November 14 19, 2010 Orlando, Florida
ABSTRACT Rapidly evolving mapping technologies
and information necessitate the practical
application and standardization of new core
technologies for the verification of map content
for updates and maintenance. Automated methods
that speed and inform the verification process
are emerging that offer the ability to more
effectively conduct absolute as well as relative
geolocation accuracy verification. These methods
combine the use of checkpoint data,
point-to-image analysis, and image-to-image
analysis for absolute and relative accuracy
verification of orthophotos. Checkpoint data
collection and offset analysis, once a tedious
and laborious process, may now be completed via
mostly automated processing. The absolute
accuracy of high resolution orthophotos may be
verified in an automated processing environment
using point-to-image (P2I) methods. As part of
the automated processing, georeferenced image
chips (iChips) may be extracted for desired areas
around each checkpoint location. The iChips would
include the full resolution and contents of the
original image and would be embedded with
marker symbols at photo identified features and
their surveyed locations. These iChips may be
used for enhanced image-to-image (I2I) relative
accuracy verification of new image
collections. Novel methods of relative accuracy
verification are presented using reference data
embedded with absolute accuracy information to
provide results closely approximating those
achieved through absolute accuracy checkpoint
analysis methods. Utilizing iChips embedded with
marker symbols at surveyed checkpoints enables
the selection of the survey marker symbols in the
iChip as the reference location and the
corresponding feature in the new imagery as the
test location for offset analysis. P2I absolute
accuracy verification are presented and compared
with iChip enabled I2I methods demonstrating the
benefits of this fused methodology for
geolocation accuracy verification.
INTRODUCTION The quantification of geolocation
accuracy and uncertainty is most frequently
conducted based on absolute or relate
measurements of offset. For absolute
quantification, survey-grade locations of ground
features are acquired and employed as checkpoints
against which image derived locations for the
feature are compared to determine offset in X and
Y. In relative assessments, similar features are
identified in a reference and a test image set
and offset is compared between the two data sets
and used to compute accuracy statistics. Both
absolute and relative accuracy and agreement are
important to determine the usefulness and quality
of data for intended purposes. The determination
of relative accuracy may be of specific interest
in cases such as the following Multi Temporal
Images Collections of Images Taken Over
Time. Multi Resolution Images Collections of
Images Collected at Different Resolutions. Multi
Source Images Images Acquired by Different
Sensor Platforms for a Common Area of Interest
(AOI). Summary of Process Workflow 1) Conduct
Absolute Accuracy Analysis Using Surveyed
Checkpoints, Offset Analysis, and Statistical
Calculation. 2) Generate Georeferenced Image
Chips (iChips) for Checkpoint Locations with
Symbol Markers Showing Survey Locations. 3)
Conduct Relative Accuracy Assessment Using
iChips as Reference Data and Identifying Photo
Locations in Test Image Data. Acknowledgements
USGS EDC Data Verification Team Tuck
Mapping IADIWG
PROCEDURE FOR USING I2I TOOLS STEP 1 Load Test
Image Data Set and iChips as Reference Select the
First iChip and Zoom for Desired Level of
Detail Note that the Data Table is Absent of
Location Data STEP 2 Zoom and
Pan on iChip as DesiredSelect Survey Checkpoint
Location on iChips Data Table X Y for Survey
Locations are Populated STEP 3
Zoom on Test Data Location as NeededSelect
Corresponding Photo Location on Ortho Test
Data Data Table X Y for Image Derived Locations
are Populated
SAMPLE TEST CASES Absolute Accuracy
Verification 0.25 Foot (GSD) RGB Image
Data Sioux Falls, South Dakota, USGS Eros Data
Center Test Range RMSE X 0.28 RMSE Y 0.17
CE90 0.48 CE95 0.55 (Feet)
Absolute Accuracy Verification 0.50 Foot (GSD)
CIR Image Data Sioux Falls, South Dakota, USGS
Eros Data Center Test Range RMSE X 0.40 RMSE
Y 0.32 CE90 0.78 CE95 0.89
(Feet) Relative Accuracy
Verification (I2I) 0.50 Foot (GSD) CIR Image
Data Sioux Falls, South Dakota, USGS Eros Data
Center Test Range RMSE X 0.47 RMSE Y 0.31
CE90 0.84 CE95 0.96 (Feet)

MULTI SOURCE I2I VERIFICATION (ORTHO LIDAR
INTENSITY) Absolute Accuracy Verification .125
Foot (GSD) RGB Image Data Tuck Mapping,
Co-Acquired Ortho and LiDAR RMSE X 0.11 RMSE
Y 0.06 CE90 0.18 CE95 0.21
(Feet) Relative Accuracy
Verification .25 Foot (GSD) Pan Intensity
Image Tuck Mapping, Co-Acquired Ortho and
LiDAR RMSE X 0.14 RMSE Y 0.08 CE90 0.23
CE95 0.27 (Feet)
SUMMARY AND CONCLUSIONS Current trends in
acquisition and use of high-resolution orthophoto
and LiDAR data present challenges in areas of
automated, efficient, consistent, and scalable
methods of data accuracy verification. Frequency
of updates as well as multi-resolution and
multi-source data means that new methods must be
validated not only for absolute verification of
data against surveyed checkpoint, but also for
relative accuracy between data sets to assure
co-alignment and agreement of data. New methods
presented for extracting iChips from
high-accuracy reference data is shown to provide
a systematic and practical basis for relative
accuracy verification. Embedding iChips with
symbols at survey checkpoints and the photo
identified locations makes use of iChip data and
extraction of necessary information easy and
effective. The methods presented provide a sound
basis for standardizing collections of iChips for
accuracy verification programs.

Collection of Image Chips (iChips) from
Orthophoto and LiDAR Intensity Data Tuck Mapping
Data Sets Co-Collected During Helicopter Data
Acquisition Mission