Resolution and Image Metrics for Digital Orthophotos

1
Resolution and Image Metricsfor Digital
Orthophotos

• Craig Molander, SVP Business Development

2
Surdex Corporation

• Located in St. Louis, MO area
• 53rd year of operation
• Photogrammetric mapping and GIS across United
  States and Canada
• Film, digital (DMC), and LiDAR acquisition
• Fleet of 6 self-maintained aircraft
• Federal, state, local, defense/security mapping,
  engineering, utilities

3
Technology in the Last 3-5 Years

• Accelerating use of digital sensors
• Direct to digital sharper imagery
• Deeper bit depth (12bpp) more latitude in
  processing
• Multispectral bands wider range of products
  (BW, color, CIR/NIR)
• Higher awareness of image quality
• Increased demand/competition for higher quality
• Image quality metrics

4
What Does This Mean?

• Image data has more information content than
  film-based results of the same geometric
  resolution
• Ability to lower costs for large-area projects by
  utilizing lower-resolution digital orthophotos
• We as a profession have the challenge of
  educating our customers and each other on the
  very subjective aspect of image quality

5
Image Metrics

• USDA NAIP Best Practices guidance developed
  with ITT Space Systems
• Based on user studies/validation
• Guidance on processing flow/techniques
• Easy-to-compute histogram-based metrics
• Calibrated display monitors
• Specification of metrics such as contrast,
  clipping, dynamic range, color, etc.
• First attempt at objective assessment of image
  quality by a major program

6
Color Balance Variations(Courtesy ITT Space
Systems)
7
User Display Monitors(Courtesy ITT Space Systems)

• Actual pictures of variances of user displays
  (same image!)

8
Major Cost Factors

• As resolution improves (eg 1 foot ? 6), costs
  increase
• Acquisition 2X
• Image processing/storage 4X
• Aerotriangulation/ground survey 1.5X
• Elevation data generation/update 1.5X
• The elevation model becomes the key issue for
  resolutions of 1 and better

9
Resolution vs. Relative Costs
USGS NED (30m/10m)
Project-Specific Elevation Model
Relative Cost
0.5/6
1.0
2.0
3.0/1m
Resolution
10
Resolution

• Users should keep an open mind for image
  resolution requirements for large areas
• 2 versus 1m little cost increase, may interpret
  to a resolution of 1-1.5
• 2/18 for largely rural areas vice 1,
  interpreting to 1-1.5
• 2007 Missouri project
• 40,000 square miles (eastern 2/3rd)
• 400,000
• Originally planned 1 meter resolution were
  able to upgrade to 2/0.6m resolution digital
  acquisition/products

11
(No Transcript)