Digital Orthorectification of Soil Surveys for GIS Applications

1
Digital Orthorectification of Soil Surveys for
GIS Applications

• Jay Bell
• Ty Wilson
• Eero Kilkson
• Matt Krusemark
• Dept. Soil, Water Climate

Mohamed Rouchdi Institut Agronomique Et
Veterinaire Hassan II Rabat, Morocco
2
Overview

• National Cooperative Soil Survey
• Current Approaches to Recompilation
• The Nature of Map Base Errors
• Photogrammetric Solutions
• Orthorectification Procedure
• Results
• Conclusions

3
Current Method Manual Recompilation

• Manual Line Transfer from map sheet (NAPP Photo)
  to hard copy orthophoto
• Time Consuming
• Expensive
• Opportunity to introduce errors
• Bottleneck for digitizing

4
Soil Survey Map Sheet
Distorted Map Base for Most Minnesota Counties Pr
e 1980 -FSA or local contractors 1980-87
-USGS NHAP (180,000) 1987-Present -USGS NAPP
(140,000)
5
(No Transcript)
6
Why are we doing this?
River appears to run upslope
7
Terrain Relief Displacement in Aerial Photographs
8
Scale Datum

• The horizontal plane at the elevation that the
  scale of the photograph is calculated
• Points with elevations above and below the scale
  will be displaced
• Set at the mean elevation of the landscape

9
Example
10
Coordinate System Transformations
11
Standard Approaches Two-Dimensional
Transformations
Common Algorithms Affine (linear) Rubber
Sheeting (non-linear)
Digital Orthophoto
12
Problems Does not directly consider
terrain relief displacement errors Inaccurate
transformations in areas of significant topographi
c relief
Digital Orthophoto
13
River appears to run up-slope
14
Digital Orthophotography
Common Algorithms Space Resection Statistical
Approximation of Orthorectification
15
Space Resection

• Geometric Reconstruction
• Analytical Solution
• Requires Fiducial Points
• Requires Camera Calibration File
• Control Points

Photo
DEM
16
Data inputs for the SYRUP orthorectification
DATA Needed
Inputs include File X File Y Ground
X Ground Y Ground Z (elevation)
17
Statistical Approximationof Orthorectification

• Fiducial points and camera calibration file not
  required
• Requires control points (use DOQ for soil survey)
• Requires DEM
• Statistically approximates solution

Digital Elevation Model
18
Orthorectification ProcedureSimple Yet Robust
Universal Parametric orthorectification(SYRUP
MODEL)
19
Scan Soil Survey Map Sheet

• 300 dpi scanner
• Soil survey mapsheet

20
Locate Control Points on Mapsheet and Orthophoto
Soil Survey Map Sheet
Digital Orthophotography
21
Orthorectification of Scanned Image
Distorted Map Sheet
Orthorectified Map Sheet
SYRUP MODEL
Digital Elevation Model
Control Points
Root Mean Squared Error
22
Extract Soil Boundaries
Orthorectified Map Sheet
Image Processing
Soil Boundaries over DOQ
23
Evaluate SYRUP Orthorectification

• RMSE
• National Standard for Spatial Data Accuracy
• Digital imaging techniques by GIS technicians
• Boundaries relative to landscape features

24
Results

• 6 Test Sites
• Goodhue, Minnesota
• Winona, Minnesota
• Washington, Minnesota
• Murray, Minnesota
• Sheboyan, Wisconsin
• Smokey Mts., North Carolina
• RMSE below Nat. Map Accuracy Standards
• RMSE below Nat. Standard for Spatial Data
  Accuracy
• Visual examination by soil scientist indicates
  successful Orthorectification

25
Whats the difference?
26
Conclusions

• Statistical approximation is a valid approach for
  the orthorectification of soil surveys on
  distorted map bases
• Considers the terrain relief displacement errors
  (base photo errors)
• Cost effective compared to current methods of
  manual recompilation

27
Current Work

• Full Minnesota County completed
• Logistics of processing
• Data storage, CPU, platform, etc.
• Evaluation of costs
• per acre or per mapsheet?