Image Indexing for Nearshore Restoration Morgan McKenzie and Dan Allen Geog 469

1
Image Indexing for Nearshore RestorationMorgan
McKenzie and Dan AllenGeog 469
2
Project Goals

• Create an Image Data Index Model
• For nearshore restoration ecologists
• Provide search tool for research time saving
• Images show important landscape attributes shore
  forms, watersheds, vegetation, structures, etc
• Index provides way to find scientific research
  question desired attributes

3
Background

• Puget Sound Nearshore Restoration Project,
  Research group
• Client Miles Logsdon, Restoration Ecologist
• Also Matt Parsons of UW Libraries and WAGDA
• Implement restoration projects and monitor their
  success/failures to improve
• Examples of restoration projects
• Remove bulkheads, plant overhanging vegetation

4
Index of Images for Restoration

• See potential areas of restoration
• Monitor areas with implemented restoration
  project
• Potential Scientific Research Questions include
• How much vegetation?
• Depositional or erosional shoreline?
• Shoreline before/after a structure removal?
• Shoreline before/after an event?

5
Data Process Diagram
Analyze Attributes (such as overhanging veg,
aquatic, veg, shorline length, etc)
Process images (manually or with image processing
software)
Download image from WAGDA
Data Entry
For PSNRP analyze geomorphic objects, woody
debris, of beach armored
6
(No Transcript)
7
(No Transcript)
8
(No Transcript)
9
(No Transcript)
10
Project Results

• Created a image data index model database
• Discovered which attributes were important and
  why
• Made database searchable by attributes

11
Image Database
IMAGE
Structures
Vegetation
ID Name Date Resolution Source
ID Bulkhead Over water structure Armored
ID Woody Debris of Clusters overhanging
Vegetation Aquatic vegetation
Image database
Bch Length
Geomorphic Objects
Location
ID Location Latitude Longitude Projection
ID Spit Embayment Erosion Deposits Watershed
ID Beach Length
12
(No Transcript)
13
Conclusions

• Indexing images for a specific use is difficult
  because of the large unique set of content
• Completing the data entry is a huge undertaking
• Once past the data entry phase, potential for
  long term time saving use
• Image data index model can be used over and over
  again for new data