STUDY TO IMPROVE ALGORITHMS FOR CONTROL POINT MATCHING

Authors
  1. Kuwahara, R.
  2. Skea, D.
  3. Barrodale, I.
Corporate Authors
Defence Research Establishment Pacific, Victoria BC (CAN);Barrodale Computing Services Ltd, Victoria BC (CAN)
Abstract
In practice, two problems with this procedure arise. First, the position and attitude data of the tow fish contain significant observational errors. This introduces errors in the geo-coding (the assignment of geographic positions to each pixel in the image) of both images and results in complex geometric distortion of one image relative to the other. Second, because side-scan sonar is an active sensor imaged from two different positions (aspects) the task of change detection is very difficult. To use control point matching for this application the side-san sonar images are first idealized as point sets. When these point sets have been selected, common points, as identified by the control point matching algorithm, can be used as a basis of a rubber-sheet transformation which warps one image onto the other. The location of points in the new image which do not appear in the original may indicate an area of change.
Report Number
DREP-CR-92-01 — Contractors Report
Date of publication
15 Mar 1992
Number of Pages
157
DSTKIM No
92-03265
CANDIS No
125719
Format(s):
Hardcopy;Originator's fiche received by DSIS

Permanent link

Document 1 of 1

Date modified: