Computer-aided Classification for a Database of Images of Minelike Objects
Abstract
In this report we describe some computer classification experiments with a database of sidescan sonar images. This database consists of 383 swaths of sidescan sonar data extracted by the authors from sea trial data collected over the last few years by DRDC Atlantic. The effects of the filtering and image segmentation processes on the resultant classification rates are considered. A number of kernel-based and nearest-neighbor classification schemes are examined. It is found that despite the complexities of the database considered in this report that high classification/low false alarm rates can be achieved.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 2005
- Accession Number
- ADA437000
Entities
People
- John A. Fawcett
- Vincent L. Myers
Organizations
- Defence Research and Development Canada