Abstract
Sonar offers a convenient and economic means of obtaining range data. However, noise in such range data can seriously affect the detection of nearby boundaries, using the conventional circular back projection method. In this paper, we present a new formulation of the problem whereby the boundary to be detected is identified with the envelop curve of the ellipses defined by the range data. Hough transform is then applied to extract the local tangents to the boundary. The method was tested on both simulated and real data from polygonal boundaries with very promising results.
Original language | English |
---|---|
Pages (from-to) | 1975-1982 |
Number of pages | 8 |
Journal | Pattern Recognition |
Volume | 34 |
Issue number | 10 |
DOIs | |
Publication status | Published - Oct 2001 |
Scopus Subject Areas
- Software
- Signal Processing
- Computer Vision and Pattern Recognition
- Artificial Intelligence
User-Defined Keywords
- Boundary location
- Circular back projection
- Data fusion
- Hough transform
- Range data
- Sonar