The detection of unresolved targets using the Hough transform
Abstract
Clutter resulting from drift and vibration in a sensor line-of-sight and time-varying components in the background can seriously hinder the detection of faint, unresolved (i.e., point) targets moving against a background. These targets, however, produce well-defined patterns of motion referred to as tracks, whereas the effects of noise and clutter are more random. Frame-to-frame differencing can be used to reject clutter and enhance the detection of moving targets. These difference frames can then be combined recursively to form target tracks. The tracks of nonmaneuvering (i.e., constant velocity) targets appear as line segments that can be detected using the Hough transform. In addition, the current position and heading of a target can be determined using a technique based on matched filtering. An example demonstrating the performance of this technique is also presented.
- Publication:
-
Computer Vision Graphics and Image Processing
- Pub Date:
- February 1983
- Bibcode:
- 1983CVGIP..21..222C
- Keywords:
-
- Clutter;
- Fir Filters;
- Image Processing;
- Pattern Recognition;
- Target Acquisition;
- Frequency Response;
- Linear Filters;
- Communications and Radar