Search Radar Track-Before-Detect Using the Hough Transform

摘要

Detection of small targets in the presence of noise and sea clutter interferencepresents a formidable task in a radar system design. Conventional radar detection schemes, such as spectral discrimination and noncoherent integration have been employed with limited success. This thesis suggests an improved target detection scheme, applicable to search radars, using the Hough transform image processing technique. The system concept involves a track-before-detect processing method which allows previous data to help in target detection. The technique provides many advantages compared to traditional techniques. These include improved detection for moving targets, a solution to the range walk problem and automatic tracking without the need to revisit. The improved detectability results from better use of old energy and spatially separated energy which is equivalent to using a three-dimensional filter matched to the target trajectories in addition to the conventional target parameters. The questions answered by this thesis concern the effectiveness of the Hough transform in achieving improved radar target detection and system detection performance, (i.e., probability of detection and false alarm rate as a function of signal to noise ratio). System design concepts are considered and a full environment simulation including sea clutter and noise is implemented to determine the algorithm efficiency and performance in various scenarios.