Detecting dim point target in image data using adaptive prediction filter

This paper studies the performance of least mean square (LMS) adaptive filters as prewhitening filters for the detection of point target in image data. The object of interest is assumed to be pixel-size and is obscured by correlated noise of much larger spatial extent. The correlation noise is predicted and subtracted from input signal, leaving components of the point target in the residual output. The noise is suppressed and the target is enhanced relatively. The prewhitened image is then processed by a proper threshold to pick out the candidate target. Experimental results show that such a detector has better operating characteristics than a conventional matched filter in the presence of correlated clutter and noise. For very low SNP, LMS-based detection systems show a considerable reduction in the number of false alarm. Simulation results have been provided at the end of the paper.