Dim point target detection and tracking system in IR imagery

Temporal profiles of point-like dim targets and the extended cloud pixels provide useful information in detecting the targets. Among the recent methods that utilize temporal profile of pixel in detecting point target are Triple Temporal Filter (TTF) and Continuos Wavelet Transform (CWT). TTF uses two damped sinusoidal filters, an exponential averaging filter with six appropriate coefficients to deal with different aspects of clutter. TTF is recursive and efficient in detecting point targets without applying any threshold techniques. The performance of CWT is comparable to TTF but all the frames in a sequence need to be stored. Therefore it is computationally complex algorithm. This paper is also based on target and clutter profiles but proposes a multistage IIR Filter method. This method uses simple filtering and pulse analyzing scheme. Thus, it is computationally efficient compared to CWT. Simulation results on a real-world image sequence has shown that multistage IIR Filter has S/C ratio of 4.72 compared to CWT method which has SIC ratio of 5.18. However, few frames still need to be stored for processing. The multistage IIR Filter system consists of a series of band pass HR filter with different cutoff frequency arranged in parallel in order to discriminate between target and clutter regions. Filter output will contain potential target signal, which typically has a pulse shape. The pulse is extracted from its surrounding. The potential target pulse is further analyzed within a predefined width limits. In the following stage, output is rated based on the similarity to true target temporal profile. True point target temporal profile has a pulse profile different from its surrounding. Thus the shape similarity is rated as ratio of the pulse intensity to the intensity of its surrounding. In the post processing stage, a global threshold is applied to the image. The threshold level is set at three times the standard deviation from mean of spatio value. Further false alarm is suppressed with correlation method. A true point target often follows a trail with certain direction. Therefore it exist correlation between consecutive frame for target pixels. Whereas clutter and noise exist randomly and do not exhibit any trail over consecutive frames. By utilizing this property, correlation method is able to differentiate target and non-target pixels. The simulated results are presented.