Human and algorithmic imaging performance in the presence of speckle

In support of target recognition and tracking, we seek to compare the effects of speckle on human observation of moving objects with those of fixed-pattern spatial noise and Gaussian temporal noise through a dynamic minimum resolvable contrast (DMRC) test. In addition, a simple motion tracker algorithm built with a Kalman filter allows for a point of comparison between human and machine perception of speckled imagery. The results of the DRMC test show that speckle causes a greater degradation in observer performance than either temporal or fixed-pattern spatial noise, strongly correlated with target spatial frequency. The results from the simple motion tracker also show that speckle acts as neither temporal nor fixed-pattern spatial noise, and the ability to clearly track a speckled target is dependent upon the relative sizes of target and speckle diameter, which furthers the argument that speckle cannot be simply treated as conventional noise in an incoherent imaging system. Such outcomes will help inform the design and implementation of future active electro-optical systems that perform target recognition and tracking with coherent illumination.