OPTIMAL AND SUBOPTIMAL DETECTION OF A TARGET WITH RANDOM GREY LEVELS IMBEDDED IN NONOVERLAPPING NOISE

The problem of the detection of a target with random grey levels appea ring on a random background is addressed. The optimal processor for an y white statistics of the target's and the background's grey levels is determined. If the statistics of the grey levels are unknown, it is s hown that one can approach the performance of the optimal processor by correlating the silouhette of the target with preprocessed versions o f the scene and by performing the fusion of the obtained correlation p lanes. This result opens new perspectives for optical correlation.