A comparison of three methods of detecting point objects using a LLLTV system

The paper presents some results concerning the problem of detecting a point object of known shape in noise using a photoelectronic imaging (i.e., LLLTV) system. It is found that detection algorithms that use information about the signal and noise performance of photoelectronic imaging systems produce superior performance than less sophisticated algorithms. It is also shown that discrepancies between modeled and actual system parameters can significantly degrade the performance of all detectors.