The influence of temporal and fixed pattern noise in staring and scanning IR systems on the perception of MRTD test patterns is investigated. The objective of the study was to validate the U.S. FLIR92 model. Digital simulation techniques were used to provide the imagery required for the perception trials. The experimental results are compared to FLIR92 model predictions. It was found that the perception of MRTD test patterns in fixed pattern noise is a random process, with the result depending on the location of the test pattern. Predictions can be made for the average (50% probability) perception capability only. With this restriction, and using a variable threshold signal-to-noise ratio, the FLIR92 predictions were validated for staring systems and fixed pattern noise. Large discrepancies were found for fixed pattern noise in scanning systems. The threshold signal-to-noise ratio function was extracted from measurements made with temporal noise.