Abstract: A gas turbine pyrometer filtering method and system uses a pyrometer to measure the high pressure turbine blade temperatures using infrared optical detection techniques. The pyrometer signal can be cluttered by positively biased noise. The turbine speed and the pyrometer electrical signal are alternately sampled by an analog to digital converter, and several revolutions of pyrometer data are stored in the local RAM. A data compression algorithm then selects single samples, at evenly spaced intervals, which results in a given number of revolutions, containing an exact number of data points of aligned data in RAM. This data is passed to a clutter rejection filter which then passes only the lowest value for each of the points in all revolutions of data. Thus, the revolutions of acquired data are reduced into a single revolution of filtered data for the current sampling. This single revolution of filtered data in then enter into a circular queue and the last element is discarded.