The radioactive decay is a random process, and the measurement precision is ruled by statistical laws. The counting ratios of the profiles are alwaysnoisy when analyzed for short periods, such as one second per measurement. Corrections made at the end of conventional processing in the airborne gamma-ray spectrometric method data are not enough to remove and minimize, or even reduce considerably, the spectrum's originated noise. Two statistic methods that actlocally in collected data, in the spectrum domain, have been suggested by literature to remove such remaining noises, the Noise-Adjusted Singular Value Decomposition - NASVD and Maximum Noise Fraction - MNF. These methods produce a significantly noise reduction. In this work both methods were applied in an area comprehended by two blocks, I and II, of the airborne survey that covers the west area of Mineral Province of Tapajós between Pará and Amazon states. The filtered and non-filtered data with the NASVD and MNF techniques were processed with the Lasa's parameters. The comparison of results between maps and profiles shows that both methods are valuable, since there was resolution gain in these products.
gamma spectrometry; NASVD; MNF; noise reduction
