Reducing the effects of vegetation cover on airborne radiometric data using Landsat TM data

The gamma-ray spectrometry responses from bedrock in Canadian Shield areas are substantially masked by overburden and vegetation. Proper interpretation of airborne gamma-ray spectrometry data is dependent on accounting for the interference provided by surface cover. In this paper, a method is tested to correct airborne gamma-ray spectrometry, acquired over the Canadian Shield of northeastern Alberta, for vegetation screening by estimating the proportions of bedrock and vegetation cover from Landsat TM data. TM pixel values, due to the patchy network of bedrock and vegetation, result from a spectral mixture of these ground cover classes. Linear unmixing was implemented to deconvolve TM bands in abundance images to estimate proportions of bedrock and vegetation for each pixel. The outcrop abundance image, representing spatial variation in area percentage of bedrock, is used in linear regression analysis to calibrate co-registered K, eTh, and eU gamma-ray spectrometry channels to 40 per cent bedrock endmember images.