Detection of local anomalies in high resolution hyperspectral imagery using geostatistical filtering and local spatial statistics

This paper describes a methodology to detect local anomalies in high resolution hyperspectral imagery, which involves successively a multivariate statistical analysis (PCA) of all spectral bands, a geostatistical filtering of noise and regional background in the first principal components using factorial kriging, and finally the computation of a local indicator of spatial autocorrelation to detect local clusters of high or low reflectance values as well as anomalies. A case study illustrates the ability of the filtering procedure to reduce the proportion of false alarms, and its robustness under low signal to noise ratios. By leveraging both spectral and spatial information, the technique requires little or no input from the user, and hence can be readily automated.