Spectral resampling based on user-defined inter-band correlation filter: C3 and C4 grass species classification

13 In this paper, a user-defined inter-band correlation filter function was used to resample 14 hyperspectral data and thereby mitigate the problem of multicollinearity in classification 15 analysis. The proposed resampling technique convolves the spectral dependence information 16 between a chosen band-centre and its shorter and longer wavelength neighbours. Weighting 17 threshold of inter-band correlation (WTC, Pearson’s r) was calculated, whereby r = 1 at the 18 band-centre. Various WTC (r = 0.99, r = 0.95 and r = 0.90) were assessed, and bands with 19 coefficients beyond a chosen threshold were assigned r = 0. The resultant data were used in 20 the random forest analysis to classify C3 and C4 grass species. The respective WTC datasets 21 yielded improved classification accuracies (kappa = 0.82, 0.79 and 0.76) with less correlated 22 wavebands when compared to resampled Hyperion bands (kappa = 0.76). Overall, the results 23 obtained from this study suggested that resampling of hyperspectral data should account for 24 the spectral dependence information to improve overall classification accuracy as well as 25 reducing the problem of multicollinearity. 26 27