Fig. 6.

Relevance spectrum. Relevance of the individual spectral bands was derived from the structure of the Random Forest classifiers. More relevant wavelength features are used more often and hence contribute more to the final decision. The images of the two hyperspectral cameras have been processed independently and result in the blue and the red relevance profile, respectively. For each camera a number of highly relevant bands are found. Three local maxima in the relevance profiles are highlighted. Limiting classification to only the three highlighted relevant wavelengths yields mean accuracies of 0.98 (VNIR camera) and 0.99 (SWIR camera) for detached berries and in combination with textural features extracted from these image bands