If a threshold is passed to kNN_SAM_Classifier.label_unclassified_pixels(), only those pixels are labelled as unclassified where the first spectral neighbour exceeds the given threshold. Lets say, we have a pixel with 5 SAM neigbours... `[0.2, 1.2, 3.4, 5.3, 6.9] => [0.2, 1.2, 3.4, 5.3, 6.9] instead of [0.2, 1.2, 3.4, --, --]` ...then label_unclassified_pixels(threshold=5) does NOT flag the last two neighbors. Only pixels with a SAM exceeding the threshold at the first neighbor get correctly flagged: `[5.2, 6.4, 6.7, 8.4, 10.3] => [--, --, --, --, --]`