In remote sensing and digital image classification, the ParallelPipe or Parallelepiped classifier is a simple decision rule used to categorize pixels into classes.
This method defines decision boundaries in the form of multi-dimensional rectangles (parallelepipeds) in feature space.
Each class is represented by a range of values for each spectral band — the classifier checks whether a pixel’s spectral values fall within these ranges.
If they do, the pixel is assigned to that class.
Though easy to implement, the method can lead to unclassified or ambiguously classified pixels if spectral ranges overlap.
Thus, ParallelPipe is a fundamental but basic image classification approach.