Channel/Scene Integration:
Qualitative method of assessing regional change; ability to observe multiple dates in a single image; does not provide quantitative change-area information.

Multidate Composite:
Creation of change and no-change spectral clusters; requires labeling of the derived clusters from the unsupervised classification; PCA yields uncorrelated vectors; data compression technique; a single classification is required; difficult to label change classes; and no from-to class information.

Image Algebra:
Positive and negative spectral values indicating direction of change; yields a distribution approximately Gaussian, change pixels distributed at the tails; subjective placement of change thresholds standard deviation and empirical testing; provides no from-to change information.

Binary Mask:
Threshold is selected to identify change and no-change values; detailed change from-to information; reduces omission and commission errors; requires a number of processing steps.

Post Classification:
Most commonly applied approach; requires image preprocessing and classification; detailed change from-to information; classification errors affect change image results.