During the last thirty years, wide and complex algorithm developments happened in the field
of remotely sensed data classification. Majority of these algorithms rely in carrying out the
classification on color characteristics reflected from objects found on earth surface, and
neglect other useful characteristics such as those related to location (space) and time. Such
an algorithm “deficit” still exists in most of the classification procedures, and negatively
affects the classification accuracy. This fact is principally true when the goal of classification
is to determine what is called “multi-level” subclasses.
The development of the proposed algorithm collection is supposed to comply with the
principles of comparative reasoning. Consequently it is expected that the algorithm collection
can be applied with an indicative results on different land cover classes, under different
geographical and/or climatologic circumstances. It is expected that the practical
implementation of the algorithm collection will be in the form of Program Package. So beside
knowledge in the fields of remote sensing and image processing, well founded mathematical,
modeling, and programming skills are all crucial in solving the above mentioned problem.