An algorithm for phase unwrapping based on swarm intelligence is proposed. The novel approach is based on the
emergent behavior of swarms. This behavior is the result of the interactions between independent agents following a
simple set of rules and is regarded as fast, flexible and robust. The rules here were designed with two purposes. Firstly,
the collective behavior must result in a reliable map of the unwrapped phase. The unwrapping reliability was evaluated
by each agent during run-time, based on the quality of the neighboring pixels. In addition, the rule set must result in a
behavior that focuses on wrapped regions. Stigmergy and communication rules were implemented in order to enable
each agent to seek less worked areas of the image. The agents were modeled as Finite-State Machines. Based on the
availability of unwrappable pixels, each agent assumed a different state in order to better adapt itself to the surroundings.
The implemented rule set was able to fulfill the requirements on reliability and focused unwrapping. The unwrapped
phase map was comparable to those from established methods as the agents were able to reliably evaluate each pixel
quality. Also, the unwrapping behavior, being observed in real time, was able to focus on workable areas as the agents
communicated in order to find less traveled regions. The results were very positive for such a new approach to the phase
unwrapping problem. Finally, the authors see great potential for future developments concerning the flexibility,
robustness and processing times of the swarm-based algorithm.