MACIEL, Lucas da Silva; ALBERTAZZI, Armando Gonçalves
PUBLICADO EM APPLIED OPTICS, Volume 53, Issue 24, Ed. Optica Publishing Group - OSA. 2014. pp. 5502-5509.
https://doi.org/10.1364/AO.53.005502
Resumo
A novel algorithm for phase unwrapping based on swarm intelligence is proposed. The algorithm was designed based on three main goals: maximum coverage of reliable information, focused effort for better efficiency, and reliable unwrapping. Experiments were performed, and a new agent was designed to follow a simple set of five rules in order to collectively achieve these goals. These rules consist of random walking for unwrapping and searching, ambiguity evaluation by comparing unwrapped regions, and a replication behavior responsible for the good distribution of agents throughout the image. The results were comparable with the results from established methods. The swarm-based algorithm was able to suppress ambiguities better than the flood-fill algorithm without relying on lengthy processing times. In addition, future developments such as parallel processing and better-quality evaluation present great potential for the proposed method.
