An unsupervised target detection algorithm in SAR images
Lanying Cao · 2007
An unsupervised target detection algorithm in SAR image is proposed in this paper. In the algorithm, 2-D fussy entropy is used as objective function. Ant colony algorithm and genetic algorithm are used to search the optimal thresholds for SAR target detection problem separately. In the ant colony algorithm, the ant move direction is determined by the trail pheromone. Each ant in the colony will generate a path based on the relative positions of the nodes and feedback information about the best paths generated by previous colonies. In the Genetic algorithm, the near optimal results are searched from selection, cross and mutation. Tests results showed that, due to Ant Colony algorithm’s ability of both finding good search paths and escaping from local minima, the proposed method could achieve more stable target detection results than genetic algorithm though its detect performance was similar to that of genetic algorithm.