Multilevel thresholding selection based on the fireworks algorithm for image segmentation

With the increasing number of the threshold, the computation of the multilevel minimum cross entropy thresholding will increase exponentially, and the processing efficiency will be low, thus it is difficult to be applied in real-time processing. Some classical optimization algorithms, such as genetic algorithm, particle swarm algorithm has been used to deal with such problems, but it is easy for them to fall into the local optimal solution, the performance is not robust. In this paper, we use the minimum cross entropy to define the objective function of the optimal image segmentation thresholding, solve the optimization problem with the new intelligence optimization algorithm–fireworks algorithm, and compare it with other algorithm. The experimental results show that the fireworks algorithm can solve the problem of multilevel thresholds image segmentation with minimum cross entropy, which is a promising multilevel thresholding method, and it is not easy to fall into local optimal solution.