改进的生物地理学算法及其在图像分割中的应用

Improved Biogeography-Based Optimization and Its Application in Image Segmentation

为了增强生物地理学优化算法(BBO)在图像多阈值分割应用中的全局搜索能力,提高其优化性能,提出一种改进的生物地理学算法(IBBO)。首先,引入多源迁移算子,该算子能更好地从搜索空间中生成新特征值,有效提高种群的多样性;其次,创建一种新型的动态变异算子,该算子能够动态地改变变异幅度,提高算法运算效率,使算法快速收敛到全局最优解;随后,将原来的精英选择算子改为贪婪选择算子,即采用优胜劣汰的策略加快算法收敛速度;最后将其应用到基于最大熵的多阈值分割中。图像分割实验结果表明,IBBO算法运行速度远远快于穷举算法,优化性能优于标准BBO算法和PSO算法。

In order to enhance the global search ability of Biogeography-Based Optimization（ BBO） in multi-threshold image segmentation, and improve its optimization performance, an Improved BBO（ IBBO）algorithm is proposed. Firstly, a polyphyletic migration operator is introduced, which can better generate new eigenvalue from the searching space and effectively improve the population diversity. Secondly, a new dynamic mutation operator is created, which can dynamically change the mutation range and improve the operation efficiency of algorithm, enabling the algorithm to quickly converge to the global optimum. Then, a greedy selection operator is used instead of the original elitist selection operator, to accelerate the convergence process by using the strategy of survival of the fittest. Finally, IBBO algorithm is applied to the maximum entropy-based multi-threshold segmentation. Experimental results of image segmentation show that the proposed IBBO algorithm operates much faster than the exhaustive algorithm, and the optimization performance is better than that of the standard BBO algorithm and PSO algorithm.