An improved whale optimization algorithm based on multi-populations and multi-strategies for multilevel threshold image segmentation

Multilevel threshold image segmentation is one of the most widely used methods. Thus, the determination of the best thresholds is crucial. Swarm intelligence optimization algorithm has been widely used to solve the optimal thresholds. However, the quality and accuracy of segmented image are not enough. In order to improve the quality and accuracy of segmented images, this paper proposes an improved whale optimization algorithm based on multi-populations and multi-strategies(MMWOA). To verify the performance of MMWOA, two sets of experiments are designed: the first group of experiments consists of 5 benchmark functions; using Kapur and Otsu as the objective functions for the second set of experiments to execute multilevel threshold segmentation on 5 grayscale images. The experimental results show that the MMWOA is better than other three swarm intelligence algorithms in astringency and segmentation quality.