Chaotic random spare ant colony optimization for multi-threshold image segmentation of 2D Kapur entropy

Abstract

Abstract Although the continuous version of ant colony optimizer (ACOR) has been successfully applied to various problems, there is room to boost its stability and improve convergence speed and precision. In addition, it is prone to stagnation, which means it cannot step out of the local optimum (LO). To effectively mitigate these concerns, an improved method using a random spare strategy and chaotic intensification strategy is proposed. Also, its selection mechanism is enhanced in our research. Among the new components, the convergence speed is mainly boosted by using a random spare approach. To effectively augment the ability to step out of LO and to refine the convergence accuracy, the chaotic intensification strategy and improved selection mechanism are applied to ACOR. To better verify the effectiveness of the proposed method, a series of comparative experiments are conducted by using 30 benchmark functions. According to all experimental results, it is evident that the convergence rapidity and accuracy of the proposed method is better than other peers. In addition, it is observed that the capability of enhanced RCACO is more reliable than other techniques in stepping out of LO. Furthermore, an excellent multi-threshold image segmentation method is proposed in this paper. On this basis, image segmentation experiments at low threshold levels and high threshold levels are also respectively carried out. The experimental results also adequately disclose that the segmentation results of RCACO for both multi-threshold image segmentation at a low threshold level and high threshold level, are even more satisfactory compared to other studied algorithms.