Image segmentation based on improved hybrid krill herd algorithm

Abstract Image segmentation is an important problem in the field of image processing. Traditional multi-threshold image segmentation faces the problems of inefficiency and lack of real-time performance. In order to improve these problems, a Kapur multi-threshold segmentation algorithm based on the improved hybrid krill herd algorithm is proposed, which adopts a novel population iterative approach limited to the top krill to effectively alleviate the phenomenon of excessive aggregation of krill. Building on this foundation, the differential evolutionary algorithm is combined with the krill herd algorithm, which provides a sufficient search range to strengthen the ability of the algorithm to jump out of the local optimum. The experimental results of multi-threshold segmentation on three classical test images show that the proposed improved algorithm can achieve higher segmentation accuracy and quality, and the convergence rate is mostly above 90%, which verifies the reliability and effectiveness of the algorithm in multi-threshold image segmentation problems.