The detour foraging and random hiding behaviour of widespread social organisms, such as rabbits, inspires a recently developed Artificial Rabbit Optimisation (ARO). It is a well-suited method to solve complex optimisation problems. In this article, the ARO method is introduced as a solution to the image processing problem. Otsu’s interclass variance is considered a fitness function, which is optimised by the ARO method for multithresholding. The ARO method utilises a population-based approach to explore (detour foraging) and exploitation (random hiding) of the candidate solution (rabbits). This approach is based on individual search (one rabbit) and group-based search (a group of rabbits) to find the global optimum. The proposed multithresholding method adapts by partitioning image intensity histograms into multiple thresholds, facilitating the effective separation of objects and background regions. The fruitfulness of the proposed approach is justified by applying it to different benchmark images, and its effectiveness is demonstrated by comparing the statistical data obtained by the proposed approach with that obtained by other optimisation approaches.