基于改进的狼群算法的新型广义熵图像分割

Novel generalized entropy image segmentation based on improved wolf pack algorithm

为了更准确地分割出图像中需提取的目标,提出改进的狼群算法与新型广义熵结合实现图像分割。在狼群算法游走行为中引入周期性随机扰动策略动态调整算法权重,并在狼群算法攻击行为中引入混沌全局搜索,将此改进的狼群算法与新型广义熵结合完成图像分割,用峰值信噪比作为图像分割评价指标对结果进行验证。结果表明,该算法能更准确地分割出图像中需提取的目标,比基本狼群算法与新型广义熵结合的分割结果更准确清晰。

In order to segment interested objects in image accurately and quickly,this paper proposed an improved wolf pack algorithm(WPA)combined with the novel generalized entropy to realize the effective segmentation of image targets.First it introduced a periodic random disturbance strategy in the walking behavior of WPA to make it adaptively adjust the inertia weight,and introduced a chaotic global search into the attack behavior of WPA.Then it combined WPA with the novel generalized entropy to achieve image segmentation.Finally,this paper verified the results by using the peak signal to noise ratio image segmentation evaluation index.The algorithm can more accurately segment the desired target in the image,which is more accurate and clearer than the basic wolf pack algorithm combined with the novel generalized entropy.