基于改进狼群算法的三维OTSU阈值法

三维OTSU阈值法引入了邻域中值及均值信息,抗噪性明显提高,但仍存在分割效果不佳的现象,而且阈值维度的增加导致算法运算量庞大、运行时间过长。为提高算法的计算效率和寻优准确率,利用改进的狼群优化算法来搜寻最佳分割阈值。采用花授粉算法优化后计算的最佳阈值调节狼群算法的初始阈值,使狼群在算法初期具有较高的全局探索能力,提高求解速度。为避免算法后期种群聚集的现象,将高斯变异引入围攻行为,设置变异函数,将满足变异条件的个体狼进行变异,避免算法陷入局部最优,提高寻优精度。将改进的算法与传统三维OTSU算法以及CWPA算法优化的三维OTSU法进行图像分割对比,实验显示,改进的算法使分割时间显著降低,并提高了计算阈值的准确度。

A Three-dimension OTSU Threshold Algorithm Based on Improved Wolf Pack Algorithm

The 3-D OTSU threshold algorithm introduces the information of neighborhood median and mean value,and the anti-noise property is obviously improved,but the segmentation effect is still poor.Moreover,the increase of threshold dimension leads to a large amount of computation and a long running time.In order to improve the efficiency and accuracy,the improved Wolf pack optimization algorithm is used to search for the optimal segmentation threshold.The optimal threshold after the optimization of the flower pollination algorithm is adopted to adjust the initial threshold of the wolf pack algorithm,so that the wolf pack has a higher global exploration ability at the initial stage of the algorithm and improves the solution speed.To avoid population aggregation and local extreme,Gaussian mutation is introduced in beleaguering behavior,and the variable function is set to mutate the individual wolves that meet the mutation conditions,so as to avoid the algorithm to fall into the local optimal and improve the optimization accuracy.The improved algorithm is compared with the traditional three-dimensional OTSU algorithm and the three-dimensional OTSU method optimized by CWPA.The results show that the improved algorithm significantly reduces the segmentation time and improves the accuracy of calculating threshold.