基于自适应粒子群优化的三维OTSU图像分割算法

摘要

Aiming at the weakness of the huge calculation of the Three-Dimension OTSU, a method for image segmentation based on Three-Dimension OTSU and adaptive Particle Swarm Optimization (PSO) is presented. Firstly,the target region is segmented through optimal entropy method, then based on the mean gray value of the target region specify the background area ranges of Three-Dimension OTSU,finally the ultimate target image is obtained via the Three-Dimension OTSU and adaptive Particle Swarm Optimization. Comparing with the image segmentation based on 3-D maximum between cluster variance, the simulation results show that the performance of this method is superior in operation time.%针对三维OTSU分割算法运算量大、计算时间长的问题，提出了一种基于自适应粒子群优化的三维OTSU图像分割算法。首先采用最佳熵的方法初步提取图像的目标区域，根据该目标区域特征自适应地调整三维OTSU算法的背景搜索范围，然后采用三维OTSU算法并结合粒子群优化最佳分割阈值对原图像进行分割。实验结果表明,与三维OTSU阈值分割方法的递推算法相比，该方法能够进一步减少运算时间。