多核环境下的图像分割并行算法研究  

Research of Parallel Algorithm for Image Segmentation Under Multi-core Environment

作　　者：刘张桥[1] 王成良[2] 焦晓军[1,3] 

机构地区：[1]重庆大学计算机学院,重庆400044 [2]重庆大学软件学院,重庆400044 [3]重庆科技学院电气与信息工程学院,重庆401331

出　　处：《计算机工程》2011年第15期197-200,共4页Computer Engineering

基　　金：重庆市科技攻关计划基金资助项目(CSTC;2009AC2060)

摘　　要：对多核环境下的图像分割并行算法进行研究,在基于正交小波分解的多分辨率图像锥中引入模糊C-均值(FCM)算法,采用OpenMP语言设计P-FCM多核并行模型,并给出该模型的算法实现步骤。在对初始图像数据预处理时,采用矩形块数据分割法进行图像分块,将分块后的子图像数据作为并行运算时的输入数据由主线程分给不同的处理器。实验结果表明,在处理较大图像时,该算法效率较高。This paper studies the parallel algorithm for image segmentation under multi-core environment. It combines the multiresolution image pyramid based on the orthogonal wavelet decomposition with the Fuzzy C-means（FCM） clUstering algorithm, and achieves the multi-processor and multi-threaded parallel model of P-FCM with the parallel language of OpenME It gives the algorithm implementation of the model. For the original image data preprocessing, the parallel method of rectangular block is used to divide the original image data into sub-rectangular data blocks, which will be used as the input data and assigned to different processor by the mater thread for the parallel computing. Experimental results show that, in dealing with the large size images, this algorithm has high efficiency.

关 键 词：图像分割 多核处理器 OpenMP语言 模糊C-均值算法 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]