摘要
在诸多图像边缘检测的算法中,Canny算法具有边缘检测性能良好、不易受噪声的影响、可以在噪声和边缘检测间获得很好的平衡等优点,不足之处是计算量大、实时性差。针对Canny算法存在的计算量大和实时性差等问题,提出了一种并行实现策略,即基于Threading Building Block(TBB)工具和C++语言的并行Canny算法,可以有效的缓解上述问题。通过串、并行算法执行时间的比较,证明了并行算法的优越性。
Compared to other image edge detection algorithms,the Canny algorithm features better edge detection performance,stronger noise tolerance,better balance between noise and edge detection,but has deficiencies such as heavy computation and poor real- time performance. For the problems of heavy computation and bad real- time performance in the Canny algorithm,the paper proposed a parallel implementation strategy,i. e.,a solution using threading building block and the parallel Canny algorithm in C + +. An execution time comparison was made between serial algorithm and the parallel algorithm to indicate the advantage of the parallel algorithm.
出处
《南阳理工学院学报》
2014年第3期47-50,共4页
Journal of Nanyang Institute of Technology
基金
河南省基础与前沿技术研究计划项目(132300410439)
河南省基础与前沿技术研究计划项目(122300410302)
河南省基础与前沿技术研究计划项目(132300410433)
关键词
CANNY算法
TBB
并行算法
Canny algorithm
threading building block
parallel algorithm
