基于线性插值的图像分块分割算法
|
摘要
图像分割是图像处理中的一个重要问题,全局阈值法和局部阈值法是图像分割中广泛采用的两类方法。全局阈值法效率高,但是对局部信息不敏感;局部阈值法对局部信息保存较好,但是计算效率不高。针对这一问题,该文提出一种基于分块的图像分割算法。首先将图像分块,然后,每块内采用最大类间方差法计算阈值,最后,块与块之间采用线性插值的方式平滑分割图像。实验表明,该文提出的算法和全局阈值法相比能较好地保存局部信息,和局部阈值法相比有更高的运行效率,在实际中推荐使用。
Image segmentation is an important problem in digital image processing. Global thresholding and local thresholding methods are widely used in image segmentation. Global thresholding method has the advantage of high efficiency but has the weak ability to extract local feature. Local thresholding method can acquire the local feature easily but has the poor computational efficiency. To solve those problem,an approach which based on linear interpolation of blocks is proposed. Firstly,the image is divide into sub- blocks,Secondly,OTSU's method is used to calculation the threshold of each sub- blocks.Finally,linear interpolation's method is used between sub- blocks. The experiment results show that the proposed method can acquire the local feature effectively compared to global thresholding method but has higher computational efficiency compared to local thresholding method. It is recommended in practical segmentation.
Image segmentation is an important problem in digital image processing. Global thresholding and local thresholding methods are widely used in image segmentation. Global thresholding method has the advantage of high efficiency but has the weak ability to extract local feature. Local thresholding method can acquire the local feature easily but has the poor computational efficiency. To solve those problem,an approach which based on linear interpolation of blocks is proposed. Firstly,the image is divide into sub- blocks,Secondly,OTSU's method is used to calculation the threshold of each sub- blocks.Finally,linear interpolation's method is used between sub- blocks. The experiment results show that the proposed method can acquire the local feature effectively compared to global thresholding method but has higher computational efficiency compared to local thresholding method. It is recommended in practical segmentation.
引文
[1]Otsu N.A threshold selection method from grey-level histograms[J].IEEE Trans System.Man Cyberner,1979,SMC-9:62-66.
[2]Bernsen J.Dynamic Thresholding of Gray-level Images[A].Proc of 8th Intel Conf.on Patt.Recog[C].Pairs,France:IEEE Computer Society Press,1986:1251-1255.
[3]章毓晋.图像分割[M].北京:科学出版社,2001.
[4]王平.基于视觉测量的脉搏信息检测[D].兰州:兰州理工大学,2014.
[5]Xu G L,Qi X P,Zeng Q H,et al.Use of land's cooperative object to estimate UAV's pose for autonomous landing[J].Chinese Journal of Aeronautics,2013,26(6):1498-1505.
[2]Bernsen J.Dynamic Thresholding of Gray-level Images[A].Proc of 8th Intel Conf.on Patt.Recog[C].Pairs,France:IEEE Computer Society Press,1986:1251-1255.
[3]章毓晋.图像分割[M].北京:科学出版社,2001.
[4]王平.基于视觉测量的脉搏信息检测[D].兰州:兰州理工大学,2014.
[5]Xu G L,Qi X P,Zeng Q H,et al.Use of land's cooperative object to estimate UAV's pose for autonomous landing[J].Chinese Journal of Aeronautics,2013,26(6):1498-1505.