基于多尺度边缘响应函数的自适应阈值边缘检测算法
doi: 10.3724/SP.J.1146.2007.00845
MERF Based Edge Detection with Adaptive Threshold
-
摘要: 该文提出了一种基于多尺度边缘响应函数的自适应阈值边缘检测算法。首先分析二进小波变换,根据边缘和噪声随尺度变化的不同特性,设计了多尺度边缘响应函数(MERF)。通过MERF中的乘积放大作用,增大了边缘响应的幅度,同时也抑制了噪声产生的伪边缘。然后利用小波变换多尺度之间的联合分布关系,计算自适应阈值,检测MERF的梯度模值形成多尺度边缘。该算法直接在小波特征上进行多尺度合成,避免了多个边缘图合成过程的病态问题。实验表明,与LOG,Canny以及Mallat多尺度小波检测方法相比,该算法在检测和定位之间能够达到更好的平衡,既能够实现小尺度下的精确定位,也可以保留大尺度下对噪声的抑制作用。Abstract: An MERF based edge detection algorithm with adaptive threshold is presented in this paper. After analyzing dyadic wavelet transform and different behavior of edge and noise across scales, a Multi-scale Edge Response Function (MERF) is defined as the multiple scales point-wise products of the DWT to enhance significant image structures and suppress noise. Thereafter, an adaptive threshold for MERF is calculated and imposed on the module of MERF to identify edges as the local maxima of the gradient map without synthesizing the edge maps at several scales together, which was employed in many multi-scale techniques. Experiments on synthetic benchmark and natural images showed that the proposed MERF based adaptive threshold edge detection algorithm achieves better detection results than that for a single scale, especially on the localization performance; and edge and noise can be better distinguished by MERF comparing with the Laplacian of Gaussian (LOG), Canny edge detection and Mallat wavelet based edge detection algorithms.
-
Gonzalez R C, Woods R E, and Eddins S L. Digital ImageProcessing Using MATLAB. Beijing, China: PublishingHouse of Electronics Industry, 2005: 287-295.[2]Canny J. A computational approach to edge detection[J].IEEETrans. on Pattern Analysis and Machine Intelligence.1986,8(2):679-698[3]Shin M C, Goldgof D B, Bowyer K W, and Nikiforou S.Comparison of edge detection algorithms using a structurefrom motion task[J].IEEE Trans. on System, Man, andCybernetics.2001, 31(4):589-601[4]Mallat S and Hwang W L. Singularity detection andprocessing with wavelets[J].IEEE. Trans. on InformationTheory.1992, 38(2):617-643[5]Mallat S and Zhong S. Characterization of signals frommulti-scale edges[J].IEEE Trans. on Pattern Analysis andMachine Intelligence.1992, 14(7):710-732[6]Yang H and Zhang W. Research on image edge detectionbased on multi-scale wavelet transform and fuzzy clustering.Computer Science, 2006, 33(1): 174-176.[7]Zhang G and Liu Q. Robust edge detection based onstationary wavelet transform. Journal of Southeast University,2006, 22(2): 218-221.[8]Demigny D. On optimal linear filtering for edge detection[J].IEEE Trans. on Image Processing.2002, 11(7):728-737[9]Wei H and Shen L. Multi-scale edge detection by usinganti-symmetrical biorthogonal wavelets. Acta ElectronicaSinica, 2002, 30(3): 313-316.[10]Patel J K and Read C B. Handbook of The NormalDistribution. New York, USA: Marcel Dekker, 1996: 288-293.[11]Pratt W K. Digital Image Processing. 3rd Edition.[J].New York,USA: John Wiley Sons.2005,:-
计量
- 文章访问数: 3828
- HTML全文浏览量: 113
- PDF下载量: 1642
- 被引次数: 0