基于不可分加性小波与形态学梯度的图像边缘提取方法  

Image edge detection method based on nonseparable additive wavelet and morphological gradient

作　　者：刘斌[1] 孙斌[1] 关淼苗 邢倩[1] 

机构地区：[1]湖北大学计算机与信息工程学院,湖北武汉430062

出　　处：《量子电子学报》2015年第6期654-662,共9页Chinese Journal of Quantum Electronics

基　　金：国家自然科学基金(61471160);湖北省自然科学基金重点项目(2012FFA053)

摘　　要：针对传统的图像边缘提取方法只强调图像中的水平和垂直边缘的不足,提出了一种基于不可分加性小波和形态学梯度相结合的图像边缘提取方法。根据二维不可分小波理论构造了低通滤波器,利用它对原图像进行加性小波多尺度分解;对低频子图像求形态学梯度,对增强后的高频子图像取模极大值;将所得梯度图与边缘图作加性小波逆变换,得重构后的边缘梯度图;并利用二值形态学方法对其进行处理,得最终结果边缘图。实验结果表明,此算法可获得较好的边缘图像,与经典的边缘提取方法相比,它具有完整性、多方向性、平移不变性和快速性的特点。In order to solve the problem that the traditional image edge detection techniques only emphasizethe horizontal and vertical edges, a new image edge extraction method combining the multi-resolutionanalysis of nonseparable additive wavelet transformation and the mathematical morphology was proposed.A low-pass filter of nonseparable wavelet was constructed and the source, image was decomposed into a low-frequency sub-image and high-frequency sub-images. The low-frequency sub-image was filtered to obtainmorphological gradient map. The high-frequency sub-images were added to form an enhanced high-frequencyedge map by using modulus maxima value method. Inverse transform of nonseparable additive wavelettransform was performed on the gradient map and edge map and an edge-gradient map was produced.The final edge map was obtained by binarizing the edge-gradient map. The experimental results show thatthe proposed method has good visual effect. When compared with the traditional image edge detectiontechniques, it can extract image edges with integrity, multidirection, shift invariance and high speed.

关 键 词：图像处理 边缘检测 不可分加性小波 形态学 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]