基于奇Gabor滤波器与Rayleigh分布的边缘检测  

Edge Detection Based on Odd Gabor Filter and Rayleigh Distribution

作　　者：朱振峰[1] 卢汉清[1] 

机构地区：[1]中国科学院自动化研究所模式识别国家重点实验室,北京100080

出　　处：《中国图象图形学报》2005年第7期821-827,共7页Journal of Image and Graphics

基　　金：国家自然科学基金项目(60135020和60121302)

摘　　要：为了自适应地更好地进行边缘检测,提出了一个基于奇Gabor滤波器与Rayle igh分布的边缘检测方法。在边缘响应获取中,由于该方法改进了原有的基于奇Gabor滤波器的边缘响应获取方法,从而提高了计算效率,并且避免了原算法对初始梯度方向估计的依赖。同时根据奇Gabor滤波器的一阶微分属性,可对其边缘输出响应利用Rayle igh分布进行拟合。为克服已有算法中阈值选择的局限性,还提出了一种基于Rayle igh分布的非线性自适应阈值选择方法。实验结果表明,所提出的边缘检测方法具有较好的检测性能和自适应性。In order to realize edge detection adaptively and effectively,this paper presents an edge detection algorithm based on odd Gabor filter and Rayleigh distribution.To obtain the edge response,an improved odd Gabor filter based edge response obtaining method was proposed,which increases the computational efficiency and avoids the estimation of gradient direction.For the differential property of odd Gabor filter,the edge response based on odd Gabor filter was fitted by a Rayleigh distribution.To overcome the shortcomings of existing threshold selection approaches,a Rayleigh distribution based nonlinear adaptive threshold determining method was given.Both the synthetic and natural images were tested in our case.The experimental results show that the proposed edge detection method has good detection property and adaptability.

关 键 词：边缘检测 GABOR滤波器 RAYLEIGH分布 自适应阈值 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]