摘要
边缘是图像的基本特征之一,携带了大量图像信息。边缘检测能够提取边界有用的结构信息,因此边缘检测具有重要作用。论文提出了基于特征值的阳性选择的图像边缘检测算法。该算法基于阳性选择原理,根据特征值匹配规则,综合图像的梯度、非极大值抑制、最大梯度差三个特征值构造"自我集",生成动态检测器对图像进行边缘检测。实验结果表明,该算法在边缘检测中可行,且对比canny和Prewitt边缘检测算法,能得到更好的图像边缘。
Edge is one of the basic features of the image, which carries a lot of image information. Edge detection can extract useful structural information of border and plays an important role. This paper presents an image edge detection algorithm based on positive selection of eigenvalue Set. The algorithm is based on the principle of positive selection. According to a matching rule of eigenvalue, gradient, nonmaxima suppression, the maximum gradient difference, are constructed to 'self set', to generate dynamic detector for image edge detection. Experimental results show that the algorithm is feasible to detect edge and can get better result than Canny algorithm and Prewitt algorithm.
出处
《科技通报》
北大核心
2017年第11期182-185,279,共5页
Bulletin of Science and Technology
基金
国家自然科学基金(No.41261091和No.61163023)
关键词
特征值
边缘检测
阳性选择
动态检测器
eigenvalue
edge detection
positive selection
dynamic detector
