Computer Engineering and Applications ›› 2008, Vol. 44 ›› Issue (19): 182-183.

• 图形、图像、模式识别 • Previous Articles     Next Articles

Edge detection of inner defect based on cross-image

JIA Chao,MA Guo-lei   

  1. College of Information Science and Engineering,Yanshan University,Qinhuangdao,Hebei 066004,China
  • Received:2007-08-21 Revised:2008-01-04 Online:2008-07-01 Published:2008-07-01
  • Contact: JIA Chao

基于断层图像的内部缺陷边缘检测

贾 超,马国雷   

  1. 燕山大学 信息科学与工程学院,河北 秦皇岛 066004
  • 通讯作者: 贾 超

Abstract: Considering the main character of inner defect in Industry Computer Tomography(ICT) image,a new edge detection method is proposed on the basis of statistical and morphological theory.Firstly,in order to select the gradient threshold of each pixel automatically,a block of 3-by-3 centered at the pixel and the human visual system are taken into account,and so gained the prior edges of defective image.Then,erode the prior edges with multiple morphological structuring elements to obtain only one response to a single edge,or at least a fixed small number of responses and remove noise.Experiments prove that this method can not only detect edge well,but also has the strong ability to suppress the noise in the image.

Key words: inner defect, edge detection, threshold selection, multiple morphological structuring elements

摘要: 针对结构件内部缺陷形状复杂、随机性大及其计算机断层图像噪声严重等特点,提出了一种新的基于统计学和数学形态学原理的边缘检测算法。为了自动提取每一像素点的梯度阈值,选择以该点为中心的3×3区域为研究对象,并考虑到人的视觉对灰度的分辨能力限制,进而得到缺陷图像的预边缘。然后应用多个结构元素对预边缘进行形态学腐蚀操作,以滤除噪声并细化边缘。实验结果表明,所提出的算法不仅具有很好的边缘提取能力,而且抗噪性、稳定性强、鲁棒性好。

关键词: 内部缺陷, 边缘检测, 阈值选择, 多结构元素形态学