Computer Engineering and Applications ›› 2010, Vol. 46 ›› Issue (18): 164-166.DOI: 10.3778/j.issn.1002-8331.2010.18.051

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

Improving of image edge detection technique

LI Dan1,2,QIAN Jian-sheng1,LU Nan-nan1,ZHOU Jun-min3,TIAN Wen-ke4   

  1. 1.College of Information and Electric Engineering,China University of Mining and Technology,Xuzhou,Jiangsu 221008,China
    2.College of Xuhai,China University of Mining and Technology,Xuzhou,Jiangsu 221008,China
    3.Lu’neng Heze Coal and Electricity Exploitation Company,Heze,Shandong 274000,China
    4.Wangzhuang Coal Mine,Lu’an Huanneng Stock Company,Lu’an,Shanxi 046204,China
  • Received:2009-03-06 Revised:2009-06-02 Online:2010-06-21 Published:2010-06-21
  • Contact: LI Dan


厉 丹1,2,钱建生1,芦楠楠1,周均民3,田文科4   

  1. 1.中国矿业大学 信息与电气工程学院,江苏 徐州 221008
    2.中国矿业大学 徐海学院,江苏 徐州 221008
    3.鲁能菏泽煤电开发有限公司,山东 菏泽 274000
    4.潞安环能股份公司 王庄煤矿,山西 潞安 046204
  • 通讯作者: 厉 丹

Abstract: Compared with some traditional edge detection technique such as Roberts operator,Sobel operator,Gauss-Laplace operator,Canny operator,etc.This paper indicates the traditional algorithms that have drawbacks in the detection accuracy and the antinoise performance.After analyzing the limitation of the above operators,a new method is put forward.On researching for best boundary through iteration arithmetic,it enhances contrast of object and background,object edge,and abstracted object area exactly.It reduces influence of noise to gray threshold in combined with morphology.The results of the experiment pictures indicate that the improved algorithm has the better accuracy,noise against quality,and precision in the edge detection.

Key words: image processing, edge detection, iteration, erosion

摘要: 比较了Roberts算子、Sobel算子、Gauss-Laplace算子、Canny算子等几种传统的边缘检测技术,指出这些算法在边缘检测精度和抗噪声性能方面存在一定的问题。在分析以上算子的缺陷后提出改进的方法。通过迭代算法寻找最佳阈值,增强了目标和背景的对比和目标边缘,准确提取目标区域,并结合形态学进行轮廓提取,有效减少噪声对灰度门限值的影响,通过对实验图像的分析表明,改进的检测算法对图像边缘提取具有较好的检测精度,抗噪能力和准确性。

关键词: 图像处理, 边缘检测, 迭代, 腐蚀

CLC Number: