J4 ›› 2017, Vol. 34 ›› Issue (4): 414-419.

• 图像与信息处理 • 上一篇    下一篇

基于视觉神经元双拮抗机制的汽车外轮廓提取

茅正冲,韩毅,黄舒伟   

  1. 江南大学轻工过程先进控制教育部重点实验室, 江苏 无锡 214122
  • 收稿日期:2016-03-21 修回日期:2016-05-03 出版日期:2017-07-28 发布日期:2017-08-29

Vehicle outer contour extraction based on the dual antagonistic mechanism of visual neurons

  • Received:2016-03-21 Revised:2016-05-03 Published:2017-07-28 Online:2017-08-29

摘要: 针对传统的边缘检测算子一般将彩色图像转化为灰值图像处理,在对汽车轮廓进行提取时,易造成图像色彩特征的丢失,作者提出了将基于视觉神经元双拮抗机制的轮廓提取算法引入汽车外轮廓检测,该边缘提取算法可充分利用图像色彩特征。实验表明双拮抗机制提取算法可完整的提取轮廓,并且可较好的抑制虚假轮廓。进一步地,为了抑制由于汽车车身反光造成的虚假轮廓,更好地提取汽车外轮廓,作者提出了一种自适应双阈值算法,该算法结合扩展的OTSU算法和双阈值算法处理图像。实验表明,该算法可良好的抑制虚假轮廓以及内部无关轮廓,提取的汽车外轮廓较完整连续。

Abstract: Abstract: In view of the fact that the traditional edge detection operator will generally transform color image to gray image processing to extract the vehicle outline,it is easy to lost the image color features.We introduced double antagonistic mechanism of visual neurons contour extraction algorithm into the automobile outer contour detection. The proposed edge extraction algorithm can make full use of the image color feature. The experimental results show that the proposed dual antagonistic mechanism algorithm can extract the contour efficiently, and can effectively suppress the false contour. Further more, in order to suppress the false contour due to the car body reflection, We proposed an adaptive double threshold algorithm to extract better outer contour of automobile. The algorithm combined extended OTSU algorithm with dual threshold algorithm based on image processing. Experiments show that the algorithm can well inhibit the false contour and internal irrelevant contours, extraction of automobile contour is continuous.

Key words: extended OTSU algorithm

中图分类号: