A Self-organizing Fuzzy Segmentation (SOFS) Method for Road Detection from High Resolution Satellite Images

This paper proposes an image segmentation algorithm based on fuzzy logic to detect road pixels in high-resolution satellite images. Other image classification and segmentation approaches require that the number of existing classes in the image scene be specified in advance. Additionally, supervised classification algorithms require sufficient samples from an image to identify the parameters of each class. The proposed method does not require specification of the number of classes. The mean and standard deviation of road pixels are automatically calculated using from 1 to 3 samples from its surface. This is an improvement on existing classification and segmentation methods. The proposed approach is tested on both simulated and real satellite images. A QuickBird color image is segmented into road and non-road zones using mean and standard deviation values calculated using the proposed method. Lastly, advanced mathematical morphology algorithms are used to extract main road centerlines from the segmented image.

  • Availability:
  • Authors:
    • Mohammadzadeh, Ali
    • Zoej, Mohmmad Javad Valadan
  • Publication Date: 2010-1


  • English

Media Info

  • Media Type: Print
  • Features: Figures; Photos; References; Tables;
  • Pagination: pp 27-35
  • Serial:

Subject/Index Terms

Filing Info

  • Accession Number: 01150897
  • Record Type: Publication
  • Files: TRIS
  • Created Date: Feb 4 2010 7:06PM