Real-time Edge Segment Detection with Edge Drawing Algorithm


TOPAL C., ÖZŞEN Ö., Akinlar C.

7th International Symposium on Image and Signal Processing and Analysis (ISPA), Dubrovnik, Hırvatistan, 4 - 06 Eylül 2011, ss.313-318 identifier identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Basıldığı Şehir: Dubrovnik
  • Basıldığı Ülke: Hırvatistan
  • Sayfa Sayıları: ss.313-318
  • Anadolu Üniversitesi Adresli: Evet

Özet

Edge Drawing (ED) is an edge detection method that works by computing a set of anchor points, which are most likely to be edge elements, and linking them with a predefined set of rules which we call smart routing. ED can thus obtain one-pixel wide, contiguous and smooth edges in real-time. In most cases, obtaining the edges in segment form (i.e. blob of connected edgels) is desired. Because edge segments also include the connectivity information among the pixels, they contain useful features about the geometric shape of the detected edges. In contrast to conventional detectors, ED is able to output the resulting edge map as vectorial data, i.e., array of edge segments instead of a 2D binary image. In this work, this capability of ED is explained and discussed with examples.