Edge Density Measure

From OTBWiki
Revision as of 13:02, 20 February 2009 by Otmane (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

This measure is given by eq. 2 in

An Adaptive Multiscale Information Fusion Approach for Feature Extraction and Classification of IKONOS Multispectral Imagery Over Urban Areas Xin Huang; Liangpei Zhang; Pingxiang Li; Geoscience and Remote Sensing Letters, IEEE Volume 4, Issue 4, Oct. 2007 Page(s):654 - 658

Precisely:

  1. Binarize the image using a Canny edge detector (other edge detectors could be used)
  2. Compute the density of edge pixels with a sliding window

In terms of implementation, one should use the same counting framework described in SIFT density image function and filter.


Conclusions :

the following class were implemented :

* otb::BinaryDensityFunction : a density count strategy, implements the evaluate method of the ImageFunction 
* otb::BinaryImageToDensityImageFilter: this class is templated with the count strategy type.
* otb::EdgeDensityImageFilter : this class is templated with the edge detector type and the count strategy type. It uses the class below.