IMAGE SEGMENTATION WITH COLOR AND TEXTURE

Authors
  1. Boulter, J.F.
Corporate Authors
Defence Research Establishment Valcartier, Valcartier QUE (CAN)
Abstract
We examined how objects or regions in an image formed by the reflection of radiation from a scene can be segmented using differences in their color and texture. The color features (normalized red content and normalized blue content) permit good spatial resolution of the segmented regions, whereas the textural features (isotropic and linear textures) help to distinguish similarly colored objects. The color and texture features were chosen because of their low dependence on viewing conditions, such as surface orientation and illumination. As an example, we segment an olive-drab colored military truck from its background of vegetation.
Report Number
DREV-4198/81 —
Date of publication
15 Apr 1981
Number of Pages
34
DSTKIM No
81-03117
CANDIS No
33956
Format(s):
Hardcopy;Originator's fiche received by DSIS

Permanent link

Document 1 of 1

Date modified: