Skip to Main content Skip to Navigation
New interface
Journal articles

Généralisation sélective en mode image d'une classification satelittale

Abstract : The generalisation of a classified stelitte image roughly consists in smoothing the boundaries of the patches and eliminating the too small ones. This process allows adjusting the level of detail of the spacial representationto th of the document as well as to the desired output scale. It also simplifies the possible vectorization of the image, for example for input and furher processing in a vector based GIS. We first review the different techniques encountered in the litterature for raster generalisation. Then, we present the Lansat TM image classification over Vietnam taken here as example, and show that the different classes it contains do not need the same kinf of generalisation. The different processing we suggest for the image relies mainly on image flitering and mathematical morphology operators, the latter being slightly to suit the particularies of « class image » processing, as opposed to binary or grey level images. The main originality of this work relies on the specific processing applied to some classes related to their specific characteristics. For example, the « water » class, although not widespread over the image, has a high thematic importance and is also accurately delimited thanks to a good spectral discrimination. This class will have its spatial extend little modified, only slightly exaggerated by removing the small gaps in its patches. An other important feature of the proposed methodology is the aggregation of highly scattered or interspersed classes which would otherwise be very difficult to generalise without inducing biases.
Document type :
Journal articles
Complete list of metadata
Contributor : Vincent Simonneaux Connect in order to contact the contributor
Submitted on : Monday, September 14, 2015 - 7:01:32 PM
Last modification on : Wednesday, November 16, 2022 - 11:46:08 AM


  • HAL Id : ird-01199059, version 1



Vincent Simonneaux. Généralisation sélective en mode image d'une classification satelittale. Teledetection, 2000, 2 (1), pp.47-64. ⟨ird-01199059⟩



Record views