ssip 2007

Pic 1

Team 5

Texture Classification

INTRODUCTION


We wanted to define which features to extract from the textures and create the texture database system where querying and finding the most similar texture will be possible.

INPUT

Input images such as Landsat terrain, plus sample images of fields/ sea, forest etc.

OUTPUT

To gether labeled scene.

AIM

Segmentation of scene based on texture (and colour) Tools could be wavelets, ANNs, SVMs…

ADDITONAL GOAL

Identification of features such as corners, crossings, tanks, ambushes.
Registration of images from different viewpoints

REMARKS


Difficulty – Reasonably easy