 |
|
We have been pioneering new techniques in remotely sensed image analysis. This takes satellite images and classifies them according to land use: e.g. road, urban and agriculture together with each subset's constituents (grass, trees etc.). These approaches include using texture as a pre-segmenter (we had to develop a new texture technique to handle multi-spectral imagery), together with a new fuzzy framework for including texture in the segmentation process.
|
