Request: How to identify rooftops and slums from GeoEye 1

[rahmansunbeam]

Hello all,

This is me again. For the last few weeks i have been studying the GeoEye imageries and learned that it's not only little information that weakens a research, but too much of it almost kills it. The imageries are of 0.5 meter resolution with 4 primary bands (R, G, B, Nir) of Dhaka and have Google Earth like view. I have been struggling (and somewhat succeeding) to extract all the wet and dry natural features from the images, but now the human creations are eating my head.

Studying urban rooftops, slums and roads are pretty hard to identify when there lots of shadows, trees, varying materials for rooftops and sometimes atmospheric haze. I am using Imagine Objective and willing to use eCognition to work things out. I need some solid stuffs,  namely a tutorial or any detailed article describing any suggestions or guidelines which i could use to identify and classify them. Is there any?

No, i was not being lazy, but googled many sites and found that none of them actually tutorialized their actions, and how they classified these features. The usual tutorials (Imagine Objective User Guide and eCognition) merely described that with a very good LIDER images, which is not going to work here. This poor fellow needs a step by step tutor for GeoEye 1 image.