Cartosat-1 stereo data are provided with RPC (rational polynomial coefficients). You can use the Rational Function Model (available in most commercial s/w e.g. ERDAS LPS, ENVI etc.) for bundle block adjustment and subsequently generate a DEM and orthophoto. It mainly uses digital photogrammetric technique and you may refer any standard book for your understanding of the technique. However, I have one tutorial (attached) which may help you in processing the Cartosat-1 stereo data for DEM generation.
From my point of view, your question can fit within stereo vision problem area.
The main issue is to extract good features from the overlapped images and to select the match points correctly. One of my students had successfully completed a simplified version of DEM (it is a disparity map model since we did not have the scale of the images at the time and calculations are not up to the scale). You can find a description of the work in the following URL:
http://ial.iust.ac.ir/index.php/research
In my opinion the main problems in this area are feature extraction and matching.
- Badges
- Science topic
- Similar topics
- Geoscience
- Remote Sensing