Towards Robust Analysis of Satellite Images Using Map Information - Application to Urban Area Detection
Résumé
With the rapid development in remote sensing, digital image processing becomes an important tool for quantitative and statistical analysis of remotely sensed images. These images contain most often complex natural scenes. Robust interpretation of such images requires the use of different sources of information about the scenes under consideration. This paper presents an integrated approach to the robust analysis of remotely sensed images by using multi-spectral SPOT image data, as well as map knowledge and contextual information. Several techniques are proposed for the effective use of map information for urban area detection in SPOT images. The first one is concerned with the modeling of SPOT images and map information using Markov random fields, which in turn permits application of various existing energy minimization algorithms for solving image analysis problems. The second one is on a new iterative optimization algorithm which automatically adjusts the optimal valures of the parametres of our image model using a feedback control. The third one is about parameter estimation in the Markov random fields which takes into account both the map knowledge and the contextual information. The last one is concerned with the fusion of (intermediate) analysis results by using again the map knowledge in the estimation of the reliability of these results.