Multiscale Markov Random Fields for Large Image Datasets Representation

{ Future users of satellite images will be faced with a huge amount of data. The development of \Content-based image retrieval algorithms for Remote Sensing Image Archives" will allow them to eeciently use the upcoming databases of large images. Here, we present an image segmentation and feature extraction algorithm, that will enable users to search images by content. In our approach, images are modelled by multiscale Markov random elds (MSRF). This model is superior to spatial Markov random eld models in that it is able to describe the long range as well as the short range behaviour of the image data. Image information extracted at multiple scales is incorporated naturally in the model. Additionally it is computationally less costly than the spatial random eld models. The diierence to similar work 1], 2] is that the multiscale process is not only used to nd a reasonable segmentation of the image, but that the estimated parameters of the scale process serve also as image features. These image features together with the textural characteristics of the image are stored hierarchically in a pyramidal structure from large to small scales. Thereby even large datasets can be browsed fastly.