A Neural Adaptive Model for Feature Extraction and Recognition in High Resolution Remote Sensing Imagery

Contextual classification methods, which require the extraction of complex spatial information over a range of scales, from fine details in local areas to large features that extend across the image, are necessary in many remote sensing image classification studies. This work presents a supervised adaptive object recognition model which integrates scale-space filtering techniques for feature extraction within a Multilayer Perceptron neural network and the back-propagation learning task of the search of the most adequate filter parameters. The experimental evaluation of the method has been conducted in an easily controlled domain using synthetic imagery, and in the real domain coping with object recognition in high-resolution remote sensing imagery. To investigate whether the strategy can be considered an alternative to conventional procedures the results were compared with those obtained by a well known contextual classification scheme.