EPIA'03 - 11th Portuguese Conference on Artificial Intelligence

EKDB -- International Workshop on Extraction of Knowledge from Data Bases


Session: December 6, 10:0-11:30, Room A
Title: Border detection on remote sensing satellite data using self-organizing maps
Chen Ning and Nuno C. Marques
Abstract: In this paper, a new approach to border detection on remote sensing satellite data is proposed. Kohonen self-organizing maps (SOM) are used as data mining tools to cluster the pixels through an unsupervised process. Then the clusters are visualized on the SOM internal map. From the visualization, the borders can be detected through an interactive way. As a result, interesting patterns are visible on the images. To test our theories, Atlantic Ocean satellite data is used to evaluate the proposed SOM approach. We compare the results of our method with conventional gradient edge detectors.
Back to schedule.