Detection of land cover change by integrating information under lattice basing on completeness and consistency

Thuc-Trung Pham (1) , Robert Jeansoulin (1) , Viet Phan Luong (2) , Alexander J. Comber (3)

1 LSIS - Laboratoire des Sciences de l'Information et des Systèmes
2 LIF - Laboratoire d'informatique Fondamentale de Marseille
3 Department of Geography [Leicester]

This paper examines issues in detection of land cover changes. Several causes are identified and treated separately before any relevant "change" decision: (a) errors in signal processing, (b) ontological disagreement on candidate classes, (c) real world updates. Basing on these issues, we propose an information quality measure of information before detecting the land cover change. The approach to determine the change is composed the steps: (i) measurement of data quality, (ii) identification of semantic discordances to determine a possible common ontology of land cover classes, which is considered as an information lattice, (iii) integration of information integrating under a lattice based on quality.

Domaines: Intelligence artificielle [cs.AI]

Thuc-Trung Pham, Robert Jeansoulin, Viet Phan Luong, Alexander J. Comber. Detection of land cover change by integrating information under lattice basing on completeness and consistency. 3rd International Symposium on Spatial Data Quality (ISSDQ'04), Dept. for Geoinformation and Carthography, Vienna University of Technology TUW, Apr 2004, Bruck am der Leitha, Austria. 12pp. ⟨hal-00622336⟩