Improving Spatiotemporal Change Detection: A High Level Fusion Approach for Discovering Uncertain Knowledge from Satellite Image Database

This paper investigates the problem of tracking spatiotemporal changes of a satellite image through the use of Knowledge Discovery in Database (KDD). The purpose of this study is to help a given user effectively discover interesting knowledge and then build prediction and decision models. Unfortunately, the KDD process for spatiotemporal data is always marked by several types of imperfections. In our paper, we take these imperfections into consideration in order to provide more accurate decisions. To achieve this objective, different KDD methods are used to discover knowledge in satellite image databases. Each method presents a different point of view of spatiotemporal evolution of a query model (which represents an extracted object from a satellite image). In order to combine these methods, we use the evidence fusion theory which considerably improves the spatiotemporal knowledge discovery process and increases our belief in the spatiotemporal model change. Experimental results of satellite images representing the region of Auckland in New Zealand depict the improvement in the overall change detection as compared to using classical methods. Keywords—Knowledge discovery in satellite databases, knowledge fusion, data imperfection, data mining, spatiotemporal change detection.

[1]  Richard O. Duda,et al.  Pattern classification and scene analysis , 1974, A Wiley-Interscience publication.

[2]  Christos Faloutsos,et al.  Prediction and indexing of moving objects with unknown motion patterns , 2004, SIGMOD '04.

[3]  Reynold Cheng,et al.  Uncertain Data Mining: An Example in Clustering Location Data , 2006, PAKDD.

[4]  Sergio Greco,et al.  A Hierarchical Algorithm for Clustering Uncertain Data via an Information-Theoretic Approach , 2008, 2008 Eighth IEEE International Conference on Data Mining.

[5]  Mohamed Ben Ahmed,et al.  Multiapproach System Based on Fusion of Multispectral Images for Land-Cover Classification , 2008, IEEE Transactions on Geoscience and Remote Sensing.

[6]  J. Ross Quinlan,et al.  C4.5: Programs for Machine Learning , 1992 .

[7]  Padhraic Smyth,et al.  From Data Mining to Knowledge Discovery: An Overview , 1996, Advances in Knowledge Discovery and Data Mining.

[8]  Hans-Peter Kriegel,et al.  Density-based clustering of uncertain data , 2005, KDD '05.

[9]  Dimitrios Gunopulos,et al.  Efficient Mining of Spatiotemporal Patterns , 2001, SSTD.

[10]  Mohamed Ben Ahmed,et al.  Interpretation of Multisensor Remote Sensing Images: Multiapproach Fusion of Uncertain Information , 2008, IEEE Transactions on Geoscience and Remote Sensing.

[11]  Peter Clark,et al.  The CN2 induction algorithm , 2004, Machine Learning.

[12]  Yan Huang,et al.  A Framework for Mining Sequential Patterns from Spatio-Temporal Event Data Sets , 2008, IEEE Transactions on Knowledge and Data Engineering.

[13]  George Kollios,et al.  Mining, indexing, and querying historical spatiotemporal data , 2004, KDD.

[14]  Padhraic Smyth,et al.  From Data Mining to Knowledge Discovery in Databases , 1996, AI Mag..

[15]  Florian Verhein.  Mining Complex Spatio-Temporal Sequence Patterns , 2009, SDM.