Title :
Context-Dependent Fusion for mine detection using Airborne Hyperspectral Imagery
Author :
Zhang, Lijun ; Frigui, Hichem ; Gader, Paul ; Bolton, Jeremy
Author_Institution :
CECS Dept., Univ. of Louisville, Louisville, KY, USA
Abstract :
We present a method for fusing the decisions of multiple algorithms that use different hyperspectral imagery (HI) classification methods and apply it to mine detection. The proposed fusion method, called Cumulative Separation-Based (CSB) method, is embedded into our Context-Dependent Fusion for Multiple Algorithms(CDF-MA) framework. The CDF-MA is motivated by the fact that the relative performance of different algorithms can vary significantly depending on the type of the different targets and other environmental conditions. Results on real world HI data show that the proposed method can identify meaningful and coherent clusters and that different expert algorithms can be identified for the different contexts. Our initial experiments have also indicated that the proposed method outperforms all individual algorithms and the global weighted average fusion method.
Keywords :
geophysical signal processing; image classification; image fusion; landmine detection; remote sensing; airborne hyperspectral imagery; classification method; context-dependent fusion; context-dependent fusion for multiple algorithms; cumulative separation-based method; global weighted average fusion method; mine detection; Clustering algorithms; Error analysis; Hyperspectral imaging; Hyperspectral sensors; Probability distribution; Random variables; Region 4; Remote sensing; Testing; Training data; Context-dependent fusion; ensemble classifiers; hyperspectral imagery; mine detection;
Conference_Titel :
Hyperspectral Image and Signal Processing: Evolution in Remote Sensing, 2009. WHISPERS '09. First Workshop on
Conference_Location :
Grenoble
Print_ISBN :
978-1-4244-4686-5
Electronic_ISBN :
978-1-4244-4687-2
DOI :
10.1109/WHISPERS.2009.5288973