DocumentCode
2004160
Title
MOSAIC: a model-based change detection process
Author
Stossel, Bryan J. ; Dockstader, Shiloh L.
Author_Institution
Commercial & Gov. Syst., Eastman Kodak Co., Rochester, NY, USA
Volume
2
fYear
2002
fDate
8-11 July 2002
Firstpage
1113
Abstract
Eastman Kodak has developed a technique called MOSAIC, for Multi-modality Operational Site Analysis and Intelligent Change detection. This technique is based on the application of faceted site models and synthetic image generation (SIG) tools in tandem with material identification using relative reflectance properties derived from imagery and/or feature maps. This presentation will step through key accomplishments in the past two phases of the projects evolution; the technique feasibility demonstration and initial development accomplished under NRO/AS&T charter, and subsequent application of this technique towards the NIMA objective of automatic change detection and update of GIS features including buildings, roads, and river-ways using 3-D volumetric models. Performance examples will be provided, along with a discussion of relationships to other NIMA and Air Force programs.
Keywords
feature extraction; image processing; image reconstruction; sensor fusion; Eastman Kodak; MOSAIC; Markov modeling; aerial imaging; change detection; faceted site models; feature extraction; feature maps; material identification; model-based fusion; relative reflectance; remote sensing; synthetic image generation; volumetric model; Atmospheric modeling; Change detection algorithms; Face detection; Government; Image analysis; Image generation; Object detection; Performance analysis; Reflectivity; Roads;
fLanguage
English
Publisher
ieee
Conference_Titel
Information Fusion, 2002. Proceedings of the Fifth International Conference on
Conference_Location
Annapolis, MD, USA
Print_ISBN
0-9721844-1-4
Type
conf
DOI
10.1109/ICIF.2002.1020937
Filename
1020937
Link To Document