Options
2000
Conference Paper
Titel
Model based target and background characterization
Abstract
Up to now most approaches of early ROI (Region of Interest) detection algorithms (in most cases MMO (Man-Made Object) detection) use implicit target and/or background models. Examples are direction histogram analysis approaches, POI (Point of Interest) approaches using e.g. local variances, or local adaptive threshold methods. Using implicit model approaches often results in complex and time consuming training and problem domain adaption. A family of explicit model-based recognition and detection methods with a wide range of applications have been further developed. The basis is the use of general models together with adaptive control structures. Two approaches can be highlighted: ROI detection and geo-coding. The detection algorithms of ROls utilize gradient direction models that have to be matched with transformed image domain data. In most cases simple threshold calculations on the match results discriminate target object signatures from the background. The geo-coding approaches extract line-like structures (street signatures) from the image domain and match the graph constellation against a vector model extracted from a GIS (Geographical Information System) data base. Apart from geo-coding the algorithms can be also used for image-to-image registration (multisensor and data fusion) and may be used for creation and validation of geographical maps.