Updating Battlefield Situation Information Based On Remote Sensing Image

Digital battlefield is the tendency of future information collection and expression in dynamic battlefield state visualization technology. We take an intense research on problems in such field, including localization, change detection and expression for interesting information in battlefield. Main works done are list as follows:1. A brief introduction of the principle and method for image registration is given. The two different concepts, absolute registration and relative registration, are illustrated. On the basis of existed polynomial rectify method, a new partitioned image registration method which takes advantage of SURF correspondences is proposed;2. Fundamental theories about change detection are illustrated as well as some frequently used methods in this field. A dynamic threshold determination method is proposed. With analysis of general features for dynamic target, we designed a fast ROI detection strategy.3. Hu invariant moments are used to construct a database of dynamic target features, by virtue of which a automatic detection and localization procedure is realized to some extent.4. Detected targets are marked and visualized on 3D virtual territory background with OpenGL library.