Uncertainty-based 3-d Fusion of Multi-Sensor Aerial Imagery

ABSTRACT: Multi-sensor fusion is a key capability required for effective inferences for timely and accurate strategy and tactical planning. A multi-sensor fusion technology needs to appropriately weigh different image streams according to the uncertainties and errors across different inputs and parameters such as GPS and INS measurements, sensor model parameters, inherent ambiguities of the data such as localization for featureless region and errors due to registering data in a common frame in each source of data for effective fusion. Vision Systems, Inc. propose to fuse aerial imagery via 3-d models using subjective logic to represent and account for various uncertainties. Additionally, the 3-d models will be registered by computing a 3-d to 3-d transformation. BENEFIT: Remote Sensing, surveying, evacuation and emergency planning, visualization and navigation.