Sensor Fusion in a Dynamic Model-Based System

An approach to object identification and sensor fusion of images is proposed. Preprocessing techniques which perform a useful and computationally efficient transformation on the image will be considered. Once the image has been transformed to a more efficient format, mathematical morphology and neural networks are proposed to address the issues of edge detection, identification, sensor fusion and model generation from various types of sensor data. After models are generated from each sensor, a weighted average (based upon the accuracy and performance of each sensor) can determine a new or refined object model. The cooperative-competitive neural network will be used to identify the model as a particular class of object.