Detection of dim targets in high cluttered background using high order correlation neural network

Presents the development and neural network implementation of a high order spatio-temporal correlation scheme for clutter rejection and dim target track detection from infrared (IR) data. The authors first describe the problem of multiscan target detection and then formulate a model for the process. A high-order correlation method is developed to examine the data between consecutive scans. Images of point sources received from IR sensors were processed consecutively using a connectionist high-order correlation network to reject the background clutter without losing the target information. About 95% clutter rejection rate was achieved using this method