A Hybrid Approach for Real-Time Target Detection/Recognition and Tracking

This Small Business Innovation Research Phase I program will investigate the feasibility of a hybrid approach for real-time target recognition and tracking. Real-time target detection/recognition and tracking are essential for a large number of defense and commercial applications. Particlarly for applications that must be performed in rapidly changing enviroments such as surveillance, automatic target recognition, battle field monitoring, mobile platform navigation, and situation awareness in hostile environments. Amherst Systems Inc. proposes to exploit the currently available technologies to investigate the feasibility of developing a real-time target detection and tracking system based on a hybrid approach. The proprosed program is an integration of biologically inspired approaches and conventional image processing. It attempts to exploit the advantages of the conventional image processing methodologies, the active nature of active computer vision, and the robustness of neural network based approaches. The tagert recognition and classification are perfromed in four steps: detection, verification, normalization, and recognition. Conventional image processing methodologies are used to identify and to prepare normalized data sets for neural networks to achieve reliable classification. The proposed approach can use both filtering based and feature based methods to achieve target detection. The proposed technology is parallel in nature.