An Impulse Detection Methodology and System with Emphasis on Weapon Fire Detection

This dissertation proposes a methodology for detecting impulse signatures. An algorithm with specific emphasis on weapon fire detection is proposed. Multiple systems in which the detection algorithm can operate, are proposed. In order for detection systems to be used in practical application, they must have high detection performance, minimizing false alarms, be cost effective, and utilize available hardware. Most applications require real time processing and increased range performance, and some applications require detection from mobile platforms. This dissertation intends to provide a methodology for impulse detection, demonstrated for the specific application case of weapon fire detection, that is intended for real world application, taking into account acceptable algorithm performance, feasible system design, and practical implementation. The proposed detection algorithm is implemented with multiple sensors, allowing spectral waveband versatility in system design. The proposed algorithm is also shown to operate at a variety of video frame rates, allowing for practical design using available common, commercial off the shelf hardware. Detection, false alarm, and classification performance are provided, given the use of different sensors and associated wavebands. The false alarms are further mitigated through use of an adaptive, multi-layer classification scheme, leading to potential on-the-move application. The algorithm is shown to work in real time. The proposed system, including algorithm and hardware, is provided. Additional systems are proposed which attempt to complement the strengths and alleviate the weaknesses of the hardware and algorithm. Systems are proposed to mitigate saturation clutter signals and increase detection of saturated targets through the use of position, navigation, and timing sensors, acoustic sensors, and imaging sensors. Furthermore, systems are provided which increase target detection and provide increased functionality, improving the cost effectiveness of the system. The resulting algorithm is shown to enable detection of weapon fire targets, while minimizing false alarms, for real-world, fieldable applications. The work presented demonstrates the complexity of detection algorithm and system design for practical applications in complex environments and also emphasizes the complex interactions and considerations when designing a practical system, where system design is the intersection of algorithm performance and design, hardware performance and design, and size, weight, power, cost, and processing