Study On CFAR Detection Algorithm Of Performance For Vibration Signals Detected By Optical Fiber

The topic of this thesis is to deeply investigate the techniques of constant false alarm rate (CAFR) detection, which is among the demands of optical fiber vibration pre-warning system(OFVWS) designing. In order to detect vibration signals in non-stationary background noise, the false alarm rate controlling can be an effective method to denoise sampled signals. Based on the large amounts of spot data, this paper models the statistical characteristics of fiber optic vibration signals and designing the 2-d secondary detection algorithm. In addition, some key techniques, such as CFAR adaptive detection framework, are discussed in depth in this paper.Firstly, in order to solve the instability problem while the vibration signals detected by optical fiber, the detection algorithm for optical fiber vibration is studied. Classical CFAR algorithms is introduced in detail, and the main problem such as the instable detection of a performance problem and poor real-time is pointed out while the algorithm applied for vibration signals detected by optical fiber. For these problems, a detection algorithm of two-dimensional and dual threshold for vibration signals detected by optical fiber is presented based on classical CFAR algorithm. Through this algorithm, the false alarm rate is stable, the detection speed is increased, detection performance is good in nonhomogeneous environments such as multiple targets.Then, in order to solve the instable detection of a performance problem that caused by the different geologic structure and the environment, chapter 4 deals with the problem of adaptive detection technology for optical fibervibration signals. Aiming at building the adaptive detection process for optical fiber vibration signals, the adaptive detection algorithm is studied. Through the adaptive selection, the" cell averaging" can be applied in homogeneous environment, and the two-level detection algorithm that presented in chapter 3 can be applied in nonhomogeneous environments. The proposed algorithm gives better detection probability than those of "order statistics" and the two-level detection algorithm. The practicality and the performance of detection are both improved.Finally, the field experiment and the application is introduced based on the adaptive detection technology for optical fiber vibration at Rushan power supply company in Weihai of Shandong Province which as the experimental station. The signals that human knocking the optical fiber and the ground are collectted, the performance of identification for the exception event type by applying OFVWS is tested. Through the field experiment, the effectiveness practicability of proposed algorithm which can detectted the vibration signals is proofed.