混合概率统计分布混响背景下恒虚警检测方法研究

The performance of active sonar systems operating in shallow water is often limited by false alarms termed reverberation that arise from a variety of geologic, biologic, and anthropogenic sources, called clutter.To solve this problem, our project is going to carry out research on CFAR detectors under non-Gaussian reverberation background.The statistics characteristic mixture model of the shallow sea reverberation is introduced, with the theoretic analyzing and experiment data processing.And the reverberation modeling, parameter-estimation of mixture distribution, CFAR method under reverberation in shallow sea are discussed systematically.Then the difference of the statistics characteristic could be obtained.The reverberation and the clutter could be modelled separatly with the difference of the statistics characteristic.We will do research on Fuzzy-CFAR detectors under mixture distributed reverberation background to improve performance of active sonar in shallow sea.

浅海混响是影响主动声纳目标检测性能的重要因素之一。针对这一问题，本项目开展了非高斯混响背景下的恒虚警（CFAR）检测器的研究。开展了利用混合概率分布模型对浅海混响进行建模的研究。通过理论分析与数据处理相结合的手段，探讨混合分布模型的参数估计方法，建立浅海混响模型。在此基础上，提出混合概率统计分布混响背景下的CFAR检测方法。最后使用模糊逻辑的方法设计模糊隶属函数，进一步改进检测器的性能。研究成果将大幅提高浅海混响背景下主动声纳探测性能。

浅海混响是影响主动声纳目标检测性能的重要因素之一。针对这一问题，本项目开展了非高斯混响背景下的恒虚警（CFAR）检测器的研究。开展了利用混合概率分布模型对浅海混响进行建模的研究。通过理论分析与数据处理相结合的手段，探讨混合分布模型的参数估计方法，建立浅海混响模型。在此基础上，提出混合概率统计分布混响背景下的CFAR检测方法。最后使用模糊逻辑的方法设计模糊隶属函数，进一步改进检测器的性能。研究成果将大幅提高浅海混响背景下主动声纳探测性能。