【摘要】 研究了舰船辐射噪声的 Lipschitz奇异性指数的分类性能及其提取方法。提出对舰船噪声进行自适应 Gauss平滑滤波是一种可取的前置处理方法 ,它可以提高特征参数的识别与分类能力。对实测的舰船信号的仿真实验证实 ,在分类和识别舰船被动目标前 ,进行适当的滤波 ,再提取Lipschitz指数可使得所提取的特征识别和分类能力更高更多还原

【Abstract】 In Ref.[7] we proved that the energy Lipschitz singularity is quite good for recognition of underwater target. But unfortunately stochastic noise interferes with effective recognition. We propose to use adaptive Gaussian smoothing filter to reduce stochastic noise and in section 1 we discuss theoretically how the singularity feature of the real signal can be enhanced by using adaptive Gaussian smoothing filter. The measured noise signals are the same as used in Ref.[7]. There were 560 samples; the samples for three different types of noise signals were 176, 262 and 122 respectively. After pretreatment with adaptive Gaussian smoothing filter, the recognition rates were 90.91%, 83.21% and 91.80% respectively, considerably higher than recongition rates obtained in Ref.[7]. The test results show preliminarily that our pretreatment method is quite effective.更多还原