基于时频分析的薄层识别方法研究.doc

摘 要 本文研究的是用于薄层识别的时频分析方法。首先介绍短时傅里叶变换、小波变换、Wigner-Ville分布、S变换、广义S变换等时频分析方法基本理论并对比研究其优缺点。相比较而言，广义S变换具有很好的时频分析效果。 通过建立正弦模型、雷克子波模型、薄层模型，来分析其时间域特征和频率域特征，并对这三种模型对比各种时频分析方法的时频分辨率。通过综合分析，广义S变换方法，其高频曲线与薄层反射系数对应准确，适合做薄层的识别和厚度的估计。并说明时间和频率分辨率广义S变换中的k值的，使得在实际应用中找到适合的参数，从而提高薄层的预测精度。Abstract In this paper, we adopt the method of the time-frequency analysis for thin bed identification. First, we introduce the basic theory of time-frequency analysis method, such as the STFT, wavelet transform, Wigner Ville distribution, S transform and the generalized S transform.At the same time,we study its advantages and disadvantages comparatively. In comparison, the generalized S transform has better effects time-frequency analysis. Furthermore a general idea of the time-frequency analysis about the steps have been obtained. Then, we analyze its characteristics in time domain and frequency domain by means of the establishment of sinusoidal model, Ricker wavelet model and thin bed model, and comparing various time-frequency resolution of time-frequency analysis method for these three models.Finally, we choose the generalized S transform as the time-frequency analysis tool after getting compared and comprehensively.The effects of generalized S transform can be examined by using generalized S transform to extract the single frequency section for thin bed model and the wedge model.We draw a conclusion that, the high frequency curve and thin bed reflection coefficients of the generalized S transform correspond accurately by comparing with other methods.As a result,it is suitable for thin bed identification and estimation of the thickness . Indicating the time and frequency resolution along with the change of generalized S transform of k value change, we can find the right parameters in practical application, so as to improve the prediction precision of thin bed. Key words:time-frequency analysis; STFT; wavelet transform; generalized S transform; single frequency section目 录 第1章 概述 1 1.1 研究意义 1 1.2 时