- 學位論文
水下聲源訊號雜訊消除之研究
The Research of the De-nosing for Underwater Acoustic Signal
摘要
水下的各種聲源訊號在海洋傳播過程中，由於受到海洋環境之影響，與海洋各種噪音之干擾，而具有隨機及時變的特性，因此接收後必須作適當的訊號處理，才能將經過長距離傳送而能量衰減，且被環境干擾的訊息分辨出來。本研究對水下聲源訊號作雜訊消減之系統設計，主要是以小波轉換為基礎加上基因演算法之參數選擇之方式作雜訊消減之處理。此方法可分為三個步驟：(1) 水下聲源訊號做小波轉換 (2) 對小波係數進行縮移 (3) 利用反小波轉換重建調整過的信號。而在第二步驟中，則利用基因演算法來求取最佳臨界值，對小波係數進行縮移。 在實作部分，利用兩種不同含有雜訊的訊號來測試此水下聲源雜訊消減系統。第一種是基本的測試訊號，如方波及鳥啾聲等；第二種則是實際的水下聲源訊號。並且用最小平方差和雜訊比來評估此系統和另外兩種典型小波轉換去雜訊方法的效能。由實驗結果顯示，本論文提出的改進方法，可以達到較佳的去雜訊效果。
並列摘要
During propagation of the underwater acoustic signal is affected by ocean interference and ambient noise disturbance, it adds random process and time vary characteristics to the signal. Therefore, in order to distinguish the weaken signal caused by long distance propagation loss the received signal must be processed properly. This research takes wavelet-based with choosing thresholding value by genetic algorithms for de-noising. It can be divided into three stages: (1) Wavelet transform of the underwater acoustic signals (2) Thresholding of wavelet coefficients (3) Inverse wavelet transform to reconstruct modified signals. And in second stage, this research makes use of Genetic Algorithms to obtain the optimal threshold value for shrinking to wavelet coefficients. In experiments, this research demonstrates two different types of noisy signals on the de-noising underwater acoustic signals system. First type is basic test signal, such as Blocks and Chirp so on. Second type is the actual underwater acoustic signals. Then, mean-square-error and signal-to-noise ratio are used to estimate this system and the other two traditional wavelet transform for de-noising methods. According to the outcomes of experiments, the proposed approach can achieve better performance on de-noising.