Reseach On Denoising For Broadband Signal

Either in life or military, the applications of broadband signal are more and more widely. But in the application process, the noise is inevitably to be mixed in a clean signal, in some cases it even make the signal useless. So denoising for broadband signal becomes a topic with great value and significance.The current noise reduction algorithms can de-noise for narrowband signal very well. On the basis of a comprehensive understanding of current algorithms, this thesis chooses the wavelet threshold de-noising algorithm for wideband LFM signal noise reduction. Then, on the basis of this algorithm, this thesis proposed two algorithms to improve two aspects, and make simulations to analysis the advantages and disadvantages of each algorithm. Finally, this thesis found a noise reduction method for wideband LFM signal noise reduction which is the combination of the two algorithms achieved very good results.The works of this thesis are:1. Introduces the basic theory of wavelet by three levels: the continuous wavelet transform, dyadic wavelet transform and wavelet series. Indicating the principle of the decomposition progressive of wavelet analysis the signal,multiresolution and mallet fast algorithm.2. Introduces the principle and process of wavelet threshold de-noising algorithm, which is very important in three parts: the selection of wavelet basis, the selection and estimation of thresholding function and the evaluation system for the noise reduction effect. Then this thesis selects the appropriate parameters for the wideband LFM signal.3. Through the simulation contrast of the common wavelet threshold noise reduction algorithm, analysis out the limitations of this algorithm. For this limitation, from two aspects that filtering the low-frequency noise and time division, this thesis improved algorithms. Through experiments and simulation proved that the improved method really works.