Research And Application On Technology Of Wavelet Denoising Based On DSP

Signal noise reduction is a very important part of signal processing in engineering test.Traditional noise reduction will cause a loss of high frequency and a distortion of the signalfrequency. Wavelet is a new theory, which is developed in recent decades. Wavelet transformhas the characteristic of multi-resolution, which window size is unchanged but the shape canchange, can describe the signal’s local characteristics in the time domain and frequencydomain. The technology of wavelet noise reduction can not only suppress the signal’s noise,but also, it can overcome the shortcomings of conventional noise reduction methods. Aimingat the methods of wavelet reduction noise and its software, this paper carried out a study:1、By simulation and data analysis, the paper selects wavelet threshold as a de-noisingmethod. According to the shortcomings of the hard threshold and the soft threshold, paperadjusts the threshold by introducing an adjustment factor to decrease the adverse effects thatcaused by method of threshold noise reduction.2、According to the internal structural characteristics of the DSP, based on the algorithmimprovement of threshold noise reduction, the paper uses integrated software developmentenvironment of DSP to exploit the system’s noise reduction program.3、In this paper, using the high performance DSP (TMS320C6713) as a signalprocessing tool to reduce the noise of vibration signal which collected from headstock oflathe CA6140, and compared the de-noising result that simulated by MATLAB.Experimental result show that the effect of de-noising is consistent and it prove that thewavelet noise reduction technology for engineering signal that based on DSP is effective andcredible.