The Research Of Denoising Algorithm For Pipleline Leak Detection

In recent years, pipeline leaks is increasing attract attentions of people,it not only brings huge economic losses, but also causes great waste of water resources.So,timely and accurately to find leaks and to repair has great significance for the conservation of water resources.The denoising algorithm of pipeline leaks detecting this design proposed plays an important role in the foundation for the detection and location of pipeline leak points.Methods to define pipeline leak points is calculating the correlation of two leakage signals.But this two leakage signals must be pure signals,we can just make sure the accuracy of leak points.So,the process of de-noising the collected leakage signals is very important.This design studies the denoising algorithm of leakage signals and emphasis on the theoretical study and simulation of denoising algorithm the leakage signals used.Because of different surroundings, noise Included in leakage signals of water supply pipes is varied. In this design,we use multi-channel filter banks and wavelet noise filtering to denoise signals. Multi-channel filter banks mainly aims at stationary noises and wavelet noise filtering mainly aims at non-stationary noises. In design method of multi-channel filter banks,we use Parks_McClellan law and in design method of wavelet noise filtering,we use wavelet threshold denoising method.After the research of related theories of FIR filter and wavelet Transform,we respectively verified the correctness of both noise filtering methods and then designed appropriate C language program and achieved the simulation of both noise filtering methods by simulating DSP chip TMS320VC5402 in CCS software platform. The experimental results shows that noise filtering methods researched in this design is feasible and valuable in practice.