Application of wavelet-based method to level detection of oil wells

Level detection of oil wells is a key problem in petroleum prospecting, on which the subsequent recovery largely depends. However, because the signal is weak and disturbance is wide-spread, conventional time-frequency methods are of little effect. For this reason, a wavelet-based denoising algorithm is designed in this article, which exploits the zooming behavior of wavelet transform and different changing trends between signal and disturbance in the wavelet transform. Proper wavelet bases and decomposing scales are chosen on actual request, along with comparison plots of original signal and denoised signal. Analysis on these results shows that the wavelet-based method increases the SNR greatly, and at the same time keeps the position information of the detected signal. Therefore, it is suitable for signal detection problems of this kind.