Denoising Method Of Seismic Data Based On Curvelet Transform

Seismic prospecting is a common method of oil-gas exploration. At present, with the deepening of petroleum exploration, the geological conditions become increasingly complicated. So I have to continuously improve the effective methods of treatment in getting the seismic data of high signal to noise ratio,suppress the noise and increase the resolution, and achieve high-fidelity.Currently, there are some common denoising methods such as fourier transform, ?- p transform, wavelet transform and so on. But these methods have their own advantages and disadvantages. In wavelet transform, although it is good for the continuous signal effect, but in the high dimensional image signal, it can not effectively remove the noise and express the edge information.This paper presents the different types of noise of generated characteristics and mechanism in seismic data, and analyzed several common denoising methods' advantages and disadvantages and conditions. Based on the wavelet transform,I introduce the curvelet transform. Curvelet transform is a method with different scales and directions, due to the curvelet transform adds directions parameter, it becomes obviously better than wavelet transform in edge characteristic of processing video signal and deep weak signal at seismic event identification.Through the research, I make the basic steps of curvelet transform in denoising and use the theoretical model to explain the curvelet transform in same direction with different scales and same scales with different direction. Using the global threshold and the local threshold to deal with the curvelet transform factor, and throu gh reconstructed images to show the denoising effect. By comparing the theoretical model and actual data of denoising dates in curvelet transform and wavelet transform, I make sure that curvelet transform has an important advantages as well as efficiency.