Font Size: a A A

Steerable Filtering By Radon Transform And Its Applications On Signal Enhancement For Seismic Data

Posted on:2015-01-14Degree:MasterType:Thesis
Country:ChinaCandidate:J W LiFull Text:PDF
GTID:2250330428497799Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
Signal enhancement and noise attenuation is an important research topic ofmodern signal processing. In the field of oil and gas exploration, researchers are oftenbased on actual seismic data acquisition, analyze the useful signal records todetermine the distribution of wells of the oil and gas exploration and locations ofdrilling. But the manual collection of seismic data is often accompanied by randomnoise. When the random noise can be excessive, it will seriously reduce the signal tonoise ratio of seismic data, thereby the continuing exploration work is adverselyaffected. Therefore, to enhance effective seismic signals and attenuate the noisecontains in seismic data, has a far-reaching significance for verifying geologicalstructure, exploration of oil and gas and other mineral resources.Steerable filters were proposed by W. T. Freeman and Edward H. Adelson in1991. It is a kind of filters that can analyze multiple directions. It is based on the polarform of the Fourier series expansion of a function. The basic idea is that the steerablefilter in an arbitrary orientation can be synthesized by taking a linear combination of aset of based filters, as if the filter can be rotated in any direction, hence the name.Important consideration is the number of the basic filters, as well as the interpolationfunction before we lead a linear combination of them. And the image filtered at anarbitrary orientation can be synthesized by a linear combination of the images filteredwith basic filters(this is because the convolution operation is linear). The steerablefilter has the advantage of simple algorithm, high speed operation. Currently steerablefilters have been applied in edge detection, image fusion, image enhancement, targettracking, feature extraction, pyramid analysis and other fields of vision and imageprocessing.Radon transform can be understood as an image projection in r-qspace, eachpoint in the space represents a straight line in Rectangular coordinate system, so it hasa role in the direction detection.Based on the difference between the effective directional signal and noise inactual seismic record, and the relations between directivity of steerable filters and the enhanced edges, for the first time, the steerable filters and Radon transform arecombined to enhance the useful signal in seismic records. Firstly, we obtain thesynthetic analog seismic records according to the characteristics of the seismic wavesfrom the geophones. Then using the Radon transform theory, accurately obtain theorientation information of the analog seismic data, and use the correspondingdirection of steerable filter to process it finally. Simulation results show that thefiltered record shows the location of the events of the original records well, and theresolution of events and the signal to noise ratio of the analog signals is greatlyimproved. Eventually this algorithm is applied to process actual seismic data, andcombined with sub-block ideas in practice, the results show that this algorithm canenhance useful seismic data, attenuate noise, and improve signal to noise ratio of theoutput and the resolution of events.
Keywords/Search Tags:Seismic data, steerable filters, Radon transform, seismic signal enhancement
PDF Full Text Request
Related items