عنوان مقاله [English]
نویسندگان [English]چکیده [English]
The presence of many types of noises such as random noise in the seismic data cause some problems; so, they must be attenuated in the processing steps. Singular value decomposition (SVD) is a coherency and linear algebra based filter, which can detect horizontal events in the first eigenimages. For random noise attenuation, after geometry assigning, in common depth point (CDP) gather, after velocity analysis and dynamic corrections and before stacking data, SVD is applied to data. The aligned reflectors are detected at first eigenimages, then they are reconstructed; hence another eigenimage, which contains random noise, is zeroed and the random noise will be attenuated. Because the SVD can detect the horizontal event, if static and dynamic corrections are not applied to data correctly and in the common depth point gather, the reflectors have fluctuations and SVD cannot separate between reflectors and random noise viable. In this paper, these steps are applied to a synthetic common depth point gather with various ratios of signal to noise and to a real common depth point gather from one of the Iranian land hydrocarbon field. According to the results, singular value decomposition can attenuate the random noise and preserves the reflectors considrably. Furthermore, this subject is shown in the synthetic data with high noise level (SNR=1).