MXPA99004032A - Method of reduction of noise from seismic data traces - Google Patents

Abstract

A method of suppression of burst coherent noise in seismic data is provided comprising:a) comparing a threshold amplitude characteristic acceptance value for a time window of a reference trace to an amplitude characteristic of a test trace of a set of traces within the window and, b) applying a non-zero scalar to the test trace in the time window if an amplitude of the test trace within the time window is not within the threshold amplitude characteristic acceptance value.

Description

METHOD OF REDUCTION OF NOISE OF RISKS OF SEISMIC DATA FIELD OF THE INVENTION This invention relates to the field of noise reduction of seismic traces, and more particularly to the reduction of "burst" noise in marine seismic data.

BACKGROUND OF THE INVENTION In some forms of marine seismic data acquisition, probes are towed behind vessels, and include seismic signal receivers (eg, hydrophones). The signals received by the receivers are recorded on the vessel after it is transmitted through cables in the probe, as is commonly known in the art. The boat also towed seismic signal sources (for example air pistols). During the operation, the vessel travels along the line of exploration, periodically firing the source and recording the reflections from the strata of the earth in the receivers. In other forms, cables are placed on the ocean floor and the fountains are towed along the scan line. In other forms, multiple probe and source vessels are used. In either of these ways, the distance between the source and the particular receiver is called the "offset" for that source-receiver pair, and the record received in Any particular receiver for any particular shot is known as a "trail". As is commonly known in the art, the scanning process results in multiple reflections received from the same reflector. Typically, those traces that have common reflections are collected in a common reflection concentrate, each trace of which has a different displacement. Additional processing of the traces is carried out within the common reflection concentrate to eliminate the error introduced by the different displacements (for example, NMO and D O, and other "emigration" algorithms). The variety of processing done in this stage is very large, and is well known to the person skilled in the art. After this processing, the traces are added to each other (stacked in a.k.a.), and the result is another trace. The first trace, which represents the receiving pair, will be referred to herein as a "shooting trace". The second trace, which represents the stacked data, will be referred to as the "stacked trail". Stacking is done in order to eliminate noise, following the theory that noise is often random, or it can be made to appear random. The signal no. As a result, reflections of seismic data must be added constructively, although noise in the shooting trails, when stacked, must be added destructively. For a large amount of noise, this process works very well. However, noise that occurs in bursts, especially noise that occurs in patterns, does not It is necessarily effected through the process of stacking common reflection concentrates. For example, such noise occurs when another exploration vessel is nearby and fires an air pistol in a periodic pattern. This seismic interference can travel a surprising distance, depending on the particular location, water depth and reflection coefficient of the water bottom. Especially when trying to receive reflections from deep layers, the interference of other seismic vessels is an acute problem. Currently, the industry has resorted to the very expensive and time-delayed method of "sharing time", whereby only one vessel in a given area operates at a given time, although many vessels could operate together in the area, if not by the interference of the sources of the respective vessels. Another source of burst noise is sometimes called "expansion" noise. This noise is not well understood. However, it occurs as the seas increase. With some modern acquisition vessels, the expansion noise becomes the limiting factor on the height of the seas in which the vessel can operate. As seen before, there is a need to reduce the amount of burst noise in seismic data. In a recent industry study, the proposed solutions to the problem included the desynchronization of the sources between any two vessels working in a given area. Also, it was observed that the most harmful noise occurs when the sources are on the side of the ship. The tests for this particular industry study identified a "dangerous" segment from noise sources located in a segment approximately 60 to approximately 120 ° from the line to be avoided. Another suggestion for dealing with pop noise is included in "Auutomatic Surgical Blanking of Burst Noise in Marine Seismic Data," by A. J.

Berni, from Shell Development Compani, S 8.2, which can be found in the following publication: SEG Annual Meeting Expanded Technical Program Abstracts With Biographies, 1987, incorporated herein by reference. As described in the article, Berni attempts to "whiten" the noise from the source of interference. Berni observes that by dividing the data into common offset distance panels, each trace on the common scroll panel is of a different shot. Therefore, the interference will appear as a burst, unless the two origins are synchronized. The lack of trace-to-trace continuity in the common displacement panel is advantageous, according to Be i, because it distinguishes the interference of all other signals in the data. The reflected energy, refracted and diffracted from the source itself, all shows lateral continuity in the common displacement panel. Berni's approach is then to bleach regions that show poor lateral continuity. It suggests doing this by dividing the panel into equal time gates, approximately 300 milliseconds in duration. The individual amplitude value is calculated for each gate in each trace. The absolute amplitude The average of the data samples and the particular gate is the preferred Bemi method, observing that RMS measurements or the sum of squares is also adequate. Bemi also represents each time gate by a single amplitude value to reduce the computational load. The value for a given gate will then be purchased with others for the same travel distance, where it is swept through the common offset distance panel, and amplitude values are detected which are anomalous for that time gate. The gates with anomalous amplitude values are selected for "surgical" bleaching. Berni mentions two methods to detect anomalous amplitude values. One method consists of laterally smoothing the sequence of amplitude values for a time horizon in the common displacement panel to obtain a reference value. This value is compared with the reference value of the close gates. The gate is bleached, or zeroed, if the ratio between a particular gate and the reference amplitude exceeds a certain threshold, typically three or more. Berni suggests that bleaching should only be applied to high-level noise bursts that can not be overcome by stacking. It also suggests completely removing weak traces, which he calls "failed shots." The second method is the use of the average of the nearby gate altitudes to obtain the reference value. Typically, 15 gate amplitudes are used on each side of the candidate trail in an average calculation of 31 points. Berni believes that a baseline calculation based on the mean is tolerant of fluctations of Occasional erroneous amplitude due to failed firings or random bit errors in the digitized samples. Unfortunately, the procedure suggested by Bemi removes too much seismic information in its bleaching and eliminates traces. In addition, it has been found that the process suggested by Berni, contrary to his assertions, eliminates real reflections. Other attempts to treat the mido are observed in several references. For example, in U.S. Patent No. 5,555,530, issued September 10, 1996 to Meehan, in an application filed December 6, 1993 and incorporated herein by reference, a method is described for improving the signal-to-mido ratio of a pair of detectors such as geophones, which each detect a mido signal comprising a signal of interest (S) and a signal of measurement (N), wherein the signal of interest (S) has a different eviction through the pair of detectors of the measured signal (N), and the measured signal (N) of a given source is detected in the first detector at a delta time t before detecting the corresponding mido signal in the second detector. Meehan delays the measured signal (S + N) detected in the first detector by means of an amount greater than the eviction of the signal of interest, but not more than delta t, and subtracts the delayed signal from that detected in the second detector by means of an adaptive filter, to reduce the power in the resulting signal. In the United States patent no. 5,424,999, issued to Manin on June 13, 1995, incorporated herein by reference, describes a method to handle an aspect of the bursting problem that requires very accurate knowledge of each vessel and source, to remove the bursts in later processing. In the United States patent no. 5,293,352, issued to Chambers on March 8, 1994 and incorporated herein by reference, describes a method for classifying common trip concentrates into common receptor concentrates, by migrating the wavefield envelopes of common receptor concentrates, using half of the near-surface velocity to provide groups of migrated data, and then reclassifying the migrated data groups into common trigger concentrates and subtracting the common trigger concentrates from the original new common trigger concentrates to provide datasets from reduced coherent mido. In the United States patent no. 5,237,538, issued to Linville, and others on August 17, 1993 and incorporated herein by reference, the removal of estimated coherent mido within a collected signal plus signal noise is handled. In the United States patent no. 5,182,729, issued to Duren et al., On January 26, 1993, and incorporated herein by reference, describes a seismic exploration and data processing method for filtering, from seismic data, any energy contribution that is not in the profile seismic line. The scanning method requires a receiver geometry in which at least one receiver is inside and at least one The receiver is outside the profile line, so that in the common displacement or firing concentrate registration graphs, seismic events along the profile line are aligned with time, and the out-of-plane energy takes the form of a series of seismic events in the form of saw teeth, arriving at different times in different lines of receptors.

Knowing the number of off-plane energies and their respective inclinations, allows the design of a filter, which is applied to the data of a computer for nullification and thereby remove the effects of energy Not wanted out of plane of seismic exploration data. U.S. Patent No. 4,937,794, issued to Marschall et al., June 26, 1990, and incorporated herein by reference, discloses a method for the suppression of coherent mido on seismic records by reforming common shot concentrates into receiver concentrates common. Seismic trace pairs of a common receptor concentration are corrected for differential normal eviction, weighted in inverse relationship to the RMS signal power, and combined to generate common compressed receptor concentrates. Compressed common receptor concentrates can be reformatted as common midpoint concentrates for postepor processing. Other earlier references found in the area search include the following, all of which are incorporated herein by reference: U.S. Patent No. 4,910,716, issued to Kiriip et al., March 20, 1990, U.S. Patent Number 4,707,812, issued to Martínez on November 17, 1987, patent of E.U.A. No. 3,704,444, issued to Schmitt on November 28, 1972, patent of E.U.A. No. 3,611, 279, issued to Hensley on October 5, 1971, patent of E.U.A. No. 3,344,395, issued to Silverman et al. On September 16, 1967, U.S. Patent No. 2,733,412, issued January 31, 1956 to Alexapder and others, and the patent of E.U.A. No. 4,210,968, issued July 1, 1980 to Lindseth. In none of the above references has a method been found to reduce the coherent or bursting mido efficiently without removing the signal together with the mido. The Industry continues to use the expensive solution of sharing time. Consequently, there is a need for a method of dealing with bursting noise that does not eliminate large amounts of real reflection data, and which reduces the need for the expensive timeshare process.

BRIEF DESCRIPTION OF THE INVENTION The object of the present invention is to handle the problems described above. Therefore, according to one aspect of the invention, a method of reducing the noise is provided from traces of seismic data, the method comprising: a) comparing a threshold acceptance value of characteristic amplitude for a time window of a reference trail with a characteristic width of a trail of a group of trails within the time window; and b) applying a non-zero scalar number to the trace of the time window, if the characteristic breadth of the trace within the time window is not within the acceptance value of the characteristic amplitude threshold, where the application of the The non-zero scalar number puts the breadth of the pmeba trail within the time window, to be less than the breadth of the reference trace. According to another aspect of the invention, a method of processing a set of seismic data traces is provided, wherein the method comprises: a) comparing a characteristic amplitude threshold acceptance value for a time window of a trace reference with a characteristic width of a trace of a trace group within the time window; b) apply a non-zero scalar number to the trace of the time window, if the characteristic amplitude of the pmeba trace within the time window is not within the characteristic amplitude threshold acceptance value, where the application of the non-zero scalar number sets the breadth of the pmeba trail within the time window to be below the breadth of the reference trace, where a group of modified data is defined; c) store the scalar numbers, together with identification information of the time window and trace to which the scalar number is applied; and d) recover the original data from the modified data and stored scalar numbers. The benefits of the invention include, in addition to handling the aforementioned problems, reduction of the low amplitude measure light quantity, reducing at the same time a greater amount of high amplitude measurement. In the application of said method, the signal is preserved in the traces. In addition, since the low amplitude noise is attenuated by a small amount, the threshold can be set very close to the expected signal level. In this way, timeshare can be eliminated, or significantly reduced.

DESCRIPTION OF THE DRAWINGS For a more complete understanding of the present invention, and for further advantages thereof, reference is made to the following detailed description taken in conjunction with the accompanying drawings in which: Figure 1 is a group of common moving traces including mido of explosion. Figure 2 is a modified group of the traces of Figure 1, after having applied an embodiment of the present invention. Figure 3 shows a graph of new shots with discharge mido. Figure 4 shows new shots of figure 3 after the application of one embodiment of the present invention. Figure 5 shows the section of difference between figures 3 and 4. However, it is noted that the attached drawings illustrate only typical embodiments of this invention, and therefore are not considered limiting of its scope, since the invention can admit other equally effective modalities.

DETAILED DESCRIPTION OF THE INVENTION Referring now to Figure 1, a group of common channel (and hence common displacement) marine traces 10 is observed, including reflections 12 of seismic reflectors (not shown) and a burst measuring region 14. Each trace 10 it is taken from a different source throughout the exploration. It will be noted that the burst measurement 14 occurs at different times along each trace. This location depends on the location of the interference source (in the case of another vessel source mido), the depth of water, the reflection coefficients of the water depth, and other parameters. However, according to the present invention, knowledge of said parameters is unnecessary, as will be evident below. The present invention operates in regions in which the amplitude of the reflections 12 is below that of the interference measurement 14. Therefore, according to one embodiment of the invention, a characteristic amplitude of the measurement 14 is used to detect the portion of trace 10 in which there is measured 14. As used herein, "characteristic amplitude" refers for example to amplitude RMS, peak amplitude, rectified amplitude, nth root of the sum of all amplitudes raised to the n power , or other characteristics that depend on the amplitude. For example, another characteristic amplitude includes a ratio of the RMS value within a window of a trace to a reference RMS value taken for example from a similar window on a reference trace. A non-zero scalar number is applied to that portion to reduce the measurement without reducing all the reflector signal that may have occurred on the trace at the same time. The result after the application of the non-zero scalar number to the trace of Figure 1, is seen in Figure 2. Since the mido does not appear on the entire trace, no portion of the traces must be scaled.

. Therefore, the portion of trace 10 that includes the mido must be identified 14. According to an example identification process, a first time window W1 of a reference trace 10a is chosen for the data at particular depth D, and a threshold acceptance value of characteristic amplitude, Athres, is chosen to be about the characteristic RMS amplitude of seismic reflection and below the characteristic RMS amplitude of mido. According to some embodiments, the characteristic amplitude of seismic reflection for window W1 derives from averaging through a distance window Wd (not shown), which corresponds to that portion of common displacement concentrate approximately within a diameter and half of the fresnel zone of the dominant frequency for a particular depth of interest. This limitation is an ideal, although if the average is carried out through a distance Wd greater than said diameter, the process will work although at the risk of increasing the damage of real reflections data. The goal of any limitation in the distance Wd on which the average is made, is to avoid averaging over changing geology regions. It is recommended to test a site, by means of which a common displacement section is displayed, differentiated with and without the process of the invention, and a difference section is inspected to determine if any geological structure appears in the section Of diference. In the case where the geology appears, the process must be run again, using a Wd window of smaller distance between the traces to average. By taking the reference trail, alternative modalities are acceptable. In an example, for each window, the characteristic trace of the average amplitude is determined and used. In other examples, the mean, the alpha-adjusted measure, or the alpha-adjusted median are used. It should be noted that the time window can be defined as a number of samples in the digital medium and the time units are used here only as a convenience. In addition, although the above procedure was described with respect to common displacement concentrates, common or common mid-point receptor concentrates are also used, according to alternative embodiments of the invention. The common displacement concentrates do not work well in cases of steep immersion structure, where the geology changes abruptly, from trace to trace. In such a case, the narrow windows (W1, W2, W3), whose sample horizontally shifts the concentrate, will see the structure as an abnormally high amplitude in only a few traces. Consequently, the acceptance value of the characteristic amplitude threshold will be set too low, and the signal will be attenuated. The use of a domain in which the signals of interest appear in the same time window or approximately in the same window through the concentrate, reduces this problem. The common midpoint domain places the structure closest to the horizontal, especially after applying NMO. However, the common midpoint domain is necessarily sensitive to inaccuracies in NMO. Consequently, the common receptor domain, which is less sensitive to inaccuracies in NMO, is another useful domain. In the common receiver domain, there is more sample than in the common displacement and therefore it is more likely that the signal appears in sufficient traces to be chosen an exact threshold value. By taking the acceptance value of characteristic amplitude threshold Athres, it is convenient to make the choice as close as possible to the highest signal resistance expected for a particular depth of interest. However, as with the process of taking the Wd distance window, it is also recommended to perform a site survey. Again, the appropriate section is deployed, differentiated both with and without the process of the invention, and a difference section is inspected to determine if any geological structure appears. In the case where the geology appears, the process must be run again using a higher Athres. In practice, a scale of between about 120% to about 500% of the characteristic RMS amplitude of the window containing the expected reflection 12 has been found acceptable. In a particular experiment, a value of approximately 170% worked well. In addition, although the RMS value of characteristic amplitude in the currently described example, it will be appreciated that peak values and other values related to the characteristic amplitude are also acceptable. It should also be noted that the window length of time depends on the measure. It will be larger or shorter, depending on the characteristics of the particular measure. For example, according to some embodiments of the invention, the time window W1 is chosen to be within about 20% the period of the dominant frequency of the expected measurement. In deep water, for interference from other vessels, a window of approximately 50 msec has been found useful. In modalities that attempt to reduce the expansion rate, which is relatively low frequency, a window of up to 500 msec is effective. In some cases of interference and expansion measurement, the measure is in the characteristic of a pulse train. In such cases, the time window used should be approximately the train's width. Also, it has been found that the bursting mido has a frequency spectrum that is different from that of the desired signal. Therefore, according to some embodiments of the invention, the Athres characteristic amplitude is used within the time window in a limited frequency band of between about 40 Hertz and about 200 Hertz. In other embodiments, a bandwidth of frequencies between about 10 Hz and 200 Hz is effective.

It has been found that a sampling rate of the data is useful, in some embodiments, of approximately 2 milliseconds, again depending on the high frequency content of the seismic interference measurement. Said fine sampling allows to register high frequencies safely. There is an acceptable scale to sample the data between approximately 1 millisecond and approximately 4 milliseconds. It was also believed that sampling rates above about 4 milliseconds would be effective in certain circumstances, depending on the bandwidth of the measure. Turning now to a further aspect of the invention, the Athres characteristic amplitude threshold acceptance value is compared to a characteristic breadth of a trace of a series of trails within the time window, and a non-scalar number is applied. zero to the traces of time in the window of time, if a characteristic amplitude of the trace of the pmeba within the time window exceeds the threshold acceptance value of characteristic amplitude. When applying the non-zero scalar number, several methods are used. In an example, the scalar number is applied evenly across the entire window. In others, the scalar number is applied in a distribution centered on the RMS peak within the window. Other methods of applying the non-zero scalar number will be apparent to the person skilled in the art. More specific examples will be detailed later. According to another modality, after having carried out the previous steps, the result is a modified series of traces, and the Same method to the modified trace. It has been found that a routine of two or more steps (that is, a "cascade routine") generally works better than a single step. According to another modality, the process, either alone or in cascade, is applied along the entire trail with sliding windows. An overlap (between about 10% and about 80%) of the windows is used in some embodiments, but such overlap is not necessary. In one example, it was found that an overlap of about 50% is effective. In a further example, the overlap is as large as possible, the window being moved only one sample along the trace at the same time. According to additional modalities, the non-zero scalar number is a multiplicative function of a relation of the characteristic amplitude of the reference trace and the characteristic amplitude of the trace of the time window. In one modality, the ratio is multiplied by a value less than 1; meanwhile, in another modality, the relationship is elevated to a positive power. In an experiment, it was found that a power of 4 is useful. According to more modalities, after the application of the scalar number to the traces, additional pre-stacking processing of the traces is performed (for example NMO, BMD, displacement migration). zero, and another filtration and deconvolution), although there is no particular order to follow. Thereafter, the traces are stacked and post-stacking processing of the stacked tracks is performed, as in the prior art, to opt for a scanning section for interpretation.

Referring now to Figure 3, a new stack with bursting mido is observed. After processing as described above, with a time window of 50 milliseconds in a 2 millisecond sample across a complete bandwidth with a threshold value of 170%, and raising the detected ratio below the threshold value of the reference value and the characteristic amplitude of the pmeba trace in the first window up to a power of 4, the new stack of figure 5 is the result. In Figure 5 a difference graph of Figure 3 and Figure 4 is observed. A specific system that has been found acceptable for use with various embodiments of the present invention is the PGS ENSBAL ™ program, run on a PGS CUBE MANAGER platform. ™ in a massively parallel processing machine (for example an Intel 1860). In such a system, the ENSBAL ™ program sweeps along a trace, reporting a correction factor (this is a characteristic amplitude) needed to be the RMS value of the trace within a window, equal to an average RMS value within from the same window of other traces, chosen from a group of common close traces of a window distance of approximately the Fresnel zone width. Note that the median RMS value for a window is not necessarily the same as the median RMS value for another window. Then, if the scalar number reported by the ENSBAL ™ is below an threshold amount (that is, a characteristic amplitude threshold acceptance value) then the correction value it is raised to a power, resulting in a scalar correction number that is applied to the trace window of pmeba. When applying the scalar number, it is convenient to avoid sudden changes. The scaled changes in the scalar number, from sample to sample, should not be greater than approximately 30% and preferably should be less than approximately 10%. Therefore, according to one embodiment of the invention, a noise reduction method of a series of seismic data traces is provided, comprising: a) comparing a characteristic amplitude threshold acceptance value to a characteristic amplitude of a trace of a trace of a trace within the time window; and b) apply a non-zero scalar number to the trace in the time window if the characteristic amplitude of the trace of the time within the time window is not within the characteristic amplitude threshold acceptance value, where the application of the number scaling not zero puts the breadth of the trail within the time window to be below the breadth of the reference trace. It should be noted that for the correction factor to be "within the characteristic amplitude threshold acceptance value", as the phrase is used here, the correction factor must be greater than the characteristic amplitude threshold acceptance value, which is modality is a relationship. In other embodiments, for example, those in which the characteristic amplitude threshold acceptance value is the RMS value of a window of a trace, the characteristic amplitude is "within the acceptance value of characteristic amplitude threshold "if it is below the characteristic amplitude threshold acceptance value." Returning now to a modality in which the characteristic amplitude comprises a correction factor, the application is made for multiple windows along the trace of the amplitude. and comprises: calculating a correction factor for each time window that, if applied to the pmeba trail within the time window, would modify the pmeba trail within the time window to equalize the characteristic amplitude of the pmeba trail in the time window with the characteristic amplitude of a corresponding reference trace window, apply a scalar number of unit correction of approximately 1, where the correction factor is above the characteristic amplitude threshold acceptance value plus a constant , where a non-mido window is defined; raising the correction factor to a power in which the correction factor is below the acceptance value of the characteristic amplitude threshold, where a scalar number of non-unitary correction is defined and where a measurement window is defined; calculate a first interpolation between the scalar correction number applied in a window adjacent to the measurement window and the scalar number of non-unitary correction; apply, between windows of adjacent noises, the first interpolation that has non-unitary scalar correction numbers; apply, between windows of adjacent mids to non-mido windows, a second interpolation between one and the characteristic threshold acceptance value, for all points on the first interpolation between the value Acceptance of characteristic amplitude threshold plus a constant, and the characteristic amplitude threshold acceptance value; and applying the first interpolation between the characteristic amplitude threshold acceptance value and the non-unitary correction scalar number for all the points on the first interpolation between the acceptance value between windows of measurement adjacent to non-measured windows. of characteristic amplitude threshold and the scalar number of non-unitary correction. According to an alternative modality that uses a correction factor, the application is made for multiple windows along the path of the pmeba and includes: calculating a correction factor for each time window that, if applied to the pmeba trail within the time window, it would modify the trace of the time window to match the characteristic amplitude of the trace of the time window with the characteristic amplitude of a corresponding reference trace window; applying a scalar number of unit correction of about 1, wherein the correction factor is above the acceptance value of characteristic amplitude threshold plus a constant, where a non-measured window is defined; dividing the correction factor for a measurement window between the characteristic amplitude threshold acceptance value, by means of which a normalized correction factor for the measurement window is defined; apply the normalized correction factor to the noise window; and apply an interpolation between the scalar correction number applied in a window adjacent to the measurement window and the scalar number of non-unit correction.

In a specific example, the characteristic Athres amplitude is chosen at a value of 0.7. For a do not measure gate, the scalar number has a value of 1. In an adjacent mode gate, the scalar number has a value of 0.3. In this case, there is a linear interpolation between 1 and 0.3. Moving down the slope of the interpolation for each sample between the gates, if the interpolation is at least Athres plus a constant (v. Gr. 0 1) then the scalar number of correction is set to one no correction is applied. However, for a position in which the interpolation is between 0.8 and 0.7 (for the example where the constant is 0.1), the scalar number is set to a value that by itself interpolates between 1 and 0.7. After the first interpolation reaches 0.7, the scalar number is set to the first interpolation value. At a gate difference of 500 msec, the values in the previous example work very well. In other embodiments, other scalar number application distributions, eg, Gaucian distribution, are also within the scope of the invention. In addition, it should be noted that other interpolation schemes will be sufficient. In another modality, the non-zero scalar numbers are saved, it is mapped with the information corresponding to the specific advantage on the specific trace to which it is applied. The purpose of this storage is to allow the recovery of the data in an unmodified form, if desired, without the need to save a completely separate series of data. The mapped scalar numbers also give an indication of how much the data was modified. Finally, in one modality, the correction factors are saved and map, which are used to determine a threshold acceptance value of appropriate characteristic amplitude. For example, a graph of the distributions of all correction factors must show two distributions. A first distribution occurs around the signal level and a second distribution occurs around the level of measurement. Among them is a region from which the characteristic amplitude threshold acceptance value is chosen. In another modality, the scalar threshold number is taken by analyzing the signal and mido levels of the series of traces on a base limited to the bandwidth, but the scalar number is applied across the entire bandwidth of the windows. Example of acceptable bandwidths for the analysis are, for seismic interference, between approximately 50 Hz and approximately 150 Hz, and for expansion measurement between approximately 5 Hz to approximately 40 Hz. The modalities described above are given by way of example only and additional modalities will be apparent to the person skilled in the art, without departing from the spirit of the present invention.

Claims (52)

1. NOVELTY OF THE INVENTION CLAIMS 1. - A method of reducing noise in a series of seismic data traces, comprising: a) comparing a threshold acceptance value of characteristic amplitude for a time window of a reference trace, with a characteristic amplitude of a trace of It shows a series of traces within the time window; and b) applying a non-zero scalar number to the trace of the time window, if the characteristic amplitude of the trace of the time within the time window is not within the characteristic amplitude threshold acceptance value, wherein the application of the The non-zero scalar number places the breadth of the trace of the pomba within the time window to be below the breadth of the reference trace.
2. 2. A method according to claim 1, characterized in that the application of the non-zero scalar number is performed uniformly through the time window.
3. 3. A method according to claim 1, characterized in that the application is performed on a distribution centered on the RMS peak value within the time window.
4. 4. A method according to claim 1, characterized in that the trace element comprises common traces of movement. 5. - A method according to claim 1, characterized in that the series of traces comprises traces of common channels. 6. A method according to claim 1, characterized in that the series of traces comprises common midpoint traces. 7. A method according to claim 1, characterized in that the series of traces comprises common receiver point traces. 8. A method according to claim 1, characterized in that the characteristic amplitude threshold acceptance value is a function of the average peak value within the time window of a reference trace. 9. A method according to claim 1, characterized in that the characteristic amplitude threshold acceptance value is a function of the RMS value of the time window of a reference trace. 10. A method according to claim 1, characterized in that the characteristic amplitude threshold acceptance value is a function of the rectified average value of the time window of a reference trace. 11. A method according to claim 1, characterized in that the characteristic amplitude threshold acceptance value is within approximately 120% and approximately 500% of the amplitude of the time window of a reference trace. 12. A method according to claim 1, characterized in that the threshold acceptance value of characteristic amplitude is a function of the amplitude of the time window of a reference trace of between about 10 to about 200 Hertz. 13. A method according to claim 1, characterized in that the characteristic amplitude threshold acceptance value 5 is a function of the amplitude of the time window of a reference trace of between about 40 to about 200 Hertz. 14. A method according to claim 1, characterized in that the time window of a reference trace is less than about 20% of the period of a dominant frequency of measurement. 15. A method according to claim 1, characterized in that the time window of a reference trace is less than approximately, and is approximately the length of a measurement train. 16. A method according to claim 1, further characterized by comprising steps a-b along multiple time windows along the trail of pmeba. 17. A method according to claim 16, characterized in that multiple windows overlap. 18. A method according to claim 17, characterized in that the multiple windows overlap between approximately 20% and approximately 80%. 19. A method according to claim 16, further characterized in that it comprises: providing a modified sequence of traces based on the application of steps a-b to the first series of traces; Y apply, at least an additional time, steps a-b to the modified series of traces. 20. A method according to claim 1, characterized in that the comparison comprises determining the difference in characteristic amplitude in the time window of the reference trail and in the trail of the pmeba, where the characteristic amplitude within the time window of the reference trace is approximately the same as the average of the characteristic amplitude of traces in the time window for the senes of seismic data traces. 21. A method according to claim 20, characterized in that the comparison is made in a bandwidth of between about 50 Hz and about 150 Hz. 22. A method according to claim 20, characterized in that the comparison is made in a bandwidth of between about 5 Hz and about 40 Hz. 23. A method according to claim 1, characterized in that the comparison comprises determining the difference in characteristic amplitude in the time window of the reference trail and in the trail of the pmeba, where the characteristic amplitude within the time window of the reference trace is approximately the same as the alpha-adjusted median of the characteristic amplitude of traces in the time window for the series of seismic data traces. 24. - A method according to claim 23, characterized in that the comparison is made in a bandwidth of between about 50 Hz and about 150 Hz. - A method according to claim 23, characterized in that the comparison is 26.- A method according to claim 1, characterized in that the comparison comprises determining the difference in characteristic amplitude in the time window of the reference trace and in the pmeba trace where the characteristic amplitude within the time window of the reference trace is approximately the same as the alpha-adjusted average of the characteristic amplitude of traces in the time window for the series of seismic data traces. 27.- A method according to claim 26, characterized in that the comparison is made in a bandwidth of between approximately 50 Hz and approximately 150 Hz. 28.- A method according to claim 26, characterized in that the The comparison is carried out in a bandwidth of between approximately 5 Hz and approximately 40 Hz. 29. A method according to claim 1, characterized in that the comparison comprises determining the difference in characteristic amplitude in the time window of the trace of reference and on the pmeba trail, where the characteristic amplitude within the window of The reference trace time is approximately the same as the median of the characteristic amplitude of traces in the time window for the series of seismic data traces. 30. A method according to claim 29, characterized in that the comparison is made in a bandwidth of between about 50 Hz and about 150 Hz. A method according to claim 29, characterized in that the comparison is made in a bandwidth of between about 5 Hz and about 40 Hz. 32. A method according to claim 1, characterized in that the non-zero scalar number consists essentially of a multiplicative function of a characteristic amplitude ratio. of the reference trail in the time window and the characteristic amplitude of the trail in the time window. 33.- A method according to claim 1, characterized in that the non-zero scalar number consists essentially of a function of a ratio of the characteristic amplitude of the reference trace in the time window, and the characteristic amplitude of the trace of the window of time, where the relationship is elevated to a power. 34.- A method according to claim 33, characterized in that the series of traces comprise common traces of displacement through approximately one half times the width of the fresnel zone of a dominant frequency. 35. - A method according to claim 33, characterized in that the characteristic amplitude threshold is a function of the RMS value of the time window of a reference trace. 36. A method according to claim 33, characterized in that the characteristic amplitude threshold acceptance value is between approximately 120% and approximately 500% of the RMS value of the time window of a reference trace. 37. A method according to claim 33, characterized in that the characteristic amplitude threshold is a function of the RMS value of the first time window of a reference trace between approximately 10 Hz and approximately 200 Hz. 38.- A method according to claim 33, characterized in that the characteristic amplitude threshold is a function of the RMS value of the first time window of a reference trace between approximately 40 Hz and approximately 200 Hz. 39.- A method according to the claim 33, characterized in that the first time window of a reference trace is below approximately 20% of the period of a dominant mode frequency. 40. A method according to claim 33, further characterized in that it comprises performing the steps a-b along multiple time windows along the trail of the target. 41. - A method according to claim 40, characterized in that the multiple windows overlap. 42. A method according to claim 41, characterized in that the multiple windows overlap between approximately 20% and approximately 80%. 43.- A method according to claim 40, further characterized in that it comprises: providing a modified series of traces based on the application of steps a-b to the ppmera series of traces; and apply, at least an additional time, steps a-b to the modified series of traces. 44. A method according to claim 1, characterized in that the non-zero scalar number consists essentially of a power of a relation between a scalar number of equalization that, if applied, could make the characteristic amplitude within the window of trace time of pmeba equal to the characteristic amplitude of the reference trace within the time window, and the threshold value of the characteristic amplitude. 45. A method according to claim 1, further characterized in that it comprises: providing a modified series of traces based on the application of steps a-b to the first series of traces; and apply at least an additional time, steps a-b to the modified trace of s. 46.- A method according to claim 1, further characterized in that it comprises: performing additional processing of preapilamento of the s traces; stacking of the s traces; and perform further processing of apiiamento of the stacked traces. 47. A method according to claim 1, wherein the application is made for multiple windows along the trail of the pmeba and comprises: calculating a correction factor for each time window that, if applied to the test trail Within the time window, you could modify the trace of the time window to match the characteristic width of the trace in the time window with the characteristic width of a corresponding reference trace window; on the other hand apply a scalar number of unit correction of approximately 1, wherein the correction factor is above the threshold value of acceptance of characteristic amplitude plus a constant, where a non-measured window is defined; raise the correction factor to a power where the correction factor is below the characteristic amplitude threshold acceptance value, where a scalar number of non-unitary correction is defined and where a measurement window is defined; calculate a first interpolation between the scalar correction number applied in a window adjacent to the measurement window and the scalar number of non-unitary correction; apply, between adjacent windows of mido, the first interpolation that has scalar numbers of non-unit correction; apply, between mido windows adjacent to non-mido windows, a second interpolation between 1 and the characteristic amplitude threshold acceptance value, for all points on the first interpolation between the amplitude threshold acceptance value characteristic plus a constant, and the characteristic amplitude threshold acceptance value; and applying the first interpolation between the characteristic amplitude threshold acceptance value and the non-unitary correction scalar number for all points on the first interpolation between the acceptance value of characteristic amplitude threshold, a scalar number of non-unitary correction. 48. A method according to claim 1, characterized in that the application is made for multiple windows along the trail of the pmeba and comprises: calculating a correction factor for each time window that, if applied to the trace of Within the time window, you could modify the trace of the time window to match the characteristic width of the trace in the time window with the characteristic width of a corresponding reference trace window; applying a scalar number of unit correction of about 1, wherein the correction factor is above the acceptance value of characteristic amplitude threshold plus a constant, where a non-measured window is defined; dividing the correction factor for a window of measurement between the threshold acceptance value of characteristic amplitude, by means of which a normalized correction factor for the measurement window is defined; apply the normalized correction factor to the measurement window; apply an interpolation between a scalar number of correction applied in a window adjacent to the measurement window and the scalar number of non-unit correction. 49. - A method according to claim 48, characterized in that, before applying the normalized correction factor, the normalized correction factor is raised to a power. 50. A method according to claim 1, further characterized in that it comprises determining the threshold acceptance value of characteristic amplitude to be between a statistical identification of a signal characteristic amplitude and a statistical identification of a measurement characteristic. 51.- A method according to claim 50, characterized in that the determination of the acceptance value of the characteristic amplitude threshold comprises: calculating a correction factor for each window of time that, if applied to the trail of the inside of the window of time, could modify the trace of pmeba within the window of time to equal the characteristic amplitude of the trail pmeba in the time window with the characteristic amplitude of a corresponding reference trace window; detect a signal-related distribution of the stored correction factors; detect a distribution related to the measured correction factors; and assigning a threshold value to the characteristic amplitude threshold acceptance value between the signal distribution and the measurement distribution. 52.- A method of processing a series of seismic data traces comprising: a) comparing a characteristic amplitude threshold acceptance value for a time window of a reference trace with a characteristic amplitude of a trace of a series of traces within the time window; and b) apply a non-zero scalar number to the trace of the time window, if the characteristic breadth of the trace within the time window is not within the threshold acceptance and characteristic amplitude value, where the application of the nonzero scalar number takes the breadth of the pmeba trail within the time window to be below the breadth of the reference trace, where a series of modified data is defined; c) store the scalar numbers, along with identification information of the time window and the trace to which the scalar number is applied; and d) recover the original data from the modified data and stored scalar numbers.