An ultrasonic rock physics data-to-time picking method and system

Abstract

The application discloses an ultrasonic rock physical data to time picking method and system, relates to the technical field of ultrasonic signal processing, and acquires rock ultrasonic waveforms, determines a guide time and a search range according to a user-specified position, and selects a wave peak, a wave trough, a zero-crossing point or a take-off to perform to time picking. Candidate points are generated in a limited range and scored, the time relationship of the candidate points with guide or reference points and the waveform similarity of adjacent records are combined, and corresponding candidate points are selected through dynamic programming. When taking off, the subsequent peak and valley are tracked to determine the reference, and then the candidate points are generated in combination with background fluctuations and subsequent signals, and the to time is supplemented according to the time difference relationship of adjacent records. The to time, record identification, result source and review state are output. The method combines user guidance, local waveform features and the connection between adjacent records to determine corresponding picking positions in each record, and utilizes subsequent peak and valley to assist in positioning when the take-off is weak; and is used for to time measurement of rock ultrasonic sequences.

Description

Technical Field

[0001] This invention relates to the field of ultrasonic signal processing technology in rock physics experiments, and in particular to a method and system for acquiring ultrasonic rock physics data over time. Background Technology

[0002] Currently, in rock physics experiments, ultrasonic wave propagation time is a crucial basis for calculating the longitudinal and transverse wave velocities of rock samples and analyzing their elastic properties. Researchers repeatedly acquire waveforms under different loading stages or other experimental conditions to study the changes in rock properties with varying experimental states. To compare the arrival time changes in different records, it is necessary to select a location from the peak, trough, zero-crossing point, and take-off point, depending on the measurement objective, to pick up the arrival time and determine its corresponding position in each record.

[0003] Existing methods utilize the relationship between the initial location and adjacent records for target tracking. For example, Chinese patent application CN111158046A discloses an automatic seismic horizon extraction device and method based on dynamic programming. This method selects a local window near the initial horizon, flattens it, and then uses dynamic programming to find the optimal path to obtain the target seismic horizon. This method focuses on reflection horizon tracking in seismic profiles, combining the initial horizon with path searching within the local window.

[0004] Rock ultrasonic experiments require distinguishing peaks, troughs, zero-crossing points, and take-off positions within selected local waveforms. As experimental conditions change, the arrival time, amplitude, and waveform of the target signal may alter; crosstalk, reflected waves, and transducer ringing can also affect local interpretation. Within the user-defined search range, one or more candidate locations may exist. When multiple locations are available, the strength of local features at each location may change with the recording, and selecting points line by line based on local features can easily lead to lost correspondences; furthermore, when the range contains multiple cycles, other cycles may be selected instead.

[0005] The determination of the takeoff position is also affected by background fluctuations. When the amplitude of the initial takeoff phase is close to the background fluctuations, the initial arrival position of the signal is difficult to identify, while subsequent peaks or troughs may be clearer. Observing only the local amplitude changes near the takeoff makes it difficult to maintain a consistent interpretation standard across different records. Manual processing requires repeated comparisons of local waveforms and adjacent records, and adjustments and verifications for positions that are difficult to determine, increasing the workload with the number of records.

[0006] Therefore, within the target range specified by the user, it is necessary to determine the corresponding picking position in each record by combining local waveform characteristics and the relationship between adjacent records, and to combine background fluctuations and subsequent signals to assist in positioning when the jump is weak, so as to reduce the selection point offset and manual correction of each record. Summary of the Invention

[0007] To address the aforementioned technical problems, this invention discloses a method and system for acquiring ultrasonic rock physics data at the time of arrival. This method combines user guidance, local waveform characteristics, and the relationship between adjacent records to determine the corresponding acquisition position in each record, and utilizes subsequent peaks and valleys to assist in positioning when the jump is weak.

[0008] To achieve the above objectives, the present invention adopts the following technical solution:

[0009] A method for acquiring ultrasonic rock physics data over time includes the following steps:

[0010] Step S1: Obtain the waveform and time axis of the ultrasonic rock physics data; determine the guidance time and search time range of each record according to the target location specified by the user; and select a location from the peak, trough, zero crossing point and jump point for time picking.

[0011] Step S2: When selecting a peak, trough, or zero-crossing point, generate corresponding candidate positions within the search time range;

[0012] When selecting a takeoff point, first track the subsequent peaks or troughs as a reference, and then generate takeoff candidates near the guide based on the background waveform and the waveform before the reference point.

[0013] The candidate score is calculated based on the waveform amplitude characteristics near the candidate and the absolute value of the difference between the candidate time and the guide time.

[0014] Step S3: Based on the candidate scores, the temporal relationship between the candidate and the guide or reference point, and the waveform similarity near the adjacent candidate records, select the corresponding candidate positions through dynamic programming.

[0015] Step S4: Determine the arrival time based on the selected candidate location, and output the arrival time, record identifier, and result source.

[0016] Furthermore, in step S1, the target positions specified by the user on two different records are connected by a straight line, or a curve is drawn along the corresponding position of the target waveform in each record;

[0017] The guiding time is determined based on the time coordinates corresponding to the straight line or curve at each record, and the search time range for each record is determined by taking the guiding time as the center and combining it with the set search half-width.

[0018] Furthermore, in step S2, when selecting a peak, trough, or zero-crossing point, the waveform within the search time range is smoothed.

[0019] When selecting a peak or trough, candidates are generated at local maxima or local minima, respectively. The vertex of the parabola is determined by the candidate sample points and their adjacent sample points, thereby refining the candidate positions and limiting the refined positions to half a sampling interval before and after the candidate sample points.

[0020] When selecting a zero-crossing point, candidates are generated at the positions where adjacent sample points change from negative to non-negative or from non-negative to negative. Linear interpolation is then performed based on the time and amplitude of the two sample points to determine the time when the waveform reaches zero amplitude.

[0021] Furthermore, in step S2, when selecting the take-off point, the peaks and troughs behind the guide are tracked separately to obtain two sets of picking results corresponding to each record;

[0022] Subtract the corresponding original guidance time from each picking time to obtain two sets of time differences; calculate the comprehensive score of the two sets of results based on the proportion of records with picking results in each set to the total number of participating records, the median of the picking result score, the median of the time difference, and the change in the time difference between adjacent records, select the set with the higher score, and use the peak or trough of that set in each record as the reference point.

[0023] The amplitude threshold is determined based on the background waveform. Combined with the waveform preceding the reference point, jump candidates are generated near the guide and scored. When subsequent signals meet the high-amplitude condition, two methods are used to find jump candidates: one checks where the low-amplitude background extends, and the other checks back to an earlier time from the subsequent high-amplitude signal. Both types of candidates are located near the original guide and earlier than the reference point. Scoring is based on the amplitude of the subsequent signal relative to the background and the degree to which the candidate is close to the original guide; the first method also considers the signal distribution between the candidate and the reference point. When multiple candidates appear at the same location, the one with the higher score is retained.

[0024] Further, in step S3, within continuous segments where each record has candidates, candidates are selected by combining candidate scores, time relationships, and waveform similarity. When selecting a peak, trough, or zero-crossing point, the time difference between the candidate and the guide and its variation among adjacent records are compared; when selecting a jump point, the time difference from the candidate to the reference point is compared among adjacent records. Waveform similarity is calculated using the correlation coefficient of waveform segments near the candidate in adjacent records. The above factors are combined into an objective function, and a candidate is selected from each record using dynamic programming to minimize the objective function value for the entire record segment.

[0025] When choosing a peak, trough, or zero-crossing point, the objective function is:

[0026] ;

[0027] When choosing the takeoff point, the objective function is:

[0028] ;

[0029] in, To select the objective function value at the peak, trough, or zero crossing point, To select the target function value at takeoff; To record the sequence number, a and b are the sequence numbers of the first and last records in the segment, respectively; , and The first The candidate time, candidate score, and guidance time of each record. The value range is from 0 to 1. ; For reference point time, ; For the first Article and No. The correlation coefficient of the waveform segments centered on their respective candidate positions in each record after removing the mean and normalizing; and A positive time scale, using the same units as the time of arrival; , , , and These are the non-negative weights corresponding to each item.

[0030] Further, in step S4, within a continuous record segment with a reference point, the time difference from the selected jump candidate to the reference point is calculated. The median of the existing time differences in each record and its neighboring records is taken, and the estimated time difference for each record is obtained through interpolation and smoothing. For records that do not have a jump candidate, whose candidate score is lower than a set threshold, or whose absolute value of the difference between the time difference and the estimated value exceeds a set tolerance, the time difference is supplemented according to the following formula:

[0031] ;

[0032] in, To record the serial number, , and The first The supplementary time, reference point time, and estimated time difference for each record;

[0033] When the absolute value of the difference between the supplementary arrival time and the guidance time does not exceed the set limit, the supplementary arrival time is used to fill the missing arrival time or replace the selected start-up candidate time.

[0034] Further, in step S4, the arrival time determined by the selected candidate location is marked as direct pickup, the supplementary arrival time calculated from the reference point time and the time difference estimate is marked as time difference estimate, and the supplementary arrival time is marked as pending review, and output together with the arrival time and record identifier.

[0035] The present invention also discloses an ultrasonic rock physics data acquisition system, wherein the processor executes the program in the memory to implement the steps of the above method.

[0036] The beneficial effects of this invention are:

[0037] 1. The candidate search range is limited according to the target location specified by the user, which is applicable to the arrival picking of different positions such as peaks, troughs, zero crossings and take-offs.

[0038] 2. Combining the temporal relationship and waveform similarity of adjacent records to select candidates helps maintain the correspondence of the picking positions, reducing point selection offset and manual adjustment of each record.

[0039] 3. Utilize background fluctuations and subsequent peak-valley references to assist in locating weaker takeoffs, and supplement necessary arrival time estimates based on the time difference relationships of neighboring records. By distinguishing between direct picking and time difference estimation, manual verification is facilitated. Attached Figure Description

[0040] Figure 1 This is a schematic diagram of the overall process of this method;

[0041] Figure 2 The diagram illustrates the guidance and search time range and different picking positions, where (a) is a diagram of the peak guidance and search time range and peak picking time, and (b) is a diagram of the peak, trough, zero crossing point and jump position.

[0042] Figure 3 The diagram shows the candidate and selected locations of peaks, troughs, and zero-crossing points, where (a) is a diagram of the candidate and selected locations of peaks, (b) is a diagram of the candidate and selected locations of troughs, and (c) is a diagram of the candidate and selected locations of zero-crossing points.

[0043] Figure 4 This is a schematic diagram of the determination of reference points during take-off pickup, where (a) is a schematic diagram of the pickup results of the peak and trough, and (b) is a schematic diagram of the scoring comparison and reference selection of the two sets of results.

[0044] Figure 5 The diagram illustrates two methods for generating jump candidates, where (a) is a diagram of determining jump candidates through low-amplitude checks, and (b) is a diagram of determining jump candidates by looking back from subsequent high-amplitude signals to earlier times.

[0045] Figure 6 The diagram shows the initial jump position based on the measured waveforms. (a) is the initial jump position of record 1, (b) is the initial jump position of record 2, (c) is the initial jump position of record 3, and (d) is the initial jump position of record 4.

[0046] Figure 7The diagram shows the time relationship between the jump candidate, the reference peak, and the supplementary arrival time. (a) shows the positional relationship between the selected jump candidate, another candidate, and the reference peak. (b) shows the comparison between the time difference from the initial jump position to the reference peak and the estimated time difference of the adjacent records, and the supplementary arrival time.

[0047] Figure 8 This is a schematic diagram of the functional modules and data transmission relationships of the system to be picked up at that time. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0049] This invention discloses a method for acquiring ultrasonic rock physics data over time, which is described below in conjunction with... Figures 1 to 8 This describes one embodiment of the present invention. The guiding time represents the approximate time corresponding to the target position specified by the user in each record; candidates are potential positions within a defined range that meet the corresponding waveform conditions. The method selects corresponding positions from the candidates in each record to form a picking result. When selecting a starting point, subsequent peaks and valleys are first tracked as references, and then the time relationship between the starting point and the reference is used to assist in candidate selection and timely supplementation. The tracking of reference peaks and valleys adopts the following candidate generation method and the path selection method in step three. The specific process is as follows:

[0050] Step 1: Data Acquisition and Guidance Determination.

[0051] Rock ultrasonic records from the same measurement channel are acquired and arranged in experimental order. In this embodiment, each record has the same start time, sampling interval, and number of sample points. The user selects a position from peak, trough, zero-crossing, and jump for time picking. Two target positions are specified on different records to form a straight line guide, or a curve guide is drawn along the corresponding position of the target waveform in each record. The guide time of each record is obtained by interpolation according to the record position, and the record interval covered by the guide is used as the participation range. When peak, trough, or zero-crossing is selected, the waveform is truncated at half width with the guide time as the center; when jump is selected, the subsequent peak and trough references are determined first, and then the jump analysis segment is determined. Candidate positions are all converted to time according to the original time axis.

[0052] Figure 2 (a) shows the guidance, search time range, and pickup time, with the markers at the baseline indicating the time coordinates; Figure 2(b) shows the positions of the peak, trough, zero-crossing point, and jump point. The zero-crossing point corresponds to the crossing of zero amplitude, and the jump point is determined based on the background fluctuations and subsequent signals.

[0053] Step 2: Candidate generation and scoring.

[0054] (1) Peak, trough and zero-crossing candidate. A symmetrical moving average is performed on the sample points within the search range and the nearby sample points. Peak and trough candidates are generated at the local maxima and local minima of the smooth waveform, respectively. Three-point parabolic interpolation is performed using the candidate sample points and the adjacent sample points before and after, and the peak time is used as the refined candidate time, with the correction range not exceeding half a sampling interval; if the peak cannot be determined, the original sample point position is retained. Zero-crossing candidate is generated when the adjacent sample value changes from negative to non-negative or from non-negative to negative, and the zero-crossing time is determined by two-point linear interpolation.

[0055] The amplitude characteristics of peaks and troughs are calculated based on the amplitude differences between their relative local reference and their neighboring regions. The amplitude characteristics at zero-crossing points are calculated based on the absolute value of the differences between adjacent samples. The local reference is the median of the smoothed samples, and the amplitude characteristics are normalized using the median of the absolute deviation. The more prominent the amplitude characteristics, the higher the amplitude score; the closer the candidate is to the guidance time, the higher the guidance score. Both scores are taken from 0 to 1, and then weighted to obtain the candidate score. The guidance score weight for curve guidance is greater than that for linear guidance. Amplitude comparison uses the samples before refinement. Figure 3 The three candidate categories and the selected positions are shown.

[0056] (2) Jump Candidates. First, shift the original guide to a later time, and generate peak and trough candidates using the methods described above. Then, select the path using the method in step three to obtain two sets of picking results. Compare the effective record ratio, median result score, median time difference relative to the original guide, and the change of this time difference between adjacent records in the two sets. After weighting the scores, select the set with the higher overall score as the reference; the longer the time difference or the greater the adjacent change, the lower the corresponding score. Missing references in the selected set are not filled. Figure 4 The comparison of the two sets of results and the process of determining the reference are shown.

[0057] The analysis waveform, including the jump and subsequent signals, is extracted based on the original guide and reference points, and the waveform immediately preceding it is extracted as the background. A linear trend of the background is fitted, and this trend is subtracted from the background and analysis waveforms, followed by a symmetrical moving average. The median of the background samples is used as the amplitude benchmark, and high and low amplitude thresholds are set based on the background amplitude fluctuations. The high threshold is used to check whether a significant subsequent signal appears, and the low threshold is used to check whether a low amplitude state is maintained before the candidate. Low thresholds are set for both candidate methods. Amplitude comparison uses the absolute value of the difference between the sample and the background benchmark.

[0058] When the proportion of samples reaching the high threshold in subsequent windows meets the requirements, candidates are generated in the following two ways. The first method checks the waveform near the original guide point by point: in the window before a certain position, if the proportion of low amplitude samples meets the requirements, the position is retained; the consecutive retained positions are grouped together, and the last position of each group is taken as a candidate.

[0059] The second method groups windows that meet the high amplitude condition by starting point intervals. Within the first window of each group, the first sample point that reaches the high threshold is taken as the starting point for the backtracking. This process is repeated point-by-point towards earlier times, selecting the first encountered position in the backtracking where all preceding samples are below the low threshold of this method as candidates. If no such position is found, the backtracking starting point is used as an alternative. Both types of candidates must be located near the original guide, earlier than the reference point, and preceded by a complete comparison window.

[0060] The first method evaluates the signal between the candidate and the reference point, while the second method evaluates the signal within the first high-amplitude window after the candidate is generated. The maximum amplitude within the corresponding interval is compared with the background fluctuation magnitude to calculate the amplitude score, and a guidance score is calculated based on the degree to which the candidate closely approximates the original guidance. The first method also considers the proportion of sample points between the candidate and the reference point that reach the lower threshold of the second method. All scores are converted to a value between 0 and 1 and then weighted, retaining the higher score for the same position before being sent to step three. Figure 5 (a) Figure 5 (b) shows the two candidate generation methods respectively.

[0061] Step 3: Select a path by combining time relationship and waveform similarity.

[0062] Within a contiguous segment of records containing candidates, select one candidate from each record. , and Let represent the candidate time, guiding time, and candidate score of the j-th record, respectively. Extract equal-length waveform segments centered on the candidates, remove the mean, normalize, and then calculate the correlation coefficient between adjacent records. The peaks, troughs, and zero-crossing segments are taken from the smooth waveform, while the jumping segments are taken from the input waveform.

[0063] When choosing a peak, trough, or zero point, let By combining candidate scores, guided offsets, adjacent changes in offsets, and waveform similarity, the objective function is made more efficient. Minimum value:

[0064] ;

[0065] In the formula, and The first and last record numbers of the section; Take values ​​from 0 to 1; To use the same positive time scale as at that time; , and The weights are non-negative.

[0066] When choosing to jump, Indicates the time at the reference point, let By comparing adjacent changes in the time difference from the takeoff point to the reference point, and combining candidate scores and waveform similarity, the objective function is optimized. Minimum value:

[0067] ;

[0068] In the formula, It uses a positive time scale and the same units as the time of arrival; and The weights are non-negative, and the other symbols are the same as in the previous formula.

[0069] The time scale is set based on the search half-width and sampling interval to convert each time difference term to a comparable scale; each weight adjusts the impact of time relationship and waveform similarity on candidate scores. When tracking reference peaks and valleys, the shifted guide time and corresponding search half-width are substituted into the peak and valley objective function.

[0070] Dynamic programming is used to accumulate costs line by line, and then backtracking from the last record yields candidate paths for the entire record segment. For peaks, troughs, and zero-crossing points, after the path is determined, the final score is calculated by combining the candidate scores with the time difference changes of the relative guidance between candidates and adjacent records. Under curve guidance, the proximity between the candidate and the guidance is also considered; under straight-line guidance, waveform correlation is also considered. The above factors are weighted according to set weights, and the time when the final score meets the requirements is retained. When tracing reference peaks and troughs, the retained results and their scores are used for the overall comparison in step two.

[0071] Step 4: Determine and output the results at the appointed time.

[0072] The arrival times of peaks, troughs, and zero-crossings are determined by the retained candidates. When selecting a start point, the time difference from the selected start point candidate to the reference point is calculated within a continuous record segment where both the original guide and reference are valid. The median of the valid time differences in the current record and its neighboring records is taken; after obtaining at least two valid medians, linear interpolation is performed according to the record number, with the most recent valid value used at the end, followed by a moving average to obtain the time difference estimate. The neighborhood range and moving average window length are set according to the changes in time difference between records.

[0073] If a record has no jump candidates, the candidate score does not meet the set requirements, or the time difference between the candidate and the reference point differs from the estimated value beyond the set allowable range, then the time difference is calculated using the following formula:

[0074] ;

[0075] In the formula, This is used to supplement the time. When the difference between the supplemented time and the original guiding time does not exceed a set range, it is used to fill in the missing results or replace the initial selected time. Here, two allowable deviations are set: the allowable deviation between the time difference from the candidate to the reference point and the estimated value, which is determined according to the sampling interval and the search half-width; and the allowable deviation between the supplemented time and the original guiding time, which is set according to the proportion of the search half-width.

[0076] Retained candidate arrival times are marked as direct picking, while supplementary arrival times are marked as time difference estimates and indicated as pending review. For direct start results, a result score is calculated based on candidate scores, time difference changes of adjacent records, guide offset, and reference result scores. If the score is below a set threshold, it is marked as pending review. The output includes arrival time, record identifier, result source, and review status. If an arrival time cannot be generated, the guide position is retained separately as a reference for pending review. The result is saved as derived data, while the input waveform and time axis remain unchanged.

[0077] Figure 6 The positional relationship between the initial jump and subsequent waveforms in the four records is shown. Figure 7 (a) in the diagram marks the selected jump candidate, another candidate, and the subsequent reference peak; Figure 7 (b) compares the time difference from the initial jump position to the reference peak with the time difference estimate given by the neighboring records, and marks the supplementary time obtained by subtracting the time difference estimate from the reference peak time. Figure 7 (a) Figure 7 (b) in the text respectively adopts Figure 6 Records 1 and 4 in Figure 7 The records are numbered 1 and 2 in sequence.

[0078] like Figure 8 As shown, the system is implemented by a computer processor executing a program stored in memory. The data acquisition module provides waveforms, a timeline, and recording order. The guidance processing module provides the user-selected peaks, troughs, zero-crossing points, or jump positions, as well as the guidance time and search range. The candidate generation and candidate selection modules interact to complete peak and trough reference determination, candidate generation, scoring, and path selection; when a jump is selected, the selected reference is returned to the candidate generation module. The results output module uses the selected candidates, references, and guidance to complete time-out supplementation, source marking, and output.

[0079] Of course, the above description is not intended to limit the present invention, and the present invention is not limited to the examples given above. Any changes, modifications, additions or substitutions made by those skilled in the art within the scope of the present invention should also fall within the protection scope of the present invention.

Claims

1. A method for timely acquisition of ultrasonic rock physical data, characterized in that, Includes the following steps: Step S1: Obtain the waveform and time axis of the ultrasonic rock physics data; determine the guidance time and search time range of each record according to the target location specified by the user; and select a location from the peak, trough, zero crossing point and jump point for time picking. Step S2: When selecting a peak, trough, or zero-crossing point, generate corresponding candidate positions within the search time range; When selecting a takeoff point, first track the subsequent peaks or troughs as a reference, and then generate takeoff candidates near the guide based on the background waveform and the waveform before the reference point. The candidate score is calculated based on the waveform amplitude characteristics near the candidate and the absolute value of the difference between the candidate time and the guide time. Step S3: Based on the candidate scores, the temporal relationship between the candidate and the guide or reference point, and the waveform similarity near the adjacent candidate records, select the corresponding candidate positions through dynamic programming. Step S4: Determine the arrival time based on the selected candidate location, and output the arrival time, record identifier, and result source.

2. The ultrasonic rock physics data acquisition method as described in claim 1, characterized in that, In step S1, the target positions specified by the user on two different records are connected by a straight line, or a curve is drawn along the corresponding position of the target waveform in each record; The guiding time is determined based on the time coordinates corresponding to the straight line or curve at each record, and the search time range for each record is determined by taking the guiding time as the center and combining it with the set search half-width.

3. The ultrasonic rock physics data acquisition method as described in claim 2, characterized in that, In step S2, when selecting a peak, trough, or zero-crossing point, the waveform within the search time range is smoothed. When selecting a peak or trough, candidates are generated at local maxima or local minima, respectively. The vertex of the parabola is determined by the candidate sample points and their adjacent sample points, thereby refining the candidate positions and limiting the refined positions to half a sampling interval before and after the candidate sample points. When selecting a zero-crossing point, candidates are generated at the positions where adjacent sample points change from negative to non-negative or from non-negative to negative. Linear interpolation is then performed based on the time and amplitude of the two sample points to determine the time when the waveform reaches zero amplitude.

4. The ultrasonic rock physics data acquisition method as described in claim 3, characterized in that, In step S2, when selecting the take-off point, first track the peaks and troughs behind the guide to obtain two sets of picking results corresponding to each record; Subtract the corresponding original guidance time from each picking time to obtain two sets of time differences; calculate the comprehensive score of the two sets of results based on the proportion of records with picking results in each set to the total number of participating records, the median of the picking result score, the median of the time difference, and the change in the time difference between adjacent records, select the set with the higher score, and use the peak or trough of that set in each record as the reference point. The amplitude threshold is determined based on the background waveform, and the jump candidates are generated and scored near the guide point by combining the waveform before the reference point.

5. The ultrasonic rock physics data acquisition method as described in claim 4, characterized in that, In step S3, within a continuous record segment with candidates, a candidate is selected from each record using dynamic programming based on the following objective function, so as to minimize the objective function value of the selected candidate combination. When choosing a peak, trough, or zero-crossing point, the objective function is: ； When choosing the takeoff point, the objective function is: ； in, To select the objective function value at the peak, trough, or zero crossing point, To select the target function value at takeoff; To record the sequence number, a and b are the sequence numbers of the first and last records in the segment, respectively; , and The first The candidate time, candidate score, and guidance time of each record. The value range is from 0 to 1. ; For reference point time, ; For the first Article and No. The correlation coefficient of the waveform segments centered on their respective candidate positions in each record after removing the mean and normalizing; and A positive time scale, using the same units as the time of arrival; , , , and These are the non-negative weights corresponding to each item.

6. The ultrasonic rock physics data acquisition method as described in claim 5, characterized in that, In step S4, within a continuous record segment with a reference point, the time difference from the selected jump candidate to the reference point is calculated. The median of the existing time differences in each record and its neighboring records is taken, and the estimated time difference for each record is obtained through interpolation and smoothing. For records that do not have a jump candidate, whose candidate score is lower than a set threshold, or whose absolute value of the difference between the time difference and the estimated value exceeds a set tolerance, the time difference is supplemented according to the following formula: ； in, To record the serial number, , and The first The supplementary time, reference point time, and estimated time difference for each record; When the absolute value of the difference between the supplementary arrival time and the guidance time does not exceed the set limit, the supplementary arrival time is used to fill the missing arrival time or replace the selected start-up candidate time.

7. The ultrasonic rock physics data acquisition method as described in claim 6, characterized in that, In step S4, the arrival time determined by the selected candidate location is marked as direct pickup, the supplementary arrival time calculated from the reference point time and the time difference estimate is marked as time difference estimate, and the supplementary arrival time is marked as pending review, and output together with the arrival time and record identifier.

8. An ultrasonic rock physics data acquisition system, characterized in that, When the processor executes a program in memory, it implements the steps of the method as described in any one of claims 1-7.

Images