A method for detecting cracks by ant tracking based on frequency filtering and frequency division

Abstract

The present invention discloses a method for detecting cracks by ant tracking based on frequency filtering frequency division bodies. The method first divides the original seismic data body into three frequency division bodies of different frequency bands through filtering, which are respectively called high-frequency, medium-frequency and low-frequency frequency division bodies; secondly, the frequency division bodies are structurally smoothed and boundary detected. By comparing with the original seismic data, a reasonable and effective boundary detection result of the frequency division body is preliminarily selected as the input of the next step of ant tracking; thirdly, the selected edge detection result is actively ant tracked to obtain the ant tracking result of the frequency division body. Fourthly, the frequency division ant tracking results are compared, and at least one better result is selected for fusion to generate the final ant tracking fusion body; finally, we automatically extract fractures from the fusion body according to the specified fracture parameters. The present invention makes full use of the sensitivity of narrowband seismic data to underground geological information, and can dig out sharper and more small-scale fractures.

Description

Technical Field

[0001] The invention relates to geophysical exploration, and in particular to a method for detecting cracks by ant tracking based on frequency filtering frequency division bodies. Background Art

[0002] The input data of traditional ant tracking technology is full-band seismic data. The cracks tracked by it show weak lateral continuity or unclear planar fracture patterns, and the ability to identify small-scale natural cracks is also limited. In order to improve the image quality of cracks, there are patented inventions that use spectrum decomposition technology for ant tracking. Although good results have been achieved, a large number of sub-frequency bodies may be obtained during the spectrum decomposition process, which not only brings a lot of workload, but also brings many difficulties in judging and distinguishing the differences between adjacent sub-frequency bodies. Summary of the invention

[0003] The purpose of the present invention is to provide a method for detecting cracks by ant tracking based on frequency filtering frequency division bodies, so as to identify small-scale cracks, improve the accuracy of crack interpretation, and save time and effort for fault interpretation.

[0004] The technical solution adopted by the present invention is:

[0005] A method for detecting cracks by ant tracking based on frequency filtering frequency division body, comprising the following steps:

[0006] like Figures 1 to 9 As shown in one, the present invention discloses a method for detecting cracks by ant tracking based on frequency filtering and frequency division, and the specific steps are as follows:

[0007] Step 1, spectrum analysis: perform spectrum analysis on seismic data to obtain effective frequency bandwidth;

[0008] Step 2, frequency filtering and frequency division: using a bandpass filter to filter out low-frequency, medium-frequency and high-frequency components from the seismic data in the effective frequency band according to the set bandwidth step size to obtain three frequency division bodies;

[0009] Step 3, structural guidance filtering smoothing: structural guidance smoothing filtering adopts the "anisotropic diffusion" smoothing algorithm, and its smoothing operation is only performed on the information parallel to the seismic event axis, and no smoothing is performed on the information perpendicular to the seismic event axis. The purpose of structural smoothing of the seismic data volume is to use effective filtering methods to remove noise along the inclination and strike of the seismic reflection interface, increase the continuity of the event axis, improve the lateral resolution of the event axis termination point (fault), and effectively preserve or improve the sharpness of the fault.

[0010] Step 4, boundary detection: perform variance volume operations on the high-frequency, medium-frequency and low-frequency component frequency division bodies respectively to detect the discontinuity points of the seismic data and strengthen this discontinuity; and open a certain time window up and down along the target layer to extract a series of slices along the layer, combine the seismic profile and other data, consider the breakpoint characteristics, and preliminarily select the frequency division body that reflects the breakpoint characteristics better, and enter the next step of ant tracking calculation.

[0011] Specifically, the boundary detection results of the three frequency division volumes may all be selected, or the boundary detection result of only one frequency division volume may be selected.

[0012] Step 5, ant tracking: Ant tracking technology is a biological technology based on ant algorithms. It searches for crack traces in the seismic data body according to the principle that ants use secretions to find food sources as quickly as possible, until the fault is tracked and identified. This tracking technology can highlight the discontinuity of seismic data, is a new attribute that strengthens the characteristics of faults, can improve the accuracy of fault prediction, and enrich the details of geological structures. It is an effective means of qualitatively characterizing natural fractures. The effect of ant tracking is controlled by the ant tracking parameter settings, which mainly have 6 parameters, including initial ant boundary, tracking deviation, search step, allowed illegal step, required legal step, and threshold value in search. This step sets the ant tracking parameters and performs ant tracking calculations on the boundary detection results of the frequency division body.

[0013] Step 6, compare and select the divided-frequency ant bodies: open a certain time window up and down along the tracking target layer, extract a series of layer slices of the divided-frequency ant body and the full-frequency ant body, combine the seismic profile and the actual drilling data, comprehensively consider the fracture morphology and the characteristics of the breakpoints on the plane and the profile, as well as the degree of matching with the actual drilling data, and select the ant body with better effect as the input data for the next step of ant body fusion.

[0014] Specifically, if there are three input data in step 5, then step 6 may select three or two; if only one is input in the previous step, then only one may be selected.

[0015] Step 7, ant body fusion: Multi-attribute fusion technology is based on certain mathematical operations, while considering the factors affecting each attribute on the geological anomaly body, combining these factors, and finally obtaining the optimal result, which can effectively solve the problem of multiple solutions for single attribute prediction of geological targets. The ant body attribute fusion in this step is to perform RGB ant body fusion on the selected ant body results to obtain the final high-precision ant body.

[0016] Step 8, automatic extraction of fractures: By adjusting the strength of the ant tracking attribute and the minimum fracture length parameter, different fragment combinations are automatically extracted to obtain fracture combinations, and the fractures in the area are three-dimensionally engraved and quantitatively described. Furthermore, the bandwidth step size set in step 2 is 15Hz, and the corresponding three frequency division bodies are low frequency 0~15Hz, medium frequency 15~30Hz, and high frequency 30~45Hz. It should be noted that the low frequency contains upper and lower limits, and the medium frequency and high frequency do not contain lower limits but contain upper limits.

[0017] Furthermore, in step 5, the ant tracking adopts the Aggressive active mode.

[0018] Furthermore, in step 7, the value range of the ant body fusion is normalized to between -1 and 1.

[0019] The present invention adopts the above technical scheme, and the construction of guided filtering smoothing is helpful to reduce the signal-to-noise ratio and improve the image quality of faults and cracks. Edge detection is to improve the continuity of cracks. The frequency filtering effect and the attribute fusion of ant tracking are to detect sharper cracks and more small-scale cracks. The automatic extraction of cracks is to improve the efficiency and accuracy of crack interpretation. Compared with the traditional ant tracking workflow based on full-band seismic volume, the workflow in the present invention makes full use of the sensitivity of narrow-band seismic data to underground geological information, and can dig out sharper and more small-scale cracks; compared with other ant tracking methods of spectrum decomposition, it requires less computing time, and it is easier to distinguish the differences between three different frequency volumes, effectively avoiding human errors, and greatly saving time and energy. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments;

[0021] Figure 1 It is a flow chart of a method for detecting cracks by ant tracking based on frequency filtering frequency division body of the present invention;

[0022] Figure 2 The spectrum analysis of the target layer section of the method for detecting cracks by ant tracking based on frequency filtering frequency division body of the present invention;

[0023] Figure 3 It is a comparison of full-band and divided-frequency seismic time slices of a method for detecting cracks by ant tracking based on frequency filtering and frequency division volume of the present invention;

[0024] Figure 4 A comparison of full-band and frequency-divided seismic profiles of a method for detecting cracks by ant tracking based on frequency filtering frequency-divided volume according to the present invention;

[0025] Figure 5It is a construction-guided filtering and smoothing of a method for detecting cracks by ant tracking based on frequency filtering frequency division volume of the present invention;

[0026] Figure 6 It is the variance slices along the layers of different frequency division bodies of the method for detecting cracks by ant tracking based on frequency filtering frequency division bodies of the present invention;

[0027] Figure 7 It is a comparison of the full-band and layer-by-layer ant tracing results of a method for detecting cracks by ant tracing based on frequency filtering and frequency division of the present invention;

[0028] Figure 8 It is a comparison of the full-band and frequency-divided-body ant tracking and seismic profile superposition diagram of a method for detecting cracks based on ant tracking of frequency-filtered frequency-divided-body in the present invention;

[0029] Fig. 9 The present invention is a method for detecting cracks by ant tracking based on frequency filtering and frequency division, and the fragments extracted by opening a certain time window along the target layer are displayed in three-dimensional space. DETAILED DESCRIPTION

[0030] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0031] like Figures 1 to 9 As shown in FIG. 1 , the present invention discloses a method for detecting cracks by ant tracking based on frequency filtering frequency division bodies. Different from the traditional ant tracking method, the present invention takes into account the different sensitivities of seismic data of different frequency bands to fractures of different scales, and uses a bandpass filter to filter out frequency division bodies of different frequency band components, and then performs ant tracking calculations on the frequency division bodies of different frequency bands. The specific steps of the present invention are as follows:

[0032] Step 1, spectrum analysis: perform spectrum analysis on seismic data to obtain effective frequency bandwidth;

[0033] Step 2, frequency filtering and frequency division: using a bandpass filter to filter out low-frequency, medium-frequency and high-frequency components from the seismic data in the effective frequency band according to the set bandwidth step size to obtain three frequency division bodies;

[0034] Furthermore, the bandwidth step is set to 15Hz, and the corresponding three frequency division bodies are low frequency 0~15Hz, medium frequency 15~30Hz, and high frequency 30~45Hz. It should be noted that the low frequency includes upper and lower limits, and the medium and high frequencies do not include lower limits but include upper limits.

[0035] Step 3: Structural guidance filtering and smoothing: Structural smoothing is performed on the original post-stack seismic data volume to reduce the impact of noise and enhance the continuity of effective seismic reflections.

[0036] Step 4, boundary detection: perform variance volume calculations on the full-band, high-frequency, medium-frequency and low-frequency component frequency division bodies, detect the discontinuity points of the seismic data, and strengthen this discontinuity. A certain time window is opened up and down along the target layer to be tracked, and a series of slices along the layer are extracted. Combined with data such as seismic profiles, considering the breakpoint characteristics, the frequency division body that reflects the breakpoint characteristics better is preliminarily selected, and the next step of ant tracking calculation is entered.

[0037] Specifically, the boundary detection results of the three frequency division volumes may all be selected, or the boundary detection result of only one frequency division volume may be selected.

[0038] Step 5, ant tracking: Ant tracking technology is a biological technology based on ant algorithms. It searches for crack traces in the seismic data body according to the principle that ants use secretions to find food sources as quickly as possible, until the fault is tracked and identified. This tracking technology can highlight the discontinuity of seismic data, is a new attribute that strengthens the characteristics of faults, can improve the accuracy of fault prediction, and enrich the details of geological structures. It is an effective means of qualitatively characterizing natural fractures. The effect of ant tracking is controlled by the ant tracking parameter settings, which mainly have 6 parameters, including initial ant boundary, tracking deviation, search step, allowed illegal step, required legal step, and threshold value in search. This step sets the ant tracking parameters and performs ant tracking calculations on the boundary detection results of the frequency division body.

[0039] Step 6, compare and select the frequency-divided ant body: open a certain time window along the tracking target layer, extract a series of slices along the layer of the frequency-divided ant body and the full-frequency ant body, combine the seismic profile and the actual drilling data, comprehensively consider the characteristics of the fracture morphology and breakpoints on the plane and the profile, and the degree of matching with the actual drilling data, and select the ant body with better effect as the input data for the next step of ant body fusion. Specifically, if there are three input data in step 5, then step 6 may select three or two; if only one is input in the previous step, then only one may be selected.

[0040] Step 7, ant body fusion: Multi-attribute fusion technology is based on certain mathematical operations, while considering the factors affecting each attribute on the geological anomaly body, combining these factors, and finally obtaining the optimal result, which can effectively solve the problem of multiple solutions for single attribute prediction of geological targets. The ant body attribute fusion in this step is to perform RGB ant body fusion on the selected ant body results to obtain the final high-precision ant body. Furthermore, the value range after ant body fusion is normalized to between -1 and 1.

[0041] Step 8, automatic extraction of fractures: by adjusting the strength of the ant tracking attribute and the minimum fracture length parameter, different fragment combinations are automatically extracted to obtain fracture combinations, and three-dimensional carving is performed on the fractures in the area. The specific principle of the present invention is described in detail below:

[0042] The H well area of ​​Sichuan shale gas was selected as the research object. It is located in the southern part of the Sichuan Basin, with complex and diverse structures. It has experienced multiple tectonic movements and structural superpositions from the Paleozoic to the Cenozoic. The overlapping relationship between faults is complex, and the faults in the area are complex and diverse, with characteristics of multiple stages, multiple scales, multiple types, and multiple trends. During the drilling process of the target layer, many well collapses and well leakages were encountered, and there were problems such as poor drillability and poor well wall stability. Existing studies have shown that the development of faults may lead to problems such as well leakage and well collapse during drilling. Therefore, improving the accuracy of fracture prediction is crucial for the subsequent shale gas exploration and development.

[0043] Step 1, spectrum analysis:

[0044] Before frequency division processing, the effective frequency band of seismic data must be obtained first. Spectral analysis of post-stack seismic data in the H well area shows that the effective frequency band of seismic data in the H well area is between 5 and 45 Hz, and the main frequency is distributed around 25 Hz ( Figure 2 ).

[0045] Step 2, frequency filtering and division:

[0046] Considering the different sensitivity of geological targets of different scales to different frequency components of seismic data, the present invention uses a 15HZ step size to divide the effective frequency band into three different frequency components. They are low frequency (0~15Hz), medium frequency (15~30Hz) and high frequency (30~45Hz). Different step sizes can be selected for frequency division processing.

[0047] like Figure 3 As shown, it is a comparison of seismic time slices of full-band and divided-frequency, where a: full-band; b: 0~15Hz; c:15~30Hz; d: 30~45Hz.

[0048] Compared with full-band seismic data, the signal-to-noise ratio of low-frequency seismic profiles is greatly reduced. On the plane, only some larger faults can be reflected in the low-frequency time slice ( Figure 3 b). The seismic amplitude energy of the low-frequency component is uneven in the vertical direction, with weaker amplitudes in shallow layers and stronger amplitudes in deep layers. Figure 4 As shown in the figure, the full-band and frequency-divided seismic profiles are compared, where a: full-band; b: 0~15Hz; c: 15~30Hz; d: 30~45Hz. Although the low-frequency component can reflect some large-scale faults in deeper layers, the overall signal-to-noise ratio is much lower, and the breakpoints and shapes are not as clear and crisp as those in the full-band. Figure 4a and 4b).

[0049] The signal-to-noise ratio and imaging quality of the intermediate frequency components are greatly improved. In the plane, some fractures are enhanced in the time slice of the intermediate frequency components ( Figure 3 c). The seismic amplitude energy in deep and shallow parts is relatively consistent in the vertical direction. It can reflect most faults of different scales, which is more prominent and clearer than the faults in the full-band data, and improves the imaging quality of the breakpoints ( Figure 4 c).

[0050] For high-frequency components, some small breaks are enhanced in the time slice ( Figure 3 d). In the vertical direction, the amplitude of the shallow layer is greater than that of the deep layer. The imaging quality of shallow small-scale faults has been improved, especially in the target layer Wufeng Formation. However, the reflection ability of deep faults is limited due to the low signal-to-noise ratio ( Figure 4 d).

[0051] In general, the signal-to-noise ratio and imaging quality of mid- and high-frequency components are improved, but the opposite is true for low-frequency components.

[0052] Step 3: construct guided filter smoothing:

[0053] The structurally guided smoothing filter adopts the "anisotropic diffusion" smoothing algorithm, and its smoothing operation is only performed on the information parallel to the seismic event axis, and no smoothing is performed on the information perpendicular to the seismic event axis. The purpose of structural smoothing of the seismic data body is to increase the continuity of the event axis along the inclination and strike of the seismic reflection interface, improve the lateral resolution of the event axis termination point (fault), and effectively preserve or improve the sharpness of the fault. Taking the 15-30Hz intermediate frequency sub-frequency body as an example, by performing structurally guided filtering on the data, random noise is suppressed, the seismic event axis is more continuous, and the breakpoints are more prominent. Figure 5 a is the initial seismic profile; Figure 5 b is the seismic section after structural smoothing filtering.

[0054] Step 4, boundary detection:

[0055] After the structural guidance filtering, the resolution and clarity of the faults have been improved to a certain extent, but some faults still have the problem of unclear breakpoints. Therefore, the variance volume attribute will be used to detect the fracture discontinuity. It is a stratigraphic discontinuity detection based on probabilistic variance analysis. It calculates the variance between adjacent seismic traces to represent the difference in reflection characteristics of each seismic trace, thereby achieving the function of identifying faults. The variance volume operation is performed on the full-band, high-frequency, medium-frequency and low-frequency component frequency division bodies respectively, and a certain time window is opened up and down along the target layer of the tracking, and a series of slices along the layer are extracted. Combined with seismic profiles and other data, considering the breakpoint characteristics, the frequency division body that reflects the breakpoint characteristics better is preliminarily selected, and the next step of ant tracking calculation is entered.

[0056] like Figure 6 is the variance slice along the layer for different frequency division bodies, where Figure 6 a: full frequency band; b: 0~15Hz; c: 15~30Hz; d: 30~45Hz;

[0057] Specifically, Figure 6 b is the variance slice along the five peaks of the 0-15Hz frequency division body. Compared with the variance slice along the layer of the full-frequency body, the effective information is seriously lost, and only a very small amount of fault reflection remains;

[0058] Figure 6 c is the variance slice along the five peaks of the 15-30Hz frequency division body, which has richer information than the full frequency band, and the intersection relationship between the faults is clearer (marked by the arrows on the right side of the figure), which can better reflect the development of medium and large-scale faults; but the present invention also notes that some faults are lost in the mid-frequency component (marked by the arrows on the left side of the figure).

[0059] Figure 6 d is the variance slice of the 30-45Hz frequency crossover along the five peaks. It can be seen that the high-frequency band variance body has more information than the full-band and mid-band, and more small-scale cracks are detected. In addition, it supplements the missing breaks in the mid-frequency slice (marked by arrows).

[0060] Therefore, the boundary detection results of medium-frequency and high-frequency data are preliminarily selected as the input data for the next step of ant tracking calculation.

[0061] Step 5, ant tracking:

[0062] In this step, ant tracking is performed on the selected boundary detection results. Before ant tracking, tracking parameters need to be set. The two commonly used tracking parameter combination modes are Aggressive active mode and Passive passive mode. Among them, the active ant tracking algorithm is similar to "hard-working ants" and is better at digging faults, but because of its strong "initiative", the noise is clearer. The initial ant boundary of the passive mode is larger, and the threshold value in the search is smaller. Therefore, the passive ant tracking algorithm is a "lazy ant". It tends to track extremely strong signals and abandon weaker signals, which helps to suppress noise and reflect the trend of large faults, but it is easy to cause local unclear low-level small faults to be discontinuous, thereby destroying their continuity. This invention recommends using the active tracking mode to perform ant tracking on each frequency division body. In order to dig more small-scale faults, three ant tracking bodies of low, medium and high frequency components are obtained respectively. Parameter setting is optional. Select the Aggressive active mode, the Passive passive mode or the self-selected ant tracking parameter combination. This invention recommends selecting the Aggressive active mode.

[0063] Step 6: Compare and optimize ant tracking results:

[0064] In this step, the traditional full-band ant tracking results and the ant tracking results of the frequency-divided body initially optimized in the previous step are compared and optimized as the input data for the next step of ant body fusion.

[0065] Figure 7 a: full frequency band; b: 15~30Hz; c: 30~45Hz; d: fusion result of 15~30Hz and 30~45Hz;

[0066] Figure 8 a: full frequency band; b: 15~30Hz; c: 30~45Hz; d: fusion result of 15~30Hz and 30~45Hz.

[0067] Specifically, Figure 7 a is a full-band crack prediction slice along Wufeng, where northwest, east-west, and northeast-oriented faults are mainly developed, and a small number of near north-south faults are developed.

[0068] Figure 7 b is the 15-30HZ crack prediction slice along Wufeng. Figure 8 b is the superposition of the 15-30 Hz ant tracking results and the seismic profile. It can be seen that there are not only many more cracks in the intermediate frequency component (taking the fracture at the arrow as an example), but also the fractures are more continuous and the intersection relationship is clearer. However, it should be noted that due to the filtering effect, the intermediate frequency component also loses some important fractures (taking the fracture at the arrow as an example).

[0069] Figure 7 c is the 30-45HZ crack prediction slice along Wufeng. Figure 8 c is the superposition of the ant tracking results and seismic profiles at 30-45 Hz. Compared with the previous two, there are more faults, especially small-scale faults, and they also effectively supplement the faults that are lost in the medium-frequency component results (take the fault at the arrow as an example).

[0070] In summary, compared with the full-band data, the fractures detected by the mid-frequency component are more prominent and continuous, and the high-frequency component is more suitable for detecting smaller-scale fractures. The two complement each other. Therefore, the mid-frequency component and the high-frequency component are finally selected as the input data for ant body fusion.

[0071] Step 7, ant body fusion: fuse the selected ant body results and normalize the value range to between -1 and 1. The fusion result takes into account the advantages of both medium and high frequency components of crack detection results, making the crack prediction results sharper, more continuous and richer in information ( Figure 7 d and Figure 8 d). Any one, two or three of the three frequency-division components can be selected as the input of ant body fusion.

[0072] Step 8, automatic extraction of crack slices:

[0073] Specifically, after integrating the ant tracking attributes, different fragment combinations can be automatically extracted by adjusting the intensity of the ant tracking attributes, the minimum fracture length and other parameters, and the fractures in the area can be three-dimensionally carved and quantitatively described. This time, a 15 millisecond time window was opened above and below the bottom of the Wufeng Formation of the target layer to extract the fragment combination with an ant body intensity greater than -0.9 and a minimum fracture length of 500 meters, in order to extract the main fractures developed in the area. Fig. 9 It is a three-dimensional display of the automatically extracted fragments in three-dimensional space. Parameters such as ant body strength and minimum fracture length can be selected to automatically extract different fragment combinations.

[0074] If the interpretation of fractures is done manually, it would take weeks or longer. However, by applying the automated extraction method, the extraction of the desired fractures can be completed in 4-5 hours. Therefore, the present invention can confidently say that this method can significantly reduce the time compared to traditional methods.

[0075] The present invention adopts the above technical scheme, and the construction of guided filtering smoothing is helpful to reduce the signal-to-noise ratio and improve the image quality of faults and cracks. Edge detection is to improve the continuity of cracks. The frequency filtering effect and the attribute fusion of ant tracking are to detect sharper cracks and more small-scale cracks. The automatic extraction of cracks is to improve the efficiency and accuracy of crack interpretation. Compared with the traditional ant tracking workflow based on full-band seismic volume, the workflow in the present invention makes full use of the sensitivity of narrow-band seismic data to underground geological information, and can dig out sharper and more small-scale cracks; compared with other ant tracking methods of spectrum decomposition, it requires less computing time, and it is easier to distinguish the differences between three different frequency volumes, effectively avoiding human errors, and greatly saving time and energy.

[0076] Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. In the absence of conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the detailed description of the embodiments of the present application is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians of the art without making creative work are within the scope of protection of the present application.

Claims

1. A method for detecting cracks by ant tracking based on frequency filtering and frequency division, characterized in that: The steps are as follows: Step 1, spectrum analysis: perform spectrum analysis on seismic data to obtain effective frequency bandwidth; Step 2, frequency filtering and frequency division: using a bandpass filter to filter out low-frequency, medium-frequency and high-frequency components from the seismic data in the effective frequency band according to the set bandwidth step size to obtain three frequency division bodies; Step 3, constructing guided filter smoothing: The "anisotropic diffusion" smoothing algorithm is used to construct guided filter smoothing for seismic data. The smoothing operation is only performed on the information parallel to the seismic event axis, and no smoothing is performed on the information perpendicular to the seismic event axis. Along the inclination and strike of the seismic reflection interface, effective filtering methods are used for denoising, increasing the continuity of the event axis, improving the lateral resolution of the event axis fault, and effectively preserving or improving the sharpness of the fault. Step 4, boundary detection: perform variance volume calculation on the high-frequency, medium-frequency and low-frequency component frequency division bodies respectively, and open a certain time window up and down along the target layer to extract a series of slices along the layer. Combined with the seismic profile data and considering the breakpoint characteristics, preliminarily select at least one frequency division body that reflects the breakpoint characteristics better, and enter the next step of ant tracking calculation; Step 5, ant tracking: using the existing mature ant tracking technology to find crack traces in the boundary detection results of the frequency division body until the fault tracking and identification are completed; Step 6, compare and select the frequency-divided ant body: open a certain time window up and down along the tracked target layer, extract a series of slices along the layer of the frequency-divided ant body, combine the seismic profile and the actual drilling data, comprehensively consider the characteristics of the fracture morphology and breakpoints on the plane and the profile, and the matching degree with the actual drilling data, and select at least one ant body with better effect as the input data for the next step of ant body fusion; Step 7, ant body fusion: fuse the selected ant body results to obtain the final high-precision ant body; Step 8, automatic extraction of fractures: By adjusting the strength of the ant tracking attribute and the minimum fracture length parameter, different fragment combinations are automatically extracted to obtain fracture combinations, and three-dimensional carving is performed on the fractures in the area.

2. The method for detecting cracks by ant tracking based on frequency filtering and frequency division according to claim 1, characterized in that: The bandwidth step size set in step 2 is 15Hz, and the corresponding three crossovers are low frequency 0~15Hz, medium frequency 15~30Hz, and high frequency 30~45Hz.

3. The method for detecting cracks by ant tracking based on frequency filtering and frequency division according to claim 1, characterized in that: The boundary detection results of the three frequency division bodies are selected to enter the next step of ant tracking calculation.

4. The method for detecting cracks by ant tracking based on frequency filtering and frequency division according to claim 3, characterized in that: Three ant bodies are selected as input data for the next step of ant body fusion.

5. The method for detecting cracks by ant tracking based on frequency filtering and frequency division according to claim 1, characterized in that: In step 5, ant tracking adopts the Aggressive active mode.

6. The method for detecting cracks by ant tracking based on frequency filtering and frequency division according to claim 1, characterized in that: In step 6, the extraction is performed by using a time window method relying on existing mature software, and the selected time window does not exceed ±20 milliseconds.

7. The method for detecting cracks by ant tracking based on frequency filtering and frequency division according to claim 6, characterized in that: The time window selected in step 6 is ±10 milliseconds.

8. The method for detecting cracks by ant tracking based on frequency filtering and frequency division according to claim 1, characterized in that: In step 7, the selected ant body results are RGB fused to obtain the final high-precision ant body.

9. The method for detecting cracks by ant tracking based on frequency filtering and frequency division according to claim 1, characterized in that: In step 7, the value range of the ant body fusion is normalized to between -1 and 1.

Images