Five-in-one oil and uranium exploration method, device and electronic equipment

Abstract

The application discloses a five-in-one oil and uranium exploration and ore prospecting technology, realizes secondary development and utilization of a large amount of oil well logging data of an oil field through comprehensive radioactive anomaly screening, screens reliable radioactive anomalies, and determines a uranium ore prospecting target area; when shallow natural gamma data of the oil field is missing and cannot be screened, old well re-measuring technology is adopted for supplement; through natural gamma and quantitative gamma conversion technology, a correlation between gamma anomalies of a uranium mineralization section verification hole and GR anomalies of an oil field hole is established, and uranium ore resource quantity is estimated; through three-dimensional seismic data application technology, according to uranium exploration geological requirements, application countermeasures and methods of sandstone type uranium mine three-dimensional seismic data fault interpretation technology are established; through typical deposit application research, favorable conditions of typical sandstone type uranium mineralization in main uranium-producing basins in northern China are analyzed, and a sandstone type uranium exploration and evaluation technology system of an oil field working area of the basin is established, so that a concave is accurately evaluated, a favorable zone is determined, and exploration deployment is guided.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of geophysical technology, and particularly relates to a five-in-one oil and uranium exploration method and device and electronic equipment. BACKGROUND

[0002] In the current mineral resource exploration, oil and gas and uranium are important energy and strategic resources, and their exploration and development are of great significance to national energy security and economic development. In recent years, a number of large and super large sandstone type uranium deposits have been discovered in the Mesozoic basins in northern China, and remarkable results have been achieved. The coexistence, interaction and exploration of sandstone type uranium deposits and oil and gas in the same basin have been widely recognized. According to the exploration idea of oil and uranium exploration, the oil and gas drilling well logging data are reinterpreted and redeveloped, the abnormal drilling holes are screened, the key exploration target areas are delineated and drilling verification is carried out, and then the uranium mineralization area is found. The key of this idea is to determine the radioactivity anomaly index of oil and gas drilling hole, to check the radioactivity anomaly characteristics of drilling hole (potential uranium ore hole, potential uranium mineralization hole, normal hole, and no parameter hole) and to carry out drilling verification. However, in actual work, the following problems are encountered:

[0003] (1) Through the collection and research of drilling data of domestic major oil fields, the developed oil fields actually retain relatively comprehensive and complete data. Through the database construction technology mastered, the drilling logging data can be exported as data files. The development personnel and technical personnel with professional background often need to spend a lot of time and energy to query and process data in daily work, and the screening of abnormal data often needs to be completed manually. This way has obvious defects, such as slow query speed, time and energy consumption, low efficiency and inability to meet the requirements of efficient office.

[0004] (2) Although most of the oil field drilling data are complete and complete, the oil reservoirs explored and developed by oil and gas fields are often deep, and the sandstone type uranium exploration target layer is shallow, so the logging data are often missing in the shallow part, resulting in the loss of radioactivity information in the shallow part, which cannot determine the ore-bearing property of the stratum, restricts the control of regional concealed ore-bearing geological body, and delays the progress of exploration.

[0005] (3) The radioactivity information of oil field logging data is in units of natural gamma (API), but it cannot effectively determine the radioactivity intensity, and cannot be used as the basis for determining the potential ore-bearing property of drilling hole, and cannot be compared with the logging information of known nuclear industry drilling hole, lacking of uniformity.

[0006] (4) As a relatively mature geophysical prospecting method, seismic exploration technology has been widely used in resource exploration of oil and gas, but it is less used in the exploration of sandstone-type uranium deposits. In addition, due to the complexity and diversity of geological structures, the results of fault identification in the current exploration of sandstone-type uranium deposits based on seismic data often depend on the experience and judgment of interpreters, and there is a certain subjectivity. The algorithm is often difficult to accurately capture the characteristics of all faults.

[0007] (5) A large number of sandstone-type uranium deposits have been discovered in the Ordos-Eren Basin, but the ore-controlling factors, enrichment regularity, and metallogenic model of each deposit are not perfect, and the understanding is not systematic and comprehensive. It is urgent to further sort out and improve the understanding.

[0008] That is, the traditional single exploration method often cannot meet the exploration needs of oil and gas and uranium deposits at the same time, and there are problems such as long exploration period, high cost, and low efficiency. Especially in the complex geological environment where sandstone-type uranium deposits and oil and gas resources coexist, single exploration technology often cannot accurately identify the distribution of ore bodies, leading to inaccurate resource assessment and affecting subsequent development decisions. SUMMARY

[0009] Therefore, the purpose of the present application is to provide a five-in-one oil and uranium exploration method, device and electronic equipment based on the five-in-one oil and uranium exploration method, device and electronic equipment to solve the problems raised in the above background art. Specifically, the present application provides a five-in-one oil and uranium exploration method, device and electronic equipment based on the five-in-one oil and uranium exploration method, device and electronic equipment based on the five-in-one oil and uranium exploration method, device and electronic equipment to solve the problems raised in the above background art.

[0010] Moreover, another purpose of the present application is to provide a deep learning-based fault identification model applied in the seismic exploration of sandstone-type uranium deposits, which significantly improves the accuracy of fault identification in sandstone-type uranium deposits to solve the problems raised in the above background art.

[0011] Moreover, another purpose of the present application is to provide a loss function applied to a deep learning-based fault identification model in the seismic exploration of sandstone-type uranium deposits, which significantly improves the convergence speed of the deep learning-based fault identification model to solve the problems raised in the above background art.

[0012] In order to achieve the above purpose, the present application adopts the following technical solutions:

[0013] In a first aspect, a five-in-one oil and uranium exploration method based on the five-in-one oil and uranium exploration method is provided, comprising:

[0014] Step S1: Conducting data screening, wherein step S1 comprises:

[0015] Step S11: Collecting, investigating and organizing drilling information of oil fields;

[0016] Step S12: Analyze the standards for borehole radioactivity anomalies and construct radioactivity anomaly screening software;

[0017] Step S13: Based on the radioactive anomaly screening software, judge according to the borehole radioactive anomaly standard, and screen out radioactive anomaly wells;

[0018] Step S2: Conduct re-survey of the old well, wherein step S2 includes:

[0019] Step S21: Establish the casing-cement ring model;

[0020] Step S22: Conduct a logging response test using cement ring casing;

[0021] Step S23: Analyze the functional relationship between the count rate and the cement ring property parameters, and determine the correction coefficient;

[0022] Step S24: Obtain accurate shallow strata radioactivity information based on the correction coefficient;

[0023] Step S3: Conduct drilling operations to verify and achieve quantitative gamma conversion from natural gamma. Step S3 includes:

[0024] Step S31: Collect gamma data from verification wells and oilfield boreholes;

[0025] Step S32: Correct the verification hole gamma logging depth to the range of abnormal depths in the oilfield borehole gamma.

[0026] Step S33: Establish the regression equation between natural gamma and quantitative gamma;

[0027] Step S34: Use the regression equation to convert between natural gamma and quantitative gamma, and evaluate mineral content;

[0028] Step S4: Conduct 3D seismic data processing, interpretation, and analysis of faults, sand bodies, and structures in the mining area;

[0029] Step S5: Conduct a comparison of typical mineral deposits. Step S5 includes:

[0030] Step S51: Collect data on typical uranium deposits;

[0031] Step S52: Analyze the structural sedimentary evolution characteristics, uranium source conditions, sand bodies and sedimentary systems, ore body characteristics, and mineralization regularity;

[0032] Step S53: Summarize the ore-controlling factors of typical sandstone-type uranium deposits;

[0033] Step S6: Based on the results of steps S3 to S5 above, determine the favorable uranium ore zones and evaluate the mineralization potential of the working area.

[0034] Furthermore, step S4 further includes:

[0035] S41: Determine the tectonic characteristics of uranium-bearing areas using seismic fine tectonic interpretation techniques and tectonic and sedimentary evolution analysis techniques;

[0036] S42: Using seismic staging and grading fault fine characterization technology to determine uranium ore faults, oil and gas faults, and oil and gas migration channels;

[0037] S43: Determine the condition characteristics of the top and bottom plates of uranium ore sand bodies using seismic identification technology for the top and bottom plates of sand bodies;

[0038] S44: Determine the distribution, thickness, and physical properties of uranium sandstone using fine-grained stratigraphic and reservoir characterization techniques.

[0039] Furthermore, step S42, which utilizes seismic staging and grading fault fine characterization technology to determine uranium ore faults, oil and gas faults, and oil and gas migration pathways, includes:

[0040] A seismic deep learning model is used for fine characterization of uranium ore fractures. The seismic deep learning model includes an encoder-decoder module, which comprises:

[0041] Encoder unit, fourth Swing Transformer unit, decoder unit, jump connection unit;

[0042] The encoder unit includes: an image block segmentation unit, an image block embedding unit, a first Swing Transformer unit, a first image block merging unit, a second Swing Transformer unit, a second image block merging unit, a third Swing Transformer unit, and a third image block merging unit. The image block segmentation unit outputs to the image block embedding unit, the image block embedding unit outputs to the first Swing Transformer unit, the first Swing Transformer unit outputs to the first image block merging unit, the first image block merging unit outputs to the second Swing Transformer unit, the second Swing Transformer unit outputs to the second image block merging unit, the second image block merging unit outputs to the third Swing Transformer unit, and the third Swing Transformer unit outputs to the third image block merging unit.

[0043] The third image block merging unit outputs to the fourth Swing Transformer unit, and the fourth Swing Transformer unit outputs to the first image block expansion unit;

[0044] The decoder unit includes: a first image block expansion unit, a fifth Swing Transformer unit, a second image block expansion unit, a sixth Swing Transformer unit, a third image block expansion unit, a seventh Swing Transformer unit, a final image expansion module, and a linear projection module. The first image block expansion unit outputs to the fifth Swing Transformer unit, the fifth Swing Transformer unit outputs to the second image block expansion unit, the second image block expansion unit outputs to the sixth Swing Transformer unit, the sixth Swing Transformer unit outputs to the third image block expansion unit, the third image block expansion unit outputs to the seventh Swing Transformer unit, the seventh Swing Transformer unit outputs to the final image expansion module, and the final image expansion module outputs to the linear projection module.

[0045] The encoder unit is connected to the decoder unit via a jumper connection unit.

[0046] Furthermore, the encoder unit is connected to the decoder unit via a jump connection unit, including:

[0047] The outputs of the first Swing Transformer unit, the second Swing Transformer unit, and the third Swing Transformer unit are concatenated and then input into the first hop connection unit. The output of the first hop connection unit is then concatenated with the output of the first image block expansion unit and input into the fifth Swing Transformer unit.

[0048] The outputs of the first, second, and third Swing Transformer units are concatenated and then input into the second hop connection unit. The output of the second hop connection unit is then concatenated with the output of the second image block expansion unit and input into the sixth Swing Transformer unit.

[0049] The outputs of the first, second, and third Swing Transformer units are concatenated and then input into the third hop connection unit. The output of the third hop connection unit is then concatenated with the output of the second image block expansion unit and input into the seventh Swing Transformer unit.

[0050] Furthermore, the jump connection unit includes: a first jump connection unit, a second jump connection unit, and a third jump connection unit, and the first jump connection unit, the second jump connection unit, and the third jump connection unit have the same structure;

[0051] The first hop connection unit includes: the eighth Swing Transformer unit and a channel mixing module;

[0052] The channel mixing module includes two branches. The first branch includes a max pooling layer and a 1×1 convolutional layer, and the second branch includes an average pooling layer and a 1×1 convolutional layer. The outputs of the first branch and the second branch are added together and then passed to the 1×1 convolutional layer and the sigmoid activation layer to obtain the output of the channel mixing module.

[0053] The output of the eighth Swing Transformer unit and the output of the channel mixing module are multiplied together to obtain the output of the first hop connection unit.

[0054] Furthermore, deep learning models include two encoder-decoder modules.

[0055] The result of the last Swing Transformer unit in the decoder unit of the first encoder-decoder module is output to the first Swing Transformer unit in the encoder unit of the second encoder-decoder module. A jump connection unit is used to establish a jump connection between the Swing Transformer unit in the decoder unit of the first encoder-decoder module and the Swing Transformer unit in the encoder unit of the second encoder-decoder module. The specific structure is as follows:

[0056] The earthquake deep learning model includes:

[0057] First encoder-decoder module, second encoder-decoder module;

[0058] The first encoder-decoder module includes:

[0059] First encoder unit, fourth Swing Transformer unit, first decoder unit, jump connection unit;

[0060] The first encoder unit includes: an image block segmentation unit, an image block embedding unit, a first Swing Transformer unit, a first image block merging unit, a second Swing Transformer unit, a second image block merging unit, a third Swing Transformer unit, and a third image block merging unit. The image block segmentation unit outputs to the image block embedding unit, the image block embedding unit outputs to the first Swing Transformer unit, the first Swing Transformer unit outputs to the first image block merging unit, the first image block merging unit outputs to the second Swing Transformer unit, and the second Swing Transformer unit outputs to the third Swing Transformer unit.

[0061] The Transformer unit outputs to the second image block merging unit, and the second image block merging unit outputs to the third Swin.

[0062] The Transformer unit and the third Swing Transformer unit output to the third image block merging unit;

[0063] The third image block merging unit outputs to the fourth Swing Transformer unit, and the fourth Swing Transformer unit outputs to the first image block expansion unit;

[0064] The first decoder unit includes: a first image block expansion unit, a fifth Swing Transformer unit, a second image block expansion unit, a sixth Swing Transformer unit, a third image block expansion unit, a seventh Swing Transformer unit, a final image expansion module, and a linear projection module. The first image block expansion unit outputs to the fifth Swing Transformer unit, the fifth Swing Transformer unit outputs to the second image block expansion unit, the second image block expansion unit outputs to the sixth Swing Transformer unit, the sixth Swing Transformer unit outputs to the third image block expansion unit, the third image block expansion unit outputs to the seventh Swing Transformer unit, the seventh Swing Transformer unit outputs to the final image expansion module, and the final image expansion module outputs to the linear projection module.

[0065] Furthermore, the output of the seventh Swing Transformer unit is sent to the eighth Swing Transformer unit;

[0066] The second encoder-decoder module includes:

[0067] Second encoder unit, eleventh Swing Transformer unit, second decoder unit, jump connection unit;

[0068] The second encoder unit includes: an eighth Swing Transformer unit, a fourth image block merging unit, a ninth Swing Transformer unit, a fifth image block merging unit, a tenth Swing Transformer unit, and a sixth image block merging unit. The eighth Swing Transformer unit outputs to the fourth image block merging unit, the fourth image block merging unit outputs to the ninth Swing Transformer unit, the ninth Swing Transformer unit outputs to the fifth image block merging unit, the fifth image block merging unit outputs to the tenth Swing Transformer unit, and the tenth Swing Transformer unit outputs to the sixth image block merging unit.

[0069] Furthermore, the output of the sixth image block merging unit is sent to the eleventh Swing Transformer unit, and the output of the eleventh Swing Transformer unit is sent to the fourth image block expansion unit;

[0070] The second decoder unit includes: a fourth image block expansion unit, a twelfth Swing Transformer unit, a fifth image block expansion unit, a thirteenth Swing Transformer unit, a sixth image block expansion unit, a fourteenth Swing Transformer unit, a final image expansion module, and a linear projection module. The output of the fourth image block expansion unit is sent to the twelfth Swing Transformer unit, the output of the twelfth Swing Transformer unit is sent to the fifth image block expansion unit, the output of the fifth image block expansion unit is sent to the thirteenth Swing Transformer unit, the output of the thirteenth Swing Transformer unit is sent to the sixth image block expansion unit, the output of the sixth image block expansion unit is sent to the fourteenth Swing Transformer unit, the output of the fourteenth Swing Transformer unit is sent to the final image expansion module, and the output of the final image expansion module is sent to the linear projection module.

[0071] The first encoder unit is connected to the first decoder unit via a jump connection unit;

[0072] The second encoder unit is connected to the second decoder unit via a jump connection unit;

[0073] The first decoder unit is connected to the first encoder unit via a jump connection unit;

[0074] Furthermore, the first encoder unit is connected to the first decoder unit via a jump connection unit, including:

[0075] The outputs of the first Swing Transformer unit, the second Swing Transformer unit, and the third Swing Transformer unit are concatenated and then input into the first hop connection unit. The output of the first hop connection unit is then concatenated with the output of the first image block expansion unit and input into the fifth Swing Transformer unit.

[0076] The outputs of the first, second, and third Swing Transformer units are concatenated and then input into the second hop connection unit. The output of the second hop connection unit is then concatenated with the output of the second image block expansion unit and input into the sixth Swing Transformer unit.

[0077] The outputs of the first, second, and third Swing Transformer units are concatenated and then input into the third hop connection unit. The output of the third hop connection unit is then concatenated with the output of the second image block expansion unit and input into the seventh Swing Transformer unit.

[0078] Furthermore, the jump connection unit includes: a first jump connection unit, a second jump connection unit, and a third jump connection unit, and the first jump connection unit, the second jump connection unit, and the third jump connection unit have the same structure;

[0079] Furthermore, the second encoder unit is connected to the second decoder unit via a jump connection unit, including:

[0080] The outputs of the eighth, ninth, and tenth Swing Transformer units are concatenated and then input into the fourth hop connection unit. The output of the fourth hop connection unit is then concatenated with the output of the sixth image block expansion unit and input into the fourteenth Swing Transformer unit.

[0081] The outputs of the eighth, ninth, and tenth Swing Transformer units are concatenated and then input into the fifth hop connection unit. The output of the fifth hop connection unit is then concatenated with the output of the fifth image block expansion unit and input into the thirteenth Swing Transformer unit.

[0082] The outputs of the eighth, ninth, and tenth Swing Transformer units are concatenated and then input into the fifth hop connection unit. The output of the sixth hop connection unit is then concatenated with the output of the fourth image block expansion unit and input into the twelfth Swing Transformer unit.

[0083] Furthermore, the first decoder unit is connected to the first encoder unit via a jump connection unit, including:

[0084] The outputs of the fifth, sixth, and seventh Swing Transformer units are concatenated and then input into the seventh hop connection unit. The output of the seventh hop connection unit is then concatenated with the output of the fourth image block merging unit and input into the ninth Swing Transformer unit.

[0085] The outputs of the fifth, sixth, and seventh Swing Transformer units are concatenated and then input into the eighth hop connection unit. The output of the eighth hop connection unit is then concatenated with the output of the fifth image block merging unit and input into the tenth Swing Transformer unit.

[0086] Furthermore, the structures of the first, second, third, fourth, fifth, sixth, seventh, and eighth jump connection units are all identical.

[0087] Furthermore, the hybrid loss function used in the earthquake deep learning model is:

[0088]

[0089]

[0090] L = ω1L1 + ω2L2, and we have: ω1 + ω2 = 1

[0091] Among them, y i =sigmoid(xi ), where xi is the input to the deep learning network, and y i ' represents the pixel label, α represents the ratio of non-fault pixels to the total number of pixels; N represents the total number of pixels in the seismic image, and i represents the pixel index. ω1 and ω2 are the weighting coefficients, respectively.

[0092] Furthermore, the weight coefficients ω1 and ω2 in the hybrid loss function are dynamically obtained as they change over time, and their calculation method is as follows: in:

[0093]

[0094] L t =ω 1,t-1 L 1,t +ω 2,t-1 L 2,t

[0095] Among them, L j,t Let u(L) be the j-th loss function at time t. j,t-1 () represents the average value of the j-th type of loss before time t-1.

[0096] σ(r i,t ) and μ(r i,t ) represent r before time t respectively j,t The standard deviation and mean of j, and j takes the value of 1 or 2.

[0097] Secondly, a five-in-one oil and uranium exploration and prospecting device is provided, including:

[0098] Data screening unit: Conduct data screening, which includes:

[0099] First screening module: collecting, surveying, and organizing borehole information from the oilfield;

[0100] The second screening module analyzes the standards for abnormal radioactivity in boreholes and constructs software for screening abnormal radioactivity.

[0101] The third screening module: Based on the radioactive anomaly screening software, it identifies wells with radioactive anomalies according to the borehole radioactive anomaly standards;

[0102] Old Well Re-survey Unit: Conducting old well re-surveys, which includes:

[0103] First retesting module: Establishing the sleeve-cement ring model;

[0104] Second retest module: Conducting logging response tests on cement ring casing;

[0105] The third retest module analyzes the functional relationship between the count rate and the cement ring property parameters to determine the correction coefficient.

[0106] Fourth retesting module: Obtain accurate shallow strata radioactivity information based on the correction coefficient;

[0107] Gamma conversion unit: Conduct drilling operations to verify and achieve quantitative gamma conversion from natural gamma. The gamma conversion unit includes:

[0108] First Gamma Module: Collects gamma data from verification boreholes and oilfield boreholes;

[0109] Second Gamma Module: Corrects the verification hole gamma logging depth to the range of abnormal depths in oilfield borehole gamma;

[0110] The third gamma module: Establishing the regression equation between natural gamma and quantitative gamma;

[0111] The fourth gamma module: uses regression equations to convert between natural gamma and quantitative gamma, and evaluates mineral content;

[0112] Seismic processing and interpretation unit: Conducts 3D seismic data processing, interpretation, and analysis of faults, sand bodies, and structures in the mining area;

[0113] Ore deposit comparison unit: Conduct comparisons of typical ore deposits, including:

[0114] First comparison module: Collecting data on typical uranium deposits;

[0115] The second comparative module analyzes the structural sedimentary evolution characteristics, uranium source conditions, sand bodies and sedimentary systems, ore body characteristics, and metallogenic regularities.

[0116] The third comparative module summarizes the ore-controlling factors of typical sandstone-type uranium deposits;

[0117] Comprehensive processing unit: Based on the results of the data screening unit, the old well re-survey unit, the gamma conversion unit, the seismic processing and interpretation unit, and the deposit comparison unit, determine the favorable uranium ore zones and evaluate the mineralization potential of the working area.

[0118] Thirdly, an electronic device is also provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of a five-in-one oil and uranium exploration method.

[0119] Fourthly, a computer storage medium is also provided, on which a computer program is stored, characterized in that the computer program, when executed by a processor, implements the steps of a five-in-one oil and uranium exploration method.

[0120] Compared with the prior art, the beneficial effects of the present invention are:

[0121] 1. This invention utilizes a five-in-one oil and uranium exploration technology. Through comprehensive radiometric anomaly screening, it enables the secondary development and utilization of a large amount of oilfield logging data, identifying reliable radiometric anomalies and determining uranium exploration target areas. When shallow natural gamma data is lacking and cannot be screened, old well re-monitoring technology is employed for supplementation. By using natural gamma and quantitative gamma conversion technology, the correlation between γ anomalies in verification wells of uranium mineralization zones and GR anomalies in oilfield wells is established to assist in estimating uranium resources. Through the application of 3D seismic data technology, based on the geological requirements of uranium exploration, strategies and methods for applying 3D seismic data fault interpretation technology to sandstone-type uranium deposits are established. Through research on typical deposits, the favorable conditions for sandstone-type uranium mineralization in major uranium-producing basins in northern my country are analyzed, establishing a sandstone-type uranium exploration and evaluation technology system for basin oilfield working areas, accurately evaluating and identifying depressions, determining favorable zones, and guiding exploration deployment.

[0122] 2. Based on the actual situation of uranium seismic exploration, this invention has developed a new deep learning neural network model for fault interpretation technology of three-dimensional seismic data. This model combines the upper and lower information to improve the resolution capability. It can not only predict the main faults, but also better avoid most non-fault discontinuities, thus improving the resolution capability of sandstone-type uranium ore faults.

[0123] 3. This invention develops a new hybrid loss function by combining a deep learning neural network model for uranium mine seismic fault identification. This hybrid loss function can accelerate the model convergence speed, improve the model's convergence adaptability, and improve the accuracy of fault identification. Attached Figure Description

[0124] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0125] Figure 1 A flowchart illustrating the implementation of the five-in-one oil and uranium exploration method according to an embodiment of the present invention;

[0126] Figure 2 This is a stratigraphic structural feature map of the depth domain of a sandstone-type uranium deposit provided according to an embodiment of the present invention;

[0127] Figure 3 A seismic profile of a sandstone-type uranium deposit provided according to an embodiment of the present invention;

[0128] Figure 4 A deep learning neural network model structure (single encoder-decoder) for fault / fracture prediction based on seismic data for sandstone-type uranium deposits provided in an embodiment of the present invention;

[0129] Figure 5 The skip connection unit structure in the deep learning neural network model for fault / fracture prediction based on seismic data for sandstone-type uranium deposits provided in the embodiments of the present invention;

[0130] Figure 6 A deep learning neural network model structure (two encoders-decoders) for fault / fracture prediction based on seismic data for sandstone-type uranium deposits provided in an embodiment of the present invention.

[0131] Figure 7 The structure of the Swin transformer module in the deep learning neural network model for fault / fracture prediction based on seismic data for sandstone-type uranium deposits provided in an embodiment of the present invention;

[0132] Figure 8 The 2D prediction results (vertical profile) of a deep learning neural network model for fault / fracture prediction based on seismic data for sandstone-type uranium deposits provided in an embodiment of the present invention.

[0133] Figure 9 The 2D prediction result (lateral profile) of a deep learning neural network model for fault / fracture prediction based on seismic data for sandstone-type uranium deposits provided in an embodiment of the present invention.

[0134] Figure 10 The 3D prediction results of a deep learning neural network model for fault / fracture prediction based on seismic data for sandstone-type uranium deposits provided in an embodiment of the present invention;

[0135] Figure 11 The graph shows the weight coefficients of the loss function of a deep learning neural network model for fault / fracture prediction based on seismic data for sandstone-type uranium deposits provided in an embodiment of the present invention, as a function of time.

[0136] Figure 12 A diagram of a five-in-one oil and uranium exploration device provided according to an embodiment of the present invention. Detailed Implementation

[0137] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0138] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0139] Figure 1 The diagram illustrates a five-in-one oil and uranium exploration method according to an embodiment of the present invention. For ease of description, only the parts relevant to the embodiments of the present invention are shown, and are detailed below:

[0140] A five-in-one method for oil and uranium exploration includes:

[0141] Step S1: Conduct data screening, wherein step S1 includes:

[0142] Step S11: Collect, investigate, and organize borehole information from the oilfield;

[0143] Step S12: Analyze the standards for borehole radioactivity anomalies and construct radioactivity anomaly screening software;

[0144] Step S13: Based on the radioactive anomaly screening software, judge according to the borehole radioactive anomaly standard, and screen out radioactive anomaly wells;

[0145] Step S2: Conduct re-survey of the old well, wherein step S2 includes:

[0146] Step S21: Establish the casing-cement ring model;

[0147] Step S22: Conduct a logging response test using cement ring casing;

[0148] Step S23: Analyze the functional relationship between the count rate and the cement ring property parameters, and determine the correction coefficient;

[0149] Step S24: Obtain accurate shallow strata radioactivity information based on the correction coefficient;

[0150] Step S3: Conduct drilling operations to verify and achieve quantitative gamma conversion from natural gamma. Step S3 includes:

[0151] Step S31: Collect gamma data from verification wells and oilfield boreholes;

[0152] Step S32: Correct the verification hole gamma logging depth to the range of abnormal depths in the oilfield borehole gamma.

[0153] Step S33: Establish the regression equation between natural gamma and quantitative gamma;

[0154] Step S34: Use the regression equation to convert between natural gamma and quantitative gamma, and evaluate mineral content;

[0155] Step S4: Conduct 3D seismic data processing, interpretation, and analysis of faults, sand bodies, and structures in the mining area;

[0156] Step S5: Conduct a comparison of typical mineral deposits. Step S5 includes:

[0157] Step S51: Collect data on typical uranium deposits;

[0158] Step S52: Analyze the structural sedimentary evolution characteristics, uranium source conditions, sand bodies and sedimentary systems, ore body characteristics, and mineralization regularity;

[0159] Step S53: Summarize the ore-controlling factors of typical sandstone-type uranium deposits;

[0160] Step S6: Based on the results of steps S3 to S5 above, determine the favorable uranium ore zones and evaluate the mineralization potential of the working area.

[0161] Furthermore, step S4 further includes:

[0162] S41: Determine the tectonic characteristics of uranium-bearing areas using seismic fine tectonic interpretation techniques and tectonic and sedimentary evolution analysis techniques;

[0163] S42: Using seismic staging and grading fault fine characterization technology to determine uranium ore faults, oil and gas faults, and oil and gas migration channels;

[0164] S43: Determine the condition characteristics of the top and bottom plates of uranium ore sand bodies using seismic identification technology for the top and bottom plates of sand bodies;

[0165] S44: Determine the distribution, thickness, and physical properties of uranium sandstone using fine-grained stratigraphic and reservoir characterization techniques.

[0166] Specifically, in step S1, data screening technology is an effective means of obtaining information on deep uranium mineralization. Based on liberating productive forces, it enables further research into anomalous data, forming a "technical system for screening and evaluating radioactive anomalies in oil and gas wells." This system allows for rapid, accurate, and efficient screening and processing of logging data, quickly acquiring anomalous information, identifying favorable exploration areas, and conducting uranium exploration through optimized radioactive monitoring areas and subsequent drilling verification. This information is then submitted to uranium development bases, ultimately leading to the development of sandstone-type uranium deposits.

[0167] For shallow oil wells with natural gamma-ray logging curves in various depressions within the oilfield, anomaly screening is based on the screening practice standards in the table below, using radioactive anomaly screening software as a foundation. Screening for anomaly sections at depths exceeding 500m is most advantageous. Based on this, wells with anomalies greater than 1000 API and a thickness generally greater than 1m are selected as extremely high anomaly wells. Experience suggests that the identified extremely high anomaly wells generally reach the level of industrial uranium ore wells.

[0168] Table 1. Criterion values ​​for data screening

[0169]

[0170] Specifically, in step S2, during the re-testing of old wells, based on the Compton effect of radiometric logging, multiple cement sheath models with different thicknesses and densities are established to simulate downhole conditions. The absorption coefficient values ​​of cement sheaths from different models are used for fitting, and the functional relationship between the radiometric count rate, cement sheath density, and cement sheath thickness is obtained through simulation calculation. Analysis shows that the radiometric count rate, cement sheath density, and thickness have a linear relationship within a certain range. The comprehensive functional relationship between the count rate and cement sheath property parameters is derived, and finally, the influence coefficient of cement sheath on gamma logging count rate is found, a specific correction coefficient is obtained, and a correction verification experiment is conducted to make the corrected data infinitely close to the true formation value. The past and upcoming old oil wells are reinterpreted and corrected to restore the true formation radiometric effect.

[0171] Based on experiments, the expression for the comprehensive correction coefficient of the cement ring is as follows:

[0172] Y=100*Y1 / (202.628-61.447*ρ(s)-0.627*h(s))

[0173] In the formula: ρ(s) represents the density of the cement ring, in g / cm3, h(s) represents the thickness of the cement ring, in mm; and satisfies 6.15≤h(s)≤50.15; 1.63≤ρ(s)≤1.94;

[0174] In particular, in step S3, when converting between natural gamma and quantitative gamma, the correlation between the natural gamma data (unit: API) from boreholes in the Ordos Basin (Changqing Exploration Area) and the Eren Basin (Huayou Exploration Area) and the quantitative gamma measurement data (unit: cps) from the nuclear industry is established. Furthermore, the uranium grade (unit: %) and uranium content (ppm) interpreted from the oilfield boreholes are calculated. This is of great significance for guiding the deployment of the next drilling work and even estimating the prospective resources within the entire Ordos-Eren Basin oilfield exploration rights area.

[0175] Analysis suggests that the uranium anomaly in mudstone is mainly controlled by the carbonaceous content and adsorption of clay minerals in the mudstone, with relatively simple influencing factors. In contrast, the uranium anomaly in sandstone exhibits stronger vertical heterogeneity, is subject to more geological interference factors, and has a higher degree of geological complexity, resulting in a complex correlation pattern. This invention establishes the correlation between GR logging data from oilfield boreholes in mudstone sections and radioactive gamma logging data from verification wells. In the study, mudstone-type uranium mineralization and anomaly clues with a thickness of approximately 3.0 m (395–398 m) and stable distribution on the profile were observed in the Saihan Formation of the block. The anomaly sections of oilfield borehole 7M72-F130 and verification well 10D72-1 were selected as the research objects. The final conversion model used is: Y = 6.147 + 0.218X1, where X1 represents the GR value of the oilfield borehole. The table below shows the analysis results of the inferred gamma from oilfield boreholes and the measured gamma from verification wells in the Erlian Basin:

[0176] Table 2. Analysis results of inferred γ and measured γ from verification wells in the Erlian Basin oilfield.

[0177] Abnormal depth m Oilfield hole GR Oilfield hole estimated cps Validation hole actual cps 243.00 249.26 60.49 51.68 243.13 285.90 68.47 61.81 243.25 317.31 75.32 81.11 243.38 326.47 77.32 98.89 243.50 307.09 73.09 91.70 243.63 266.91 64.33 71.81 243.75 223.72 54.92 50.01 243.88 192.45 48.10 34.61 244.00 171.71 43.58 36.61

[0178] Specifically, in step S4, the interpretation of seismic data for in-situ leaching sandstone-type uranium deposits can be based on the main objectives of in-situ leaching sandstone-type uranium deposit exploration, namely, the geological environment of uranium source conditions, uranium migration, and uranium precipitation. Through the interpretation of 3D seismic data, the spatial characteristics of the region's structure, tectonic history, lithology, and sedimentary facies are analyzed and studied, thereby summarizing mineralization regularities, establishing mineralization models, and guiding the selection of peripheral areas. Based on uranium mineralization elements and combined with geological needs, the main application directions of seismic technology are divided into six aspects: tectonic and sedimentary evolution analysis, fault characterization, tectonic description, analysis of top and bottom conditions, sand body distribution and physical property prediction, and evaluation of favorable areas.

[0179] Figure 2 The diagram shows the stratigraphic features in the depth region of a sandstone-type uranium deposit according to an embodiment of the present invention. For ease of description, only the parts relevant to the embodiments of the present invention are shown, and are detailed below:

[0180] This invention develops a novel deep learning neural network model based on practical experience in uranium seismic exploration. This model predicts major faults and, based on the temporal distribution and spatial characteristics of strata and faults, combined with the time-depth relationships of typical wells in the region, performs planar time-depth conversion to establish stratigraphic structural features in the depth domain. In the Zhenyuan area, the bottom interface of the ore-bearing sand body slopes from west to east, with two structural depressions in the northern and southern regions. The strata are relatively flat in the main uranium exploration area, with a maximum elevation difference of approximately 50m and a maximum fault displacement of approximately 40m. The structural features of the top interface of the Luohe Formation are similar to those of the bottom interface of the ore-bearing sand body, exhibiting a certain degree of inheritance.

[0181] Figure 3The diagram shown is a seismic profile of a sandstone-type uranium deposit according to an embodiment of the present invention. For ease of description, only the parts relevant to the embodiments of the present invention are shown, and are detailed below:

[0182] The PY-11 well area is located in a gently sloping, low-lying region. The lower boundary of the upper Huanhe Formation is between 616 and 636 meters above sea level, with a stratigraphic angle of less than 1.5°. The lower boundary of the Huanhe Formation is between 440 and 460 meters above sea level, with a stratigraphic angle of less than 1°. The well area is less than 1 km from the Tongyuan Fault. The PY-7 well area is located in the middle of two northeast-trending strike-slip faults. The stratigraphy is gently sloping and low-lying. The lower boundary of the upper Huanhe Formation is between 430 and 440 meters above sea level. It is less than 600 meters from the southern fault and less than 1 km from the northern fault. Both faults are connected to the Chang 7 Fault. The PY-11 well area is relatively close to the PY-7 well area, about 2.5 km apart. According to the regional tectonic unit division, this area is located in the southwestern part of the Ordos Basin, belonging to the Tianhuan Depression. The overall structure of the work area is relatively gentle, with well-developed faults. Analysis of currently completed industrial uranium mines shows that uranium deposits are mostly concentrated in relatively low-lying, gently sloping areas, close to fault lines. Except for well PY-6, the other well areas all meet the requirements of low-lying, gently sloping formations and proximity to faults (less than 1 km). Furthermore, compared to several key wells, well PY-6 also has the lowest uranium content per square meter among industrial wells. Therefore, it is generally agreed that low-lying areas with gently sloping formations and proximity to fault lines (less than 1 km) are conducive to uranium enrichment and mineralization.

[0183] Specifically, in step S5, the Erlian Basin is an important uranium-producing basin in northern China, as indicated by the typical deposit comparison. The main type of uranium deposits within the basin is the paleovalley sandstone type. Currently discovered deposits include the Bayanwula large-scale deposit, the Hadatu large-scale deposit, and the Saihangobi medium-scale deposit. The main uranium mineralization within the paleovalley is hosted in the upper part of the Saihan Formation, controlled by braided river sand bodies within the upper part of the Saihan Formation. The main mineralization type of the Bayanwula deposit is the paleovalley phreatic-interstratal oxidation zone type, the Saihangobi deposit is mainly the paleovalley phreatic oxidation zone type, and the Hadatu deposit is mainly the interstratal oxidation zone type.

[0184] Analysis of the mineralization conditions of typical sandstone-type uranium deposits in the Erlian Basin reveals that various mineralization conditions are directly or indirectly reflected in the main controlling factors. A summary of the enrichment patterns of typical sandstone-type uranium deposits in the Erlian Basin shows both commonalities and unique characteristics. The commonalities are reflected in paleotectonic and paleogeographical aspects, target strata, paleoclimate, and sedimentary systems; the unique characteristics are reflected in uranium source, sedimentary facies zones, hydrological conditions, and groundwater-interstratal oxidation.

[0185] The main controlling factors for uranium mineralization in sandstone-type mineralization in the Erlian Basin are considered to be: uranium source, large-scale sand bodies, and redox zones (replenishment, discharge, and redox). Therefore, the analysis and research of the Erren Nuoer and Naomugen working areas focuses on three aspects: uranium source localization, sand body stratification, and redox localization.

[0186] Figure 4 The diagram illustrates the structure of a deep learning neural network model for fault / fracture prediction based on seismic data in sandstone-type uranium deposits, according to an embodiment of the present invention. For ease of description, only the parts relevant to the embodiment of the present invention are shown, and are detailed below: The seismic deep learning model includes: an encoder-decoder module, wherein the encoder-decoder module includes:

[0187] Encoder unit, fourth Swing Transformer unit, decoder unit, jump connection unit;

[0188] The encoder unit includes: an image block segmentation unit, an image block embedding unit, a first Swing Transformer unit, a first image block merging unit, a second Swing Transformer unit, a second image block merging unit, a third Swing Transformer unit, and a third image block merging unit. The image block segmentation unit outputs to the image block embedding unit, the image block embedding unit outputs to the first Swing Transformer unit, the first Swing Transformer unit outputs to the first image block merging unit, the first image block merging unit outputs to the second Swing Transformer unit, the second Swing Transformer unit outputs to the second image block merging unit, the second image block merging unit outputs to the third Swing Transformer unit, and the third Swing Transformer unit outputs to the third image block merging unit.

[0189] The third image block merging unit outputs to the fourth Swing Transformer unit, and the fourth Swing Transformer unit outputs to the first image block expansion unit;

[0190] The decoder unit includes: a first image block expansion unit, a fifth Swing Transformer unit, a second image block expansion unit, a sixth Swing Transformer unit, a third image block expansion unit, a seventh Swing Transformer unit, a final image expansion module, and a linear projection module. The first image block expansion unit outputs to the fifth Swing Transformer unit, the fifth Swing Transformer unit outputs to the second image block expansion unit, the second image block expansion unit outputs to the sixth Swing Transformer unit, the sixth Swing Transformer unit outputs to the third image block expansion unit, the third image block expansion unit outputs to the seventh Swing Transformer unit, the seventh Swing Transformer unit outputs to the final image expansion module, and the final image expansion module outputs to the linear projection module.

[0191] The encoder unit is connected to the decoder unit via a jumper connection unit.

[0192] In particular, the detailed information of this deep learning neural network model is still... Figure 5 The description is not elaborated here. The principle is explained as follows: Specifically, the input seismic image x is first segmented into image patch partitions and embedded with fixed-length label vectors (patch embeddings). The label vector is the basic information unit in the network architecture. For two-dimensional and three-dimensional network architectures, we set the image patch size to 4×4 and 4×4×4 respectively, and the length of the embedding vector C is 64. Without loss of generality, taking a two-dimensional seismic image as an example, the input dimension of the encoder branch is H / 4×W / 4×C. The encoder and decoder branches have a similar three-layer structure. A single level of the encoder consists of two consecutive Swing Transformer unit blocks to implement the self-attention mechanism, followed by an image patch merging block for downsampling the output. The image patch merging block connects the features of two adjacent patches, which expands the receptive field of the next (lower) level of the encoder, allowing the next level of the encoder to focus on larger-scale features. The input and output dimensions of the i-th level of the encoder are W / (4×2) and (4×4)×C respectively. i-1 )×2 i-1 C and H / (4×2) i-1 )×W / (4×2 i )×2 1 Similarly, a single level of the decoder consists of two consecutive Swing-transformer blocks, followed by a patch-expanding block that upsamples the output to restore the patch resolution of the upper layers. Between the encoder and decoder are two more consecutive Swing transformer blocks and a patch-expanding block. At the top of the decoder branch, the final patch-expanding is followed by a fully connected linear projection layer and a sigmoid activation layer, outputting an image mask displaying pixel-level fault probabilities. The output size (H×W) is the same as the input seismic image.

[0193] Furthermore, in the neural network deep learning model of this invention, for each level of decoder, the input is a combination of the output of the next level decoder and the output of the skip connection unit. This invention does not use simple skip connections to pass tomographic information between the encoder and decoder at the same level. This invention innovatively provides the decoder with multi-dimensional information from various levels of the encoder to enhance the neural network. This architecture enables the network to learn tomographic features from different dimensions simultaneously.

[0194] Figure 5 The diagram illustrates the skip connection unit structure in a deep learning neural network model for fault / fracture prediction based on seismic data in sandstone-type uranium deposits, as provided in an embodiment of the present invention. For ease of description, only the parts relevant to the embodiments of the present invention are shown, and are detailed below:

[0195] Furthermore, the encoder unit is connected to the decoder unit via a jump connection unit, including:

[0196] The outputs of the first Swing Transformer unit, the second Swing Transformer unit, and the third Swing Transformer unit are concatenated and then input into the first hop connection unit. The output of the first hop connection unit is then concatenated with the output of the first image block expansion unit and input into the fifth Swing Transformer unit.

[0197] The outputs of the first, second, and third Swing Transformer units are concatenated and then input into the second hop connection unit. The output of the second hop connection unit is then concatenated with the output of the second image block expansion unit and input into the sixth Swing Transformer unit.

[0198] The outputs of the first, second, and third Swing Transformer units are concatenated and then input into the third hop connection unit. The output of the third hop connection unit is then concatenated with the output of the second image block expansion unit and input into the seventh Swing Transformer unit.

[0199] Furthermore, the jump connection unit includes: a first jump connection unit, a second jump connection unit, and a third jump connection unit, and the first jump connection unit, the second jump connection unit, and the third jump connection unit have the same structure;

[0200] The first hop connection unit includes: the eighth Swing Transformer unit and a channel mixing module;

[0201] The channel mixing module includes two branches. The first branch includes a max pooling layer and a 1×1 convolutional layer, and the second branch includes an average pooling layer and a 1×1 convolutional layer. The outputs of the first branch and the second branch are added together and then passed to the 1×1 convolutional layer and the sigmoid activation layer to obtain the output of the channel mixing module.

[0202] The output of the eighth Swing Transformer unit and the output of the channel mixing module are multiplied together to obtain the output of the first hop connection unit.

[0203] The principle is explained as follows: Specifically, the output of all encoder levels is first downsampled, upsampled, or directly passed to the Concat module, and then enters the i-th stage of the jump connection unit module. Then, it is copied and split into two inputs: one feeds to two Swin transformer modules for calculating spatial self-attention, and the other feeds to a channel mixing module for calculating channel self-attention. Finally, the output of the Swin transformer block is projected through a fully connected linear projection layer to the same size H / (4×2) of the output of the i-th encoder. i-1 )×W / (4×2 i )×2 1 The output is C. The channel blending block has two branches, one with max pooling and the other with average pooling to reduce spatial information, and then a 1×1 convolutional layer is applied to perform channel self-attention. The outputs of the two branches are added together and then passed to another 1×1 convolutional layer and a sigmoid activation layer to obtain the channel attention result. Finally, the output of the hop connection unit is the multiplication of the channel attention and spatial attention.

[0204] Figure 6 This illustration shows another structure of a deep learning neural network model for fault / fracture prediction based on seismic data in sandstone-type uranium deposits, provided by an embodiment of the present invention. For ease of description, only the parts relevant to the embodiments of the present invention are shown, and are detailed below:

[0205] The network architecture includes Figure 6 The two encoder-decoder modules are shown.

[0206] The result of the last Swing Transformer unit in the decoder unit of the first encoder-decoder module is output to the first Swing Transformer unit in the encoder unit of the second encoder-decoder module. A jump connection unit is used to establish a jump connection between the Swing Transformer unit in the decoder unit of the first encoder-decoder module and the Swing Transformer unit in the encoder unit of the second encoder-decoder module. The specific structure is as follows:

[0207] The earthquake deep learning model includes:

[0208] First encoder-decoder module, second encoder-decoder module;

[0209] The first encoder-decoder module includes:

[0210] First encoder unit, fourth Swing Transformer unit, first decoder unit, jump connection unit;

[0211] The first encoder unit includes: an image block segmentation unit, an image block embedding unit, a first Swing Transformer unit, a first image block merging unit, a second Swing Transformer unit, a second image block merging unit, a third Swing Transformer unit, and a third image block merging unit. The image block segmentation unit outputs to the image block embedding unit, the image block embedding unit outputs to the first Swing Transformer unit, the first Swing Transformer unit outputs to the first image block merging unit, the first image block merging unit outputs to the second Swing Transformer unit, and the second Swing Transformer unit outputs to the third Swing Transformer unit.

[0212] The Transformer unit outputs to the second image block merging unit, and the second image block merging unit outputs to the third Swin.

[0213] The Transformer unit and the third Swing Transformer unit output to the third image block merging unit;

[0214] The third image block merging unit outputs to the fourth Swing Transformer unit, and the fourth Swing Transformer unit outputs to the first image block expansion unit;

[0215] The first decoder unit includes: a first image block expansion unit, a fifth Swing Transformer unit, a second image block expansion unit, a sixth Swing Transformer unit, a third image block expansion unit, a seventh Swing Transformer unit, a final image expansion module, and a linear projection module. The first image block expansion unit outputs to the fifth Swing Transformer unit, the fifth Swing Transformer unit outputs to the second image block expansion unit, the second image block expansion unit outputs to the sixth Swing Transformer unit, the sixth Swing Transformer unit outputs to the third image block expansion unit, the third image block expansion unit outputs to the seventh Swing Transformer unit, the seventh Swing Transformer unit outputs to the final image expansion module, and the final image expansion module outputs to the linear projection module.

[0216] Furthermore, the output of the seventh Swing Transformer unit is sent to the eighth Swing Transformer unit;

[0217] The second encoder-decoder module includes:

[0218] Second encoder unit, eleventh Swing Transformer unit, second decoder unit, jump connection unit;

[0219] The second encoder unit includes: an eighth Swing Transformer unit, a fourth image block merging unit, a ninth Swing Transformer unit, a fifth image block merging unit, a tenth Swing Transformer unit, and a sixth image block merging unit. The eighth Swing Transformer unit outputs to the fourth image block merging unit, the fourth image block merging unit outputs to the ninth Swing Transformer unit, the ninth Swing Transformer unit outputs to the fifth image block merging unit, the fifth image block merging unit outputs to the tenth Swing Transformer unit, and the tenth Swing Transformer unit outputs to the sixth image block merging unit.

[0220] Furthermore, the output of the sixth image block merging unit is sent to the eleventh Swing Transformer unit, and the output of the eleventh Swing Transformer unit is sent to the fourth image block expansion unit;

[0221] The second decoder unit includes: a fourth image block expansion unit, a twelfth Swing Transformer unit, a fifth image block expansion unit, a thirteenth Swing Transformer unit, a sixth image block expansion unit, a fourteenth Swing Transformer unit, a final image expansion module, and a linear projection module. The output of the fourth image block expansion unit is sent to the twelfth Swing Transformer unit, the output of the twelfth Swing Transformer unit is sent to the fifth image block expansion unit, the output of the fifth image block expansion unit is sent to the thirteenth Swing Transformer unit, the output of the thirteenth Swing Transformer unit is sent to the sixth image block expansion unit, the output of the sixth image block expansion unit is sent to the fourteenth Swing Transformer unit, the output of the fourteenth Swing Transformer unit is sent to the final image expansion module, and the output of the final image expansion module is sent to the linear projection module.

[0222] The first encoder unit is connected to the first decoder unit via a jump connection unit;

[0223] The second encoder unit is connected to the second decoder unit via a jump connection unit;

[0224] The first decoder unit is connected to the first encoder unit via a jump connection unit;

[0225] Furthermore, the first encoder unit is connected to the first decoder unit via a jump connection unit, including:

[0226] The outputs of the first Swing Transformer unit, the second Swing Transformer unit, and the third Swing Transformer unit are concatenated and then input into the first hop connection unit. The output of the first hop connection unit is then concatenated with the output of the first image block expansion unit and input into the fifth Swing Transformer unit.

[0227] The outputs of the first, second, and third Swing Transformer units are concatenated and then input into the second hop connection unit. The output of the second hop connection unit is then concatenated with the output of the second image block expansion unit and input into the sixth Swing Transformer unit.

[0228] The outputs of the first, second, and third Swing Transformer units are concatenated and then input into the third hop connection unit. The output of the third hop connection unit is then concatenated with the output of the second image block expansion unit and input into the seventh Swing Transformer unit.

[0229] Furthermore, the jump connection unit includes: a first jump connection unit, a second jump connection unit, and a third jump connection unit, and the first jump connection unit, the second jump connection unit, and the third jump connection unit have the same structure;

[0230] Furthermore, the second encoder unit is connected to the second decoder unit via a jump connection unit, including:

[0231] The outputs of the eighth, ninth, and tenth Swing Transformer units are concatenated and then input into the fourth hop connection unit. The output of the fourth hop connection unit is then concatenated with the output of the sixth image block expansion unit and input into the fourteenth Swing Transformer unit.

[0232] The outputs of the eighth, ninth, and tenth Swing Transformer units are concatenated and then input into the fifth hop connection unit. The output of the fifth hop connection unit is then concatenated with the output of the fifth image block expansion unit and input into the thirteenth Swing Transformer unit.

[0233] The outputs of the eighth, ninth, and tenth Swing Transformer units are concatenated and then input into the fifth hop connection unit. The output of the sixth hop connection unit is then concatenated with the output of the fourth image block expansion unit and input into the twelfth Swing Transformer unit.

[0234] Furthermore, the first decoder unit is connected to the first encoder unit via a jump connection unit, including:

[0235] The outputs of the fifth, sixth, and seventh Swing Transformer units are concatenated and then input into the seventh hop connection unit. The output of the seventh hop connection unit is then concatenated with the output of the fourth image block merging unit and input into the ninth Swing Transformer unit.

[0236] The outputs of the fifth, sixth, and seventh Swing Transformer units are concatenated and then input into the eighth hop connection unit. The output of the eighth hop connection unit is then concatenated with the output of the fifth image block merging unit and input into the tenth Swing Transformer unit.

[0237] In particular, the structures of the first, second, third, fourth, fifth, sixth, seventh, and eighth jump connection units are all the same.

[0238] In particular, Figure 6 The model principle of each encoder-decoder unit in the two encoder-decoder units shown is the same as... Figure 4 The principle of the single encoder-decoder unit model shown is the same, and will not be repeated here.

[0239] Experiments show that by setting up this two encoder-decoder module, the features of the fault can be better extracted for learning and training, thereby achieving high-precision fault feature extraction.

[0240] Figure 7 The diagram illustrates the Swin transformer module structure in a deep learning neural network model for fault / fracture prediction based on seismic data in sandstone-type uranium deposits, as provided in an embodiment of the present invention. For ease of description, only the parts relevant to the embodiments of the present invention are shown, and are detailed below:

[0241] Figure 7 The codes LN, W-MSA, SW-MSA, MLP-GELU, etc., are explained as follows:

[0242] Layer Normalization (LN): In the Swin Transformer, LayerNorm is used to normalize the input of layers to stabilize and accelerate the training process. It is applied before each Multi-Head Self-Attention (MSA) module and each MLP, helping to improve the training efficiency and performance of the model.

[0243] Window-based Multi-Head Self-Attention (W-MSA): W-MSA is one of the core components of the Swin Transformer. It segments the image into non-overlapping windows and applies a self-attention mechanism independently within each window. This approach reduces computation because self-attention is performed only within a local window, rather than across the entire image. This design allows the model to capture local features while maintaining computational efficiency.

[0244] Shifted Window Multi-Head Self-Attention (SW-MSA): To overcome the information isolation problem between windows in W-MSA, the Swin Transformer introduces SW-MSA. By sliding between adjacent windows, SW-MSA allows information to be passed between windows, thereby enhancing the model's ability to capture global information. W-MSA and SW-MSA are used alternately in the Swin Transformer to balance the processing of local and global information.

[0245] Multi-Layer Perceptron (MLP-GELU): In each Transformer block of the Swin Transformer, the MLP is used to further process features processed by the self-attention mechanism. The MLP typically contains two fully connected layers, with GELU (Gaussian Error Linear Unit) as the activation function in between. The MLP is applied after the self-attention layer to increase the model's non-linear expressive power. The GELU activation function is widely used due to its excellent performance in deep learning, helping the model to better learn complex feature representations.

[0246] Here, the MLP layer has two fully connected linear layers with Gaussian error and linear unit activation functions. A skip connection adding the input of the Swing transformer module and the W-MSA / SW-MSA output is designed to pass information to the second LN and the MLP. Another skip connection adds the input of the second LN and the output of the MLP, and passes information to the next Swing transformer module.

[0247] The key components of the Swin transformer block are W-MSA and SW-MSA. The input to W-MSA and SW-MSA is a large token matrix. To reduce computational cost, the token matrix is ​​divided into 8×8 windows based on the actual locations in the original image. Multi-head self-attention is computed in parallel within each window. The self-attention mechanism is as follows:

[0248]

[0249] Where Q, K, and V are the query matrix, key matrix, and value matrix, respectively, calculated from the three linear projections of the tag matrix. The linear projection layers have different weights W. Q W K and W V A fully connected layer. Where d k This is the number of features used for normalization. The self-attention mechanism computes the consistency of each feature across different embedded label vectors; therefore, it can naturally measure the global coherence between label vectors. To incorporate coherence between image patches in different windows, a shifted window mechanism is implemented. Here, the window used to compute multi-head self-attention is shifted cyclically down and to the right four levels. If the window moves out of bounds, it will pick up missing patches from the top and left. This cyclic shifting mechanism ensures that the self-attention mechanism is not limited to the local window and maintains long-distance and global connectivity between image patches.

[0250] Figure 8 The diagram shows a 2D prediction result (vertical profile) of a deep learning neural network model for fault / fracture prediction based on seismic data in sandstone-type uranium deposits, according to an embodiment of the present invention. For ease of description, only the parts relevant to the embodiments of the present invention are shown, and are detailed below:

[0251] Figure 8 (a) shows the tomographic prediction results of a single encoder-decoder. Figure 8 (b) Tomographic prediction results of two encoder-decoder pairs. As can be seen from the figure, both single encoder-decoder and two encoder-decoder pairs achieved good tomographic identification results.

[0252] Figure 9 The diagram shows a 2D prediction result (lateral profile) of a deep learning neural network model for fault / fracture prediction based on seismic data in sandstone-type uranium deposits, according to an embodiment of the present invention. For ease of description, only the parts relevant to the embodiments of the present invention are shown, and are detailed below:

[0253] Figure 9 (a) shows the tomographic prediction results of a single encoder-decoder. Figure 9 (b) Tomographic prediction results of two encoder-decoder pairs. As can be seen from the figure, both single encoder-decoder and two encoder-decoder pairs achieved good tomographic identification results.

[0254] Figure 10 The image shows 3D prediction results of a deep learning neural network model for fault / fracture prediction based on seismic data in sandstone-type uranium deposits, according to an embodiment of the present invention. For ease of description, only the parts relevant to the embodiments of the present invention are shown, and are detailed below:

[0255] Figure 10(a) shows the tomographic prediction results of a single encoder-decoder. Figure 10 (b) Tomographic prediction results of two encoder-decoder pairs. As can be seen from the figure, both single encoder-decoder and two encoder-decoder pairs achieved good tomographic identification results.

[0256] The loss function used in this invention will be introduced below:

[0257] In deep learning, the loss function is a crucial component that measures the difference between the model's predictions and the actual results. In other words, it evaluates the model's performance by calculating the degree of inconsistency between the predicted and true values. During training, the model aims to minimize the loss function by adjusting its parameters. The model updates its parameters using the backpropagation algorithm to reduce the loss between the true and predicted values. In this way, the model's predicted values ​​gradually converge towards the true values, thus achieving the learning objective.

[0258] Loss functions are mainly classified into several types according to task type, including:

[0259] Distance-based loss functions: These loss functions typically map the input data to a distance-based feature space (such as Euclidean space, Hamming space, etc.), and then use an appropriate loss function to measure the distance between the true values ​​of the samples and the model's predicted values ​​in the feature space. The smaller the distance, the better the model's predictive performance. Common distance-based loss functions include MSE, LAE loss, and LSE loss.

[0260] MSE: Calculates the average of the sum of squares of the differences between the predicted and actual values. MSE penalizes larger errors more heavily and is sensitive to outliers.

[0261] LAE loss, also known as minimizing absolute error (LAE), calculates the sum of the absolute values ​​of the differences between the predicted and true values. LAE loss is relatively insensitive to outliers, and its gradient is discontinuous at zero, which helps to produce sparse solutions (i.e., some features have zero weights).

[0262] LSE loss, also known as least squares error (LSE), minimizes the sum of squares of the differences between the target value and the estimated value. LSE loss is sensitive to outliers, and its gradient is continuous at zero.

[0263] Loss functions based on probability distribution metrics: These loss functions transform the similarity between samples into the probability of random events occurring, judging the similarity between the two by measuring the distance between the true distribution of the samples and their estimated distribution. They are particularly commonly used in classification problems. Common loss functions based on probability distribution metrics include cross-entropy loss and the KL divergence function.

[0264] Cross-entropy loss measures the difference between the predicted probability distribution and the true label probability distribution. In binary classification problems, cross-entropy loss can be expressed as the sum of the negative logarithms of the product of the predicted probability of the positive class and the true label probability (the same applies to the negative class). In multi-class classification problems, cross-entropy loss compares the predicted probability of each class with the probability distribution of the true label and sums the results. The smaller the value of cross-entropy loss, the better the model performance.

[0265] The KL divergence function, also known as relative entropy or information divergence, measures the difference between two probability distributions. KL divergence is non-negative and takes the value zero if and only if the two distributions are identical. In deep learning, KL divergence is often used in regularization terms to encourage the model's output probability distribution to approximate the prior distribution.

[0266] Based on the need for a deep learning model for fault identification in seismic exploration of uranium sandstone-type uranium deposits, this invention proposes a hybrid loss function for the seismic deep learning model:

[0267]

[0268] L = ω1L1 + ω2L2, and we have: ω1 + ω2 = 1

[0269] Among them, y i =sigmoid(x i ), where xi is the input to the deep learning network, and y i ' represents the pixel label, α represents the ratio of non-fault pixels to the total number of pixels; N represents the total number of pixels in the seismic image, and i represents the pixel index. ω1 and ω2 are the weighting coefficients, respectively.

[0270] To accurately measure the predictive performance of the network model, this embodiment of the invention uses four accuracy evaluation metrics—precision, recall, F1 score, and mean intersection over union (MLOU)—to determine the weights.

[0271] Precision is the proportion of samples predicted as positive that are actually positive. It measures how accurately the model predicts positive samples when they are actually positive. The formula is: Where: TP (True Positive) represents the number of samples correctly predicted as positive by the model. FP (False Positive) represents the number of samples incorrectly predicted as positive by the model.

[0272] Recall is the proportion of actual positive samples that are predicted as positive. It measures the model's ability to capture all positive samples. The calculation formula is: FN (False Negative) represents the number of positive samples that the model incorrectly predicts as negative.

[0273] The F1 score is the harmonic mean of precision and recall, attempting to find a balance between the two. Calculation formula: mIoU is the average Intersection over Union (IoU) of multiple classes, used to measure the segmentation performance of a model, especially in multi-class segmentation tasks. Calculation formula: Where N is the number of categories, and IoUi is the intersection-union ratio of the i-th category.

[0274] Table 3 shows the comparison results of different weight values ​​for different hybrid loss functions:

[0275]

[0276] As can be seen from Table 3, the first set of weights has the worst effect, obviously having a side effect, resulting in poor network fitting. When the weight ratios are 0.2 and 0.8, all evaluation indicators are the highest, that is, when the L1 loss function accounts for 0.2 and the L2 loss function accounts for 0.8, the network convergence is the best.

[0277] Furthermore, this invention found in its research that when the weight coefficients in the hybrid loss function remain constant, using fixed weight coefficients may cause the model to favor the majority class when dealing with imbalanced data, thus neglecting the feature learning of the minority class. This can lead to overfitting of the model to certain classes during training, especially when the class distribution is uneven. The model may overemphasize the majority class while ignoring the minority class, resulting in good performance on the training set but poor generalization ability on broader datasets. Therefore, this invention proposes that the weight coefficients ω1 and ω2 in the hybrid loss function are variable, and their calculation method is as follows:

[0278] in:

[0279]

[0280] L t =ω 1,t-1 L 1,t +ω 2,t-1 L 2,t

[0281] Among them, L j,t Let u(L) be the j-th loss function at time t. j,t-1 () represents the average value of the j-th type of loss before time t-1.

[0282] σ(r i,t ) and μ(r i,t ) represent r before time t respectively j,t The standard deviation and mean of j, and j takes the value of 1 or 2.

[0283] The weighting coefficients can vary according to the prediction difficulty of data instances, allowing the model to adjust its focus on instances of different difficulty during training. This improves the model's generalization ability, enhances its calibration performance, and makes the probabilities output by the model more reliable.

[0284] Figure 11 The diagram shows the weight coefficients of the loss function of a deep learning neural network model for fault / fracture prediction based on seismic data in sandstone-type uranium deposits, according to an embodiment of the present invention, as a function of the present invention. For ease of description, only the parts relevant to the embodiment of the present invention are shown, and are detailed below:

[0285] Figure 11 This demonstrates how the weights of two loss functions (loss1(L1) and loss2(L2)) adjust with the number of training epochs during the training of a deep learning model. Initial Stage: In the initial stage of training, the weight of loss1 decreases rapidly, from nearly 0.5 to approximately 0.4. Simultaneously, the weight of loss2 increases rapidly, from nearly 0.5 to approximately 0.6. This indicates that in the initial stage, the model tends to optimize loss2, as loss2 is more helpful for the model's learning in the early stages. Mid-Stage Stage: In this stage, the weight of loss1 continues to decrease slowly, but at a slower rate. The weight of loss2 reaches a plateau in the mid-stage and then continues to increase slowly. This means that as training progresses, the contribution of loss1 to the model gradually decreases, while loss2 continues to maintain its importance. Late Stage Stage: In the later stages of training, the weight of loss1 continues to decrease, but the change is small, eventually stabilizing at approximately 0.2. The weight of loss2 continues to increase in the later stages, eventually approaching 0.8, showing its importance in the later stages of model training. This trend may indicate that as the model approaches convergence, loss2 becomes more critical for improving model performance. Figure 11 This demonstrates how the weights of the two components in the hybrid loss function dynamically change during training to optimize model performance. This adjustment strategy helps the model better balance the contributions of different loss functions during training, thereby improving the final model performance.

[0286] Figure 12 The illustration shows a five-in-one oil and uranium exploration device according to an embodiment of the present invention. For ease of description, only the parts relevant to the embodiment of the present invention are shown, and the details are as follows: The five-in-one oil and uranium exploration device includes:

[0287] Data screening unit: Conduct data screening, which includes:

[0288] First screening module: collecting, surveying, and organizing borehole information from the oilfield;

[0289] The second screening module analyzes the standards for abnormal radioactivity in boreholes and constructs software for screening abnormal radioactivity.

[0290] The third screening module: Based on the radioactive anomaly screening software, it identifies wells with radioactive anomalies according to the borehole radioactive anomaly standards;

[0291] Old Well Re-survey Unit: Conducting old well re-surveys, which includes:

[0292] First retesting module: Establishing the sleeve-cement ring model;

[0293] Second retest module: Conducting logging response tests on cement ring casing;

[0294] The third retest module analyzes the functional relationship between the count rate and the cement ring property parameters to determine the correction coefficient.

[0295] Fourth retesting module: Obtain accurate shallow strata radioactivity information based on the correction coefficient;

[0296] Gamma conversion unit: Conduct drilling operations to verify and achieve quantitative gamma conversion from natural gamma. The gamma conversion unit includes:

[0297] First Gamma Module: Collects gamma data from verification boreholes and oilfield boreholes;

[0298] Second Gamma Module: Corrects the verification hole gamma logging depth to the range of abnormal depths in oilfield borehole gamma;

[0299] The third gamma module: Establishing the regression equation between natural gamma and quantitative gamma;

[0300] The fourth gamma module: uses regression equations to convert between natural gamma and quantitative gamma, and evaluates mineral content;

[0301] Seismic processing and interpretation unit: Conducts 3D seismic data processing, interpretation, and analysis of faults, sand bodies, and structures in the mining area;

[0302] Ore deposit comparison unit: Conduct comparisons of typical ore deposits, including:

[0303] First comparison module: Collecting data on typical uranium deposits;

[0304] The second comparative module analyzes the structural sedimentary evolution characteristics, uranium source conditions, sand bodies and sedimentary systems, ore body characteristics, and metallogenic regularities.

[0305] The third comparative module summarizes the ore-controlling factors of typical sandstone-type uranium deposits;

[0306] Comprehensive processing unit: Based on the results of the data screening unit, the old well re-survey unit, the gamma conversion unit, the seismic processing and interpretation unit, and the deposit comparison unit, determine the favorable uranium ore zones and evaluate the mineralization potential of the working area.

[0307] Embodiments of the present invention also provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of a five-in-one oil and uranium exploration method.

[0308] Embodiments of the present invention also provide a computer storage medium storing a computer program, characterized in that the computer program, when executed by a processor, implements the steps of a five-in-one oil and uranium exploration method.

[0309] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The present invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, produce implementations of the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0310] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0311] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0312] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A five-in-one method for oil and uranium exploration, characterized in that, include: Step S1: Conduct data screening, wherein step S1 includes: Step S11: Collect, investigate, and organize borehole information from the oilfield; Step S12: Analyze the standards for borehole radioactivity anomalies and construct radioactivity anomaly screening software; Step S13: Based on the radioactive anomaly screening software, judge according to the borehole radioactive anomaly standard, and screen out radioactive anomaly wells; Step S2: Conduct re-survey of the old well, wherein step S2 includes: Step S21: Establish the casing-cement ring model; Step S22: Conduct a logging response test using cement ring casing; Step S23: Analyze the functional relationship between the count rate and the cement ring property parameters, and determine the correction coefficient; Step S24: Obtain accurate shallow strata radioactivity information based on the correction coefficient; Step S3: Conduct drilling operations to verify and achieve quantitative gamma conversion from natural gamma. Step S3 includes: Step S31: Collect gamma data from verification wells and oilfield boreholes; Step S32: Correct the verification hole gamma logging depth to the range of abnormal depths in the oilfield borehole gamma. Step S33: Establish the regression equation between natural gamma and quantitative gamma; Step S34: Use the regression equation to convert between natural gamma and quantitative gamma, and evaluate mineral content; Step S4: Conduct 3D seismic data processing, interpretation, and analysis of faults, sand bodies, and structures in the mining area; Step S5: Conduct a comparison of typical mineral deposits. Step S5 includes: Step S51: Collect data on typical uranium deposits; Step S52: Analyze the structural sedimentary evolution characteristics, uranium source conditions, sand bodies and sedimentary systems, ore body characteristics, and mineralization regularity; Step S53: Summarize the ore-controlling factors of typical sandstone-type uranium deposits; Step S6: Based on the results of steps S1 to S5 above, determine the favorable uranium ore zones and evaluate the mineralization potential of the working area; Step S4 further includes: S41: Determine the tectonic characteristics of uranium-bearing areas using seismic fine tectonic interpretation techniques and tectonic and sedimentary evolution analysis techniques; S42: Using seismic staging and grading fault fine characterization technology to determine uranium ore faults, oil and gas faults, and oil and gas migration channels; S43: Determine the condition characteristics of the top and bottom plates of uranium ore sand bodies using seismic identification technology for the top and bottom plates of sand bodies; S44: Determine the distribution, thickness, and physical properties of uranium sandstone using fine-grained stratigraphic and reservoir characterization techniques; The step S42, which utilizes seismic staging and grading fault fine characterization technology to determine uranium ore faults, oil and gas faults, and oil and gas migration channels, includes: A seismic deep learning model is used to finely characterize uranium ore fractures. The seismic deep learning model includes an encoder-decoder module.

2. The five-in-one oil and uranium exploration method according to claim 1, characterized in that, The encoder-decoder module includes: Encoder unit, fourth Swing Transformer unit, decoder unit, jump connection unit; The encoder unit includes: an image block segmentation unit, an image block embedding unit, a first Swing Transformer unit, a first image block merging unit, a second Swing Transformer unit, a second image block merging unit, a third Swing Transformer unit, and a third image block merging unit. The image block segmentation unit outputs to the image block embedding unit, the image block embedding unit outputs to the first Swing Transformer unit, the first Swing Transformer unit outputs to the first image block merging unit, the first image block merging unit outputs to the second Swing Transformer unit, the second Swing Transformer unit outputs to the second image block merging unit, the second image block merging unit outputs to the third Swing Transformer unit, and the third Swing Transformer unit outputs to the third image block merging unit. The third image block merging unit outputs to the fourth Swing Transformer unit, and the fourth Swing Transformer unit outputs to the first image block expansion unit; The decoder unit includes: a first image block expansion unit, a fifth Swing Transformer unit, a second image block expansion unit, a sixth Swing Transformer unit, a third image block expansion unit, a seventh Swing Transformer unit, a final image expansion module, and a linear projection module. The first image block expansion unit outputs to the fifth Swing Transformer unit, the fifth Swing Transformer unit outputs to the second image block expansion unit, the second image block expansion unit outputs to the sixth Swing Transformer unit, the sixth Swing Transformer unit outputs to the third image block expansion unit, the third image block expansion unit outputs to the seventh Swing Transformer unit, the seventh Swing Transformer unit outputs to the final image expansion module, and the final image expansion module outputs to the linear projection module. The encoder unit is connected to the decoder unit via a jumper connection unit.

3. The five-in-one oil and uranium exploration method according to claim 2, characterized in that, The encoder unit is connected to the decoder unit via a jump connection unit, including: The outputs of the first, second, and third Swing Transformer units are concatenated and then input into the first hop connection unit. The output of the first hop connection unit is then concatenated with the output of the first image block expansion unit and input into the fifth Swing Transformer unit. The outputs of the first, second, and third Swing Transformer units are concatenated and then input into the second hop connection unit. The output of the second hop connection unit is then concatenated with the output of the second image block expansion unit and input into the sixth Swing Transformer unit. The outputs of the first, second, and third Swing Transformer units are concatenated and then input into the third hop connection unit. The output of the third hop connection unit is then concatenated with the output of the second image block expansion unit and input into the seventh Swing Transformer unit.

4. The five-in-one oil and uranium exploration method according to claim 3, characterized in that, The jump connection unit includes: a first jump connection unit, a second jump connection unit, and a third jump connection unit, and the first jump connection unit, the second jump connection unit, and the third jump connection unit have the same structure; The first hop connection unit includes: the eighth Swing Transformer unit and a channel mixing module; The channel mixing module includes two branches: the first branch includes a max pooling layer and a 1×1 convolutional layer, and the second branch includes an average pooling layer and a 1×1 convolutional layer. The outputs of the first and second branches are added together and then passed to the 1×1 convolutional layer and the sigmoid activation layer to obtain the output of the channel mixing module. The output of the eighth Swing Transformer unit and the output of the channel mixing module are multiplied together to obtain the output of the first hop connection unit.

5. The five-in-one oil and uranium exploration method according to claim 1, characterized in that, The earthquake deep learning model includes: First encoder-decoder module, second encoder-decoder module; The first encoder-decoder module includes: First encoder unit, fourth Swing Transformer unit, first decoder unit, jump connection unit; The first encoder unit includes: an image block segmentation unit, an image block embedding unit, a first Swing Transformer unit, a first image block merging unit, a second Swing Transformer unit, a second image block merging unit, a third Swing Transformer unit, and a third image block merging unit. The image block segmentation unit outputs to the image block embedding unit, the image block embedding unit outputs to the first Swing Transformer unit, the first Swing Transformer unit outputs to the first image block merging unit, the first image block merging unit outputs to the second Swing Transformer unit, the second Swing Transformer unit outputs to the second image block merging unit, the second image block merging unit outputs to the third Swing Transformer unit, and the third Swing Transformer unit outputs to the third image block merging unit. The third image block merging unit outputs to the fourth Swing Transformer unit, and the fourth Swing Transformer unit outputs to the first image block expansion unit; The first decoder unit includes: a first image block expansion unit, a fifth Swing Transformer unit, a second image block expansion unit, a sixth Swing Transformer unit, a third image block expansion unit, a seventh Swing Transformer unit, a final image expansion module, and a linear projection module. The first image block expansion unit outputs to the fifth Swing Transformer unit, the fifth Swing Transformer unit outputs to the second image block expansion unit, the second image block expansion unit outputs to the sixth Swing Transformer unit, the sixth Swing Transformer unit outputs to the third image block expansion unit, the third image block expansion unit outputs to the seventh Swing Transformer unit, the seventh Swing Transformer unit outputs to the final image expansion module, and the final image expansion module outputs to the linear projection module. Furthermore, the output of the seventh Swing Transformer unit is sent to the eighth Swing Transformer unit; The second encoder-decoder module includes: Second encoder unit, eleventh Swing Transformer unit, second decoder unit, jump connection unit; The second encoder unit includes: an eighth Swing Transformer unit, a fourth image block merging unit, a ninth Swing Transformer unit, a fifth image block merging unit, a tenth Swing Transformer unit, and a sixth image block merging unit. The eighth Swing Transformer unit outputs to the fourth image block merging unit, the fourth image block merging unit outputs to the ninth Swing Transformer unit, the ninth Swing Transformer unit outputs to the fifth image block merging unit, the fifth image block merging unit outputs to the tenth Swing Transformer unit, and the tenth Swing Transformer unit outputs to the sixth image block merging unit. Furthermore, the output of the sixth image block merging unit is sent to the eleventh Swing Transformer unit, and the output of the eleventh Swing Transformer unit is sent to the fourth image block expansion unit; The second decoder unit includes: a fourth image block expansion unit, a twelfth Swing Transformer unit, a fifth image block expansion unit, a thirteenth Swing Transformer unit, a sixth image block expansion unit, a fourteenth Swing Transformer unit, a final image expansion module, and a linear projection module. The output of the fourth image block expansion unit is sent to the twelfth Swing Transformer unit, the output of the twelfth Swing Transformer unit is sent to the fifth image block expansion unit, the output of the fifth image block expansion unit is sent to the thirteenth Swing Transformer unit, the output of the thirteenth Swing Transformer unit is sent to the sixth image block expansion unit, the output of the sixth image block expansion unit is sent to the fourteenth Swing Transformer unit, the output of the fourteenth Swing Transformer unit is sent to the final image expansion module, and the output of the final image expansion module is sent to the linear projection module. The first encoder unit is connected to the first decoder unit via a jump connection unit; The second encoder unit is connected to the second decoder unit via a jump connection unit; The first decoder unit is connected to the first encoder unit via a jump connection unit; Furthermore, the first encoder unit is connected to the first decoder unit via a jump connection unit, including: The outputs of the first, second, and third Swing Transformer units are concatenated and then input into the first hop connection unit. The output of the first hop connection unit is then concatenated with the output of the first image block expansion unit and input into the fifth Swing Transformer unit. The outputs of the first, second, and third Swing Transformer units are concatenated and then input into the second hop connection unit. The output of the second hop connection unit is then concatenated with the output of the second image block expansion unit and input into the sixth Swing Transformer unit. The outputs of the first, second, and third Swing Transformer units are concatenated and then input into the third hop connection unit. The output of the third hop connection unit is then concatenated with the output of the second image block expansion unit and input into the seventh Swing Transformer unit. Furthermore, the jump connection unit includes: a first jump connection unit, a second jump connection unit, and a third jump connection unit, and the first jump connection unit, the second jump connection unit, and the third jump connection unit have the same structure; Furthermore, the second encoder unit is connected to the second decoder unit via a jump connection unit, including: The outputs of the eighth, ninth, and tenth Swing Transformer units are concatenated and then input into the fourth hop connection unit. The output of the fourth hop connection unit is then concatenated with the output of the sixth image block expansion unit and input into the fourteenth Swing Transformer unit. The outputs of the eighth, ninth, and tenth Swing Transformer units are concatenated and then input into the fifth hop connection unit. The output of the fifth hop connection unit is then concatenated with the output of the fifth image block expansion unit and input into the thirteenth Swing Transformer unit. The outputs of the eighth, ninth, and tenth Swing Transformer units are concatenated and then input into the fifth hop connection unit. The output of the sixth hop connection unit is then concatenated with the output of the fourth image block expansion unit and input into the twelfth Swing Transformer unit. Furthermore, the first decoder unit is connected to the first encoder unit via a jump connection unit, including: The outputs of the fifth, sixth, and seventh Swing Transformer units are concatenated and then input into the seventh hop connection unit. The output of the seventh hop connection unit is then concatenated with the output of the fourth image block merging unit and input into the ninth Swing Transformer unit. The outputs of the fifth, sixth, and seventh Swing Transformer units are concatenated and then input into the eighth hop connection unit. The output of the eighth hop connection unit is then concatenated with the output of the fifth image block merging unit and input into the tenth Swing Transformer unit.

6. The five-in-one oil and uranium exploration method according to claim 4 or 5, characterized in that, The hybrid loss function used in the earthquake deep learning model is: ； ； ； ； in, xi is the input to the deep learning network. Here, α represents the pixel label, α is the ratio of non-fault pixels to the total number of pixels, N is the total number of pixels in the seismic image, and i is the pixel index. , These are the weighting coefficients.

7. The five-in-one oil and uranium exploration method according to claim 6, characterized in that, The weight coefficients in the hybrid loss function To obtain the data dynamically over time, the calculation method is as follows: ,in: ； ； ； Where is the j-th loss function at time t, represents the average value of the j-th loss before time t-1, and and are the standard deviation and average value before time t, respectively, and j takes the value of 1 or 2.

8. An apparatus for a five-in-one oil and uranium exploration method according to any one of claims 1 to 7, characterized in that, include: Data screening unit: Conduct data screening, which includes: First screening module: collecting, surveying, and organizing borehole information from the oilfield; The second screening module analyzes the standards for abnormal radioactivity in boreholes and constructs software for screening abnormal radioactivity. The third screening module: Based on the radioactive anomaly screening software, it identifies wells with radioactive anomalies according to the borehole radioactive anomaly standards; Old Well Re-survey Unit: Conducting old well re-surveys, which includes: First retesting module: Establishing the sleeve-cement ring model; Second retest module: Conducting logging response tests on cement ring casing; The third retest module analyzes the functional relationship between the count rate and the cement ring property parameters to determine the correction coefficient. Fourth retesting module: Obtain accurate shallow strata radioactivity information based on the correction coefficient; Gamma conversion unit: Conduct drilling operations to verify and achieve quantitative gamma conversion from natural gamma. The gamma conversion unit includes: First Gamma Module: Collects gamma data from verification boreholes and oilfield boreholes; Second Gamma Module: Corrects the verification hole gamma logging depth to the range of abnormal depths in oilfield borehole gamma; The third gamma module: Establishing the regression equation between natural gamma and quantitative gamma; The fourth gamma module: uses regression equations to convert between natural gamma and quantitative gamma, and evaluates mineral content; Seismic processing and interpretation unit: Conducts 3D seismic data processing, interpretation, and analysis of faults, sand bodies, and structures in the mining area; Ore deposit comparison unit: Conduct comparisons of typical ore deposits, including: First comparison module: Collecting data on typical uranium deposits; The second comparative module analyzes the structural sedimentary evolution characteristics, uranium source conditions, sand bodies and sedimentary systems, ore body characteristics, and metallogenic regularities. The third comparative module summarizes the ore-controlling factors of typical sandstone-type uranium deposits; Comprehensive processing unit: Based on the results of the data screening unit, the old well re-survey unit, the gamma conversion unit, the seismic processing and interpretation unit, and the deposit comparison unit, determine the favorable uranium ore zones and evaluate the mineralization potential of the working area.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the five-in-one oil and uranium exploration method according to any one of claims 1 to 7.

Images