A method for identifying stratigraphic structure in red soil areas based on shallow seismic refraction wave method

Abstract

The present invention relates to a method for identifying stratigraphic structure in red soil areas based on a shallow seismic refraction wave method. First, by designing the execution of steps A to K, optimal prediction models corresponding to the stratigraphic structure depths of each target operation area are obtained; then step i is executed to obtain the spatial distribution of the stratigraphic structure depths of each target operation area according to the optimal prediction models corresponding to the stratigraphic structure depths; the design scheme fully retains the advantages of existing seismic exploration technology, improves the shortcomings of limited application of the shallow seismic refraction wave method in red soil areas, and organically combines geographic information systems with spatial prediction, thereby realizing the universal shallow seismic refraction wave method technology of "traditional technology, improved operation; spatial prediction, improved efficiency", and has broad industrial application prospects in geological exploration, hydrogeological survey and other departments.

Description

Technical Field

[0001] The invention relates to a method for identifying stratum structure in red soil areas based on a shallow seismic refraction wave method, and belongs to the technical field of geophysical exploration. Background Art

[0002] Seismic refraction technology was first used in the field of oil exploration in the early 20th century, and then gradually turned to civilian use. In the early days of application, the main source excitation method was blasting, and the relevant technicians at that time were able to design a delay time calculation method to solve the target problem. With the rapid development of the economy and society, countries around the world were forced to meet the social needs of a large number of infrastructure construction, and seismic wave exploration technology was gradually applied in the field of engineering surveys. During this period, data acquisition technology and signal interpretation methods were further developed.

[0003] In the 1930s and 1950s, seismographs were mainly single-channel counting types, which could stimulate the source by artificial hammering. However, this method often had problems when dealing with multiple points and sections, and the efficiency of field operations was low. In the 1950s and 1970s, the main instrument model used was the multi-channel seismograph, which recorded data signals on photosensitive materials in the form of light spot focusing, and could provide dynamic characteristic information such as earthquake wave amplitude and frequency, greatly improving the accuracy and speed of data processing and interpretation. Because the intelligence level of the operating system technology is low, the waveform collection range is small, the signal-to-noise ratio is low, and the accuracy of data analysis is relatively high, which depends to a large extent on the professional level of the staff. Since the 1980s, digital seismographs based on signal enhancement have become the main operating instruments, equipped with better lightweight materials and more efficient signal enhancement technology. At the same time, based on computer automatic processing technology, the instrument operation steps have been greatly reduced, and ordinary people can use it after simple training.

[0004] The widespread application of shallow seismic refraction wave technology in my country was relatively lagging. It was not until the 1980s that relevant seismic wave technology began to be applied to engineering construction, and the production technology of high-precision seismograph host and detector equipped with microcomputer control system also matured. In recent years, my country's seismic instruments and equipment have been continuously updated and upgraded, and the prediction accuracy of the equipment has become higher and higher. Many progress has been made in the exploration of the earth's surface in different fields.

[0005] With the continuous advancement of seismic wave instruments and equipment, the interpretation level of seismic data has also been continuously improved, and many analysis methods and improved technologies have emerged at home and abroad. In the 1930s, scholars designed the wavefront method based on Huygens' principle and used Snell's law to analyze seismic wave signals. This method has poor application effect in areas where the underlying strata fluctuate violently. In the 1940s, the time term method appeared to deal with the uncertain path problem of seismic interpretation; in the 1950s, the time intercept method appeared, which used travel time curves to measure the apparent velocity and intercept time of seismic waves. It can be applied to multi-layer geological structures. This method is suitable for situations where the stratum dip is <10°. In the 1980s, the seismic ray interchange method was widely used by scholars. This method can predict the structure of simple strata and the change of seismic wave velocity; the later improved method, the generalized interchange method, can calculate the optimal distance between signals, and then deduce the seismic wave propagation velocity and depth of irregular strata.

[0006] At present, the shallow seismic refraction wave method has been widely used in the fields of hydrogeological survey, petroleum exploration, engineering geology, scientific research and environmental geological exploration. The shallow seismic refraction wave method refers to the refraction phenomenon of seismic waves emitted by artificial ground in the near-surface stratigraphic structure. The instrument can analyze the spatial distribution characteristics of the stratigraphic structure through the time from the emission to the arrival of the refracted wave at the detector and the reflected seismic wave signal. In general, the analysis results of the shallow seismic refraction wave method show that the shallow surface can usually be divided into three layers: the cover layer (soil, loose accumulation or loose sediment), the semi-weathered layer (loose rock with different degrees of cracks) and the bedrock layer (slightly weathered or non-cracked dense rock). By combining with geographic information system technology, multiple stratigraphic sequences based on seismic wave refraction interpretation can construct the three-dimensional structural hierarchy of the region, and then combine the geochemical model to calculate the weathering rate of the bedrock.

[0007] However, different industries have different business requirements for seismographs. Due to budgetary constraints, traditional drilling-based geological exploration is difficult to meet the needs of large-scale stratigraphic structure surveys. Unlike general stratigraphic structures, a large number of Quaternary red clay (red weathering crust) deposits are distributed in southern my country, including homogeneous red soil layer, reticular red soil layer and red sandstone from top to bottom. The homogeneous red soil layer has a shorter geological age, is mainly reddish brown, does not contain gravel, and is relatively loose, suitable for crop growth; the reticular red soil layer was formed longer than the homogeneous red soil layer, has a relatively compact texture, is full of red and yellow spots or stripes, has a heavy texture, is usually relatively compacted, and precipitation in the rainy season is difficult to infiltrate, which can easily aggravate the scouring of the slope and cause serious soil erosion. The structural difference between the homogeneous red soil layer and the reticular red soil layer can be easily distinguished by visual methods in the field, but the contact between the two usually presents a gradient feature, which is difficult to directly distinguish based on traditional seismic refraction methods. Especially in some hilly areas, due to the high terrain undulation, the variation of the stratigraphic structure is relatively high. For example, in some areas, long-term soil erosion has resulted in the exposure of red sandstone or the lack of a homogeneous red soil layer.

[0008] At present, the identification of stratigraphic structure in subtropical humid red soil areas based on shallow seismic refraction wave method has certain limitations in field operation process and seismic wave signal interpretation, including:

[0009] (1) Through the analysis of relevant technical reports, journals, patents and national standards at home and abroad, it is found that the existing technology rarely pays attention to the identification of shallow stratigraphic structure in subtropical humid red soil areas, especially when it comes to the different distribution characteristics of stratigraphic structure at different slope positions. The traditional drilling and profile excavation used to calibrate seismic signal interpretation is time-consuming and labor-intensive, and it is difficult to promote it on a large scale.

[0010] (2) The shallow seismic refraction wave method is mainly based on the propagation speed of artificial seismic waves in different strata to explore the structural stratification of the underground. The core link of the operation is to analyze the changes in the waveform, vibration intensity and propagation path of the seismic wave in different media. The contact between the homogeneous red soil layer and the reticular red soil layer in the red soil area of ​​my country usually presents a gradient feature, which may lead to the inability to fully invert the depth information of the strata based only on the velocity and travel time data of the seismic wave, or lead to high uncertainty in the interpretation results. In particular, the lack of stratigraphic structure in some areas may lead to the staff's signal interpretation experience in red soil areas cannot be accurately applied to similar red soil areas, especially in some areas where the red sandstone is directly exposed and the detector cannot be directly buried, resulting in the inability to obtain the velocity of the seismic wave.

[0011] (3) After interpretation, the profile obtained by the seismic refraction wave method only represents the distribution of the underlying stratigraphic structure in the vertical direction of one survey line. The several survey lines deployed in the operating area can only provide an understanding of the stratigraphic structure of these survey lines in the area, and the stratigraphic structure of the uninvestigated area is still not clear enough.

[0012] The above-mentioned deficiencies of the existing seismic refraction wave analysis technology have affected the field operation efficiency and indoor analysis accuracy of relevant application departments, brought losses to the national economy to a certain extent, and affected the economic benefits of my country's basic geological surveys. Summary of the invention

[0013] The technical problem to be solved by the present invention is to provide a method for identifying the stratigraphic structure of red soil areas based on the shallow seismic refraction wave method, innovatively combining geophysical exploration with spatial prediction, and making up for the technical deficiencies of traditional geological exploration technology.

[0014] In order to solve the above technical problems, the present invention adopts the following technical solutions: the present invention designs a method for identifying the stratigraphic structure of the red soil area based on the shallow seismic refraction wave method, which is used to realize the identification of the stratigraphic structure of the red soil area in the target operation area, and obtains the optimal prediction model corresponding to each stratigraphic structure depth in the target operation area through the following steps A to K; then, step i is executed to obtain the spatial distribution of each stratigraphic structure depth in the target operation area according to the optimal prediction model corresponding to each stratigraphic structure depth;

[0015] Step A. For each preset survey line in the target operation area, based on the shallow seismic refraction wave method, the horizontal distribution of different stratum structure depths below each survey line is obtained to form the seismic profile data corresponding to each survey line, and then proceed to step B;

[0016] Step B. For all the survey lines in the target operation area, press Randomly select m1 survey lines to form an independent validation set Prof_IndepValid, and the remaining survey lines to form a spatial modeling set Prof_SpatiModel, and then proceed to step C; wherein M represents the total number of survey lines in the target operation area, a represents the preset first ratio value, Indicates rounding up;

[0017] Step C. For each survey line in the spatial modeling set Prof_SpatiModel, based on the length of the survey line, according to the preset horizontal sampling distance DisH, and combined with the two end points of the survey line, construct each sample point located on the survey line; then obtain each sample point located on each survey line in the spatial modeling set Prof_SpatiModel, and then enter step D;

[0018] Step D. For each sample point on each survey line in the spatial modeling set Prof_SpatiModel, select the average value of the depth of each stratigraphic structure below the sample point within the distance range of b·DisH on both sides of the sample point along the survey line set for the sample point as the depth of each stratigraphic structure directly below the sample point; then obtain the depth of each stratigraphic structure directly below each sample point on each survey line in the spatial modeling set Prof_SpatiModel, and then enter step E; wherein b represents the preset second ratio value;

[0019] Step E. For each sample point on each survey line in the spatial modeling set Prof_SpatiModel, obtain the data value of each preset primary discrete environmental variable corresponding to the position of each sample point, and then apply the preset first analysis method to screen and obtain the primary discrete environmental variables associated with each stratigraphic structure depth as the discrete environmental variables associated with each stratigraphic structure depth;

[0020] At the same time, for each sample point on each survey line in the spatial modeling set Prof_SpatiModel, the data values ​​of each preset primary continuous environmental variable corresponding to each sample point position are obtained, and then the preset second analysis method is applied to screen and obtain the primary continuous environmental variables associated with each stratigraphic structure depth as the continuous environmental variables associated with each stratigraphic structure depth;

[0021] Then, the discrete environmental variables and continuous environmental variables respectively associated with each stratigraphic structure depth are obtained to form a set of available environmental variables respectively associated with each stratigraphic structure depth, and then the process goes to step F;

[0022] Step F. Divide all the sample points in the spatial modeling set Prof_SpatiModel into a training sample point set CaliPProf and a verification sample point set ValiPProf, wherein the number of sample points in the training sample point set CaliPProf is Nc, and the number of sample points in the verification sample point set ValiPProf is Nv, Nc>Nv, and then proceed to step G;

[0023] Step G. For each stratigraphic structure, based on each sample point in the training sample point set CaliPProf, the data value of each environmental variable in the available environmental variable set associated with the stratigraphic structure depth corresponding to the sample point is used as input, and the stratigraphic structure depth corresponding to the sample point is used as output. The preset models to be trained are trained respectively to obtain the prediction models corresponding to the stratigraphic structure depth, and then the prediction models corresponding to each stratigraphic structure depth are obtained, and then the process goes to step H;

[0024] Step H. For each stratum structure, based on each sample point in the verification sample point set ValiPProf, perform accuracy tests on each prediction model corresponding to the stratum structure depth, obtain the first temporary optimal prediction model corresponding to the stratum structure depth, and then obtain the first temporary optimal prediction model corresponding to each stratum structure depth, and then enter step I;

[0025] Step I: For the m1 measurement lines in the independent validation set Prof_IndepValid, perform equal-interval discretization and random sampling in turn to obtain a total of Ni sample points on each measurement line, and Ni = L m1 ×NcNt, Nt represents the number of sample points in the spatial modeling set Prof_SpatiModel, and then proceed to step J;

[0026] Step J. For Ni sample points in the independent validation set Prof_IndepValid, according to the method in step D, the depth of each stratigraphic structure directly below each sample point is obtained, and the data value of each environmental variable in the available environmental variable set associated with each stratigraphic structure depth corresponding to each sample point is obtained. Based on this, the accuracy of each stratigraphic structure and each prediction model corresponding to the stratigraphic structure depth is tested, and the second temporary optimal prediction model corresponding to the stratigraphic structure depth is obtained, and then the second temporary optimal prediction model corresponding to each stratigraphic structure depth is obtained, and then the process goes to step K;

[0027] Step K. for each stratum structure, determine whether the first temporary optimal prediction model and the second temporary optimal prediction model corresponding to the stratum structure depth are the same. If yes, the prediction model is the optimal prediction model corresponding to the stratum structure depth; otherwise, the first temporary optimal prediction model and the second temporary optimal prediction model are respectively assigned a weight of 0.5, and the optimal prediction model corresponding to the stratum structure depth is constructed by weighting; and then the optimal prediction models corresponding to each stratum structure depth are obtained;

[0028] Step i. For each stratigraphic structure, obtain the data value of each environmental variable in the set of available environmental variables associated with the stratigraphic structure depth corresponding to each position in the target operating area, apply the optimal prediction model corresponding to the stratigraphic structure depth, and obtain the stratigraphic structure depth corresponding to each position in the target operating area; then obtain the stratigraphic structure depth corresponding to each position in the target operating area, and constitute the spatial distribution of the stratigraphic structure depths in the target operating area.

[0029] As a preferred technical solution of the present invention: in the step E, the variance analysis method as the preset first analysis method is applied, and the following steps Ea-1 to Ea-2 are performed for each stratigraphic structure respectively to obtain each discrete environmental variable associated with each stratigraphic structure depth;

[0030] Step Ea-1. For each sample point on each survey line in the spatial modeling set Prof_SpatiModel, according to the depth of the stratum structure corresponding to each sample point position, and the data value of each preset primary discrete environmental variable corresponding to each sample point position, perform variance analysis to obtain the difference analysis results between the stratum structure depth and each primary discrete environmental variable, and then enter step Ea-2;

[0031] Step Ea-2. For each preliminarily selected discrete environmental variable, determine whether the difference analysis result between the preliminarily selected discrete environmental variable and the stratigraphic structure depth is less than a preset difference threshold; if so, determine that the stratigraphic structure depth is not affected by the preliminarily selected discrete environmental variable and do not perform any operation; otherwise, associate the preliminarily selected discrete environmental variable with the stratigraphic structure depth, that is, construct a discrete environmental variable associated with the stratigraphic structure depth; and then obtain each discrete environmental variable associated with the stratigraphic structure depth.

[0032] As a preferred technical solution of the present invention: in the step E, the correlation analysis method as the preset second analysis method is applied, and the following steps Eb-1 to Eb-2 are performed for each stratigraphic structure respectively to obtain each continuous environmental variable associated with each stratigraphic structure depth;

[0033] Step Eb-1. For each sample point on each survey line in the spatial modeling set Prof_SpatiModel, according to the depth of the stratum structure corresponding to each sample point position, and the data value of each preset primary continuous environmental variable corresponding to each sample point position, perform the Pearson correlation coefficient analysis method, according to the following formula:

[0034]

[0035] Obtain the Pearson correlation coefficient Pearson between the stratigraphic structure depth and each of the preliminary selected continuous environmental variables, and then proceed to step Eb-2; wherein PProf represents the stratigraphic structure depth corresponding to the sample point location, VPVari represents the data value of the preliminary selected continuous environmental variable corresponding to the sample point location, and Nt represents the number of sample points in the spatial modeling set Prof_SpatiModel;

[0036] Step Eb-2. For each preliminary continuous environmental variable, determine whether the Pearson correlation coefficient Pearson between the preliminary continuous environmental variable and the stratigraphic structure depth is less than a preset difference threshold. If so, determine that the stratigraphic structure depth is not affected by the preliminary continuous environmental variable and do not perform any operation; otherwise, associate the preliminary continuous environmental variable with the stratigraphic structure depth to form a continuous environmental variable associated with the stratigraphic structure depth; and then obtain each continuous environmental variable associated with the stratigraphic structure depth.

[0037] As a preferred technical solution of the present invention: the initially selected discrete environmental variables include soil parent material, land use, and hydrogeology; the initially selected continuous environmental variables include elevation, average annual sunshine time, average annual evapotranspiration, surface roughness, average annual rainfall, plane curvature, profile curvature, slope, slope aspect, terrain humidity index, average annual temperature, and average annual wind speed.

[0038] As a preferred technical solution of the present invention: it also includes step Fa as follows, after executing step F, entering step Fa;

[0039] Step Fa. For each stratigraphic structure respectively, according to the correlation between environmental variables, use the stepwise linear regression analysis method to screen the environmental variables in the set of available environmental variables associated with the stratigraphic structure depth, and update the set of available environmental variables associated with the stratigraphic structure depth; then update the set of available environmental variables associated with each stratigraphic structure depth, and then enter step G.

[0040] As a preferred technical solution of the present invention: in the step G, each model to be trained is preset as a random forest model, a multivariate linear regression model, a geographically weighted regression model, a support vector machine model, and an artificial neural network.

[0041] As a preferred technical solution of the present invention: in the steps H and J, the root mean square error is used as an accuracy test indicator to perform accuracy tests on each prediction model corresponding to the depth of the formation structure.

[0042] As a preferred technical solution of the present invention, each preset survey line in the target operation area in step A is constructed according to the following method 1, method 2, or a combination of method 1 and method 2:

[0043] Method 1: Based on the preset preliminary survey lines in the target operation area, according to the accessibility of the roads in the target operation area to the survey lines, the preliminary survey lines that are not accessible by roads are deleted, and the preliminary survey lines that are accessible by roads are retained, that is, the preset survey lines in the target operation area are formed;

[0044] Method 2: Based on the grid division of the target operation area and the road distribution in the target operation, survey lines are arranged at the position closest to the surrounding roads in each grid, that is, the preset survey lines in the target operation area are formed.

[0045] As a preferred technical solution of the present invention, in step A, for each preset survey line in the target operation area, hardware operation is performed in the following manner to realize the application of shallow seismic refraction wave method;

[0046] Based on the detectors arranged in sequence along the survey line, a hammer point is set at the middle detector position, a hammer point is set at the two end detector positions, and a hammer point is set at a preset distance outside the two end detector positions. An artificial hammer source is operated at each hammer point, and the signal is detected by each detector and uploaded to the seismograph equipment. The seismograph equipment analyzes and processes the collected signals to obtain the horizontal distribution of different stratum structure depths below the survey line, thereby forming the seismic profile data corresponding to each survey line.

[0047] As a preferred technical solution of the present invention, the seismograph equipment sequentially performs head placement, static correction, pre-stack noise suppression, amplitude compensation, pre-stack deconvolution, CMP gather extraction, velocity analysis, dynamic correction and initial stacking, residual static correction, DMO-pre-stack time migration, post-stack time migration processing, extraction of first arrival wave, creation of initial velocity model, and feedback for the collected signals from the detector, thereby obtaining the horizontal distribution of different stratigraphic structure depths below the survey line, and then forming the seismic profile data corresponding to each survey line.

[0048] The method for identifying the stratum structure in red soil areas based on the shallow seismic refraction wave method described in the present invention has the following technical effects compared with the prior art by using the above technical solution:

[0049] (1) The present invention designs a method for identifying the stratigraphic structure of red soil areas based on the shallow seismic refraction wave method, which can ensure the acquisition of seismic wave signals in standardized red soil areas, which is conducive to improving the acquisition accuracy and resolution of seismic signals. In addition, the survey line layout technology based on typical terrain can greatly reduce the number of survey line layouts, thereby improving the survey time and economic cost of the operating area.

[0050] (2) The fluctuation or depth change of the stratigraphic structure is of great significance to soil erosion, agriculture and industrial management in local areas. The typical survey line layout and typical sample point screening method proposed in the design of the present invention can calculate the correlation between the stratigraphic structure depth and elevation in the target operation area, and can effectively predict the future changes at different levels, which is of great significance for engineering safety precautions.

[0051] (3) The depth of the stratigraphic structure in red soil areas has been significantly affected by other environmental factors for a long time. The stratigraphic structure depth prediction method based on environmental variables proposed in the design of the present invention can make full use of the correlation between the stratigraphic structure depth and environmental variables, directly improve the accuracy of the compilation of the stratigraphic structure depth spatial distribution map, directly serve the monitoring and investigation of the regional stratigraphic structure, and also provide technical reference for the ecological models and surface process simulations of different regions.

[0052] (4) The method for identifying the stratigraphic structure in red soil areas based on the shallow seismic refraction wave method designed in the present invention fully retains the advantages of existing seismic exploration technology, improves the shortcomings of the limited application of the shallow seismic refraction wave method in red soil areas, and organically combines the geographic information system with spatial prediction, thus realizing the universal shallow seismic refraction wave method technology of "traditional technology, improved operation; spatial prediction, improved efficiency", which has broad industrial application prospects in geological exploration, hydrogeological survey and other departments. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a flow chart of the steps of the method for identifying the stratum structure in red soil area based on the shallow seismic refraction wave method designed by the present invention;

[0054] Figure 2 It is a schematic diagram of the field survey line arrangement based on the artificially set sampling method in the present invention;

[0055] Figure 3 It is a schematic diagram of the field survey line arrangement based on the grid sampling method in the present invention;

[0056] Figure 4 It is a schematic diagram of the seismic exploration method of unilateral pursuit excitation and multiple coverage in the present invention;

[0057] Figure 5 This is the standard processing process of seismic data in the present invention;

[0058] Figure 6 It is a schematic diagram of the horizontal distribution of different stratigraphic structural depths in the interpretation results of the present invention;

[0059] Figure 7 It is a schematic diagram of sample points including different stratigraphic structure depths obtained from the interpretation results in the present invention;

[0060] Figure 8 The interpreted depth map and drilling information of the stratigraphic structure in the implementation case of the present invention;

[0061] Fig. 9 is a spatial distribution diagram of the depth of the first layer of stratum structure in the implementation case of the present invention;

[0062] Fig.10 is a spatial distribution diagram of the depth of the second layer of stratum structure in the implementation case of the present invention;

[0063] Fig.11 It is a spatial distribution diagram of the depth of the third layer of stratum structure in the implementation case of the present invention. DETAILED DESCRIPTION

[0064] The specific implementation modes of the present invention will be further described in detail below in conjunction with the accompanying drawings.

[0065] The present invention designs a method for identifying the stratigraphic structure in red soil areas based on the shallow seismic refraction wave method. This technology is aimed at subtropical humid red soil areas, especially the stratigraphic structure identification involving reticular layers and red sandstones, and then combines geographic information system technology with spatial prediction technology to perform spatial prediction of the stratigraphic structure depth.

[0066] The basic idea of ​​the present invention is to lay out a certain number of survey lines according to the terrain undulations based on a certain understanding of the operating area. After analyzing the seismic signals covering different stratigraphic structures, a standardized method is used to interpret the seismic longitudinal wave velocities at different levels. Based on the spatial variation characteristics of the terrain and the depth of each stratigraphic structure, the correlation analysis method is used to set the number of sample points to be selected from the survey line. In combination with the geographic information system, the set of environmental variables at the locations of different sample points is extracted according to the longitude and latitude of each sample point. A variety of spatial prediction models are trained, and the Holdout verification and independent verification mechanisms are used to evaluate the accuracy of the constructed prediction models to generate a spatial distribution map of the stratigraphic structure of the operating area.

[0067] Therefore, the present invention designs a method for identifying the stratigraphic structure of red soil areas based on shallow seismic refraction wave method, which is used to realize the identification of the stratigraphic structure of red soil areas in the target operation area. In specific practical applications, such as Figure 1 As shown, through the following steps A to K, the optimal prediction model corresponding to each stratigraphic structure depth in the target operation area is obtained.

[0068] Step A. For each preset survey line in the target operation area, based on the shallow seismic refraction wave method, obtain the horizontal distribution of different stratum structure depths below each survey line, form the seismic profile data corresponding to each survey line, and then enter step B.

[0069] In practical applications, the arrangement of survey lines can be simple random sampling, systematic sampling, and stratified random sampling. These methods are mainly influenced by the probability statistics theory in traditional sampling design, assuming that the spatial variation characteristics of attributes can be expressed and measured in a random manner. Therefore, given that users have different experiences in field sampling, the preset survey lines in the target operation area in step A above are constructed according to the following method 1, method 2, or a combination of method 1 and method 2.

[0070] Method 1: Based on the preset preliminary survey lines in the target operation area, according to the accessibility of the roads in the target operation area to the survey lines, the preliminary survey lines that are not accessible by roads are deleted, and the preliminary survey lines that are accessible by roads are retained, that is, the preset survey lines in the target operation area are formed, such as Figure 2 shown.

[0071] Method 1 here can often give full play to the work experience and historical data of operators, and can obtain more seismic data at a lower cost. However, this method often leads to a denser spatial distribution of survey line positions, with fewer or missing survey lines in some areas.

[0072] Method 2: Based on the grid division of the target operation area and the road distribution in the target operation, the survey lines are arranged at the position closest to the surrounding roads in each grid, that is, the preset survey lines in the target operation area are formed, such as Figure 3 shown.

[0073] The advantage of method 2 here is that it can achieve full coverage of the survey line in the operating area. The disadvantage is that the location of some survey lines may be far away from the road and may not be reached. It is also possible that a square in the operating area is urban land, a factory or a lake, making it impossible to collect survey lines.

[0074] And in the above step A, for each preset survey line in the target operation area, the hardware operation is performed in the following manner to realize the application of the shallow seismic refraction wave method.

[0075] Based on the geophones arranged in sequence along the survey line, a hammer point is set at the middle geophone position, a hammer point is set at each of the two end geophones, and a hammer point is set at a preset distance outside the two end geophones. Each hammer point is operated in an artificial hammer source manner, and the detection is uploaded to the seismograph equipment through each geophone. The seismograph equipment analyzes and processes the collected signals from the geophones to obtain the horizontal distribution of different stratum structure depths below the survey line, thereby forming the seismic profile data corresponding to each survey line.

[0076] The operational actions executed by the above design are implemented according to the following process.

[0077] (1.a) Before conducting seismic exploration in the operating area, prepare the seismic signal source. Explosive seismic sources are unsafe and inconvenient to use. The stratum structure in the red soil area is shallow, and the artificial hammer seismic source is sufficient to meet business needs. In view of the small background of regional noise interference, a 10-20kg copper hammer hitting a steel plate can be considered as a signal source. The size of the steel plate is generally 20-40cm wide and can be round or rectangular in shape.

[0078] (1.b) Other materials: Global Positioning System (to record the latitude and longitude information of the starting and ending points of the survey line), tape measure (the length needs to be greater than the survey line length) and notebook (to record the working environment, survey line length, soil moisture conditions and weather, etc.).

[0079] (1.c) Check whether the seismic equipment is complete and ensure that the battery power of the equipment is sufficient to meet the 8-hour / day operation requirements. According to the survey line location layout in the first step, use the appropriate transportation to transport the seismograph equipment to the designated location.

[0080] (1.d) Install relevant seismic exploration equipment and use single-side chasing excitation and multiple coverage methods, such as Figure 4 As shown in the figure, the excitation point layout uses 5 source points for excitation: one hammer point is set for the geophone in the middle of the array, one hammer point is set for each end of the array, and one hammer point is set for each end at a certain distance (maximum offset) from the array. In order to enhance the signal-to-noise ratio and improve the accuracy of seismic exploration, three repetitions are set for each hammer point.

[0081] Operators should select detectors of corresponding frequencies according to the depth of the geological structure in the operating area. If the geological structure in the operating area is shallow, a 40Hz detector can be used.

[0082] Note: In order to achieve the best observation effect, the specific survey line location selected in the field should cover typical land use, such as dry land, paddy field, bare land and forest land. Because the energy of artificial seismic sources is limited, in order to ensure accurate signal recognition and improve the signal-to-noise ratio of seismic signals, the area where the survey line is operated should avoid high noise. Unless there are special circumstances, try not to operate in rainy and snowy weather conditions.

[0083] (1.e) In particular, in order to improve the accuracy of stratigraphic structure identification, operators should try to find a detection environment with exposed rock in or near the operating area, that is, the rock on the surface is the lowest layer of the stratigraphic structure in the area due to natural or human reasons. The purpose of arranging the bare rock survey line is to calculate the seismic wave velocity of the bedrock in the operating area.

[0084] (1.f) After the operation is completed, carefully count the instruments and accessories to avoid loss of equipment. Check the power supply to ensure that there is enough power for the next use.

[0085] (1.g) Several wells are deployed in several survey lines in the operation area to record the changes in the formation structure at different depths of the wells. In order to achieve better result analysis, the wells are deployed as close as possible to the locations where the seismic survey lines pass.

[0086] Setting the empirical value of seismic wave longitudinal wave propagation velocity. Strata structures often have different seismic wave propagation velocities and wave impedances. Therefore, based on the differences in the propagation laws of seismic waves in stratigraphic materials, the properties of the rock layers (such as rock composition, density, burial depth, porosity, etc.) can be comprehensively reflected. The empirical values ​​of seismic wave propagation velocities of different stratigraphic structures in my country's subtropical humid red soil region are shown in Table 1.

[0087] Table 1

[0088]

[0089]

[0090] If the field seismic data are collected correctly, these digital data should contain information about the stratigraphic structure. Due to the superposition of seismic signals, interference from other signals and inconsistent strength of artificial seismic sources, there may be some interweaving between seismic signals. The original field seismic signals need to be processed by standard procedures, such as Figure 5 In view of the shallow burial depth of the stratigraphic structure in the red soil region of my country, two-dimensional filtering technology is used in the indoor processing of seismic wave signals. It is necessary to generate time-distance curves first and obtain the velocity diagrams of various stratigraphic sections through time-depth conversion.

[0091] Along with the above operations, the detection is performed by each detector and uploaded to the seismograph equipment. The seismograph equipment sequentially performs channel header placement, static correction, pre-stack noise suppression, amplitude compensation, pre-stack deconvolution, CMP gather extraction, velocity analysis, dynamic correction and initial stacking, residual static correction, DMO-pre-stack time migration, post-stack time migration processing, extraction of first arrival wave, creation of initial velocity model, feedback, etc. on the collected signals from the detector to realize analysis and processing, and obtain the horizontal distribution of different stratigraphic structure depths below the survey line, thereby forming the seismic profile data corresponding to each survey line.

[0092] Among them, static correction: eliminates the interference of seismic data caused by terrain changes in the operating area. This operation controls the vertical resolution and signal-to-noise ratio of the stacked section to a certain extent, affecting the quality of stacking velocity analysis. The main static correction methods include first-arrival refraction wave method, micro-logging static correction and elevation static correction.

[0093] Prestack noise suppression: suppress and attenuate noise caused by multiple factors such as external conditions, construction factors or instruments. Noise suppression methods include coherence enhancement, radial filtering, polynomial fitting and random noise attenuation.

[0094] Amplitude compensation: To enhance the reflected signal of seismic waves that weakens as the propagation distance increases in different strata structures, methods such as spherical diffusion and earth absorption compensation, and surface consistency amplitude compensation can be used.

[0095] Velocity is one of the most important parameters in seismic exploration. Velocity analysis can fit the propagation velocity of seismic waves. Furthermore, using layer velocity data or the vertical and horizontal velocity variation patterns, it is possible to study the sedimentary characteristics of the strata or predict the distribution of lithology. Users can perform velocity analysis through human-computer interaction, that is, using the computer graphics display function and combined with operating experience to determine the optimal stacking velocity.

[0096] Note: Velocity analysis and residual statics are usually an iterative process that needs to be performed multiple times to determine whether the residual statics are less than 0.5 ms. Only when the residual statics are less than 0.5 ms can DMO or prestack time migration operations be performed.

[0097] Dynamic Correction: In order to eliminate the influence of normal time difference, the reflection event axis based on one point reflection information is flattened, which can provide superposition input data for the common center point.

[0098] According to the fitting of the propagation velocity of longitudinal waves on the exposed rock surface, the propagation velocity of seismic waves on the bedrock surface of the operation area and the propagation velocity of the interface between the overburden and the semi-weathered layer are determined, and all the collected seismic data are interpreted. Assuming that the number of survey lines in the actual field survey is M, the interpreted seismic profile data set is Prof = {L1, L2, ...LM}, Li is a one-dimensional vector, which contains the horizontal distribution of different stratigraphic structural depths in the survey line i and the longitude and latitude of the starting and ending points of the survey line, such as Figure 6 shown.

[0099] Step B. For all the survey lines in the target operation area, press Randomly select m1 survey lines to form an independent validation set Prof_IndepValid, and the remaining survey lines to form a spatial modeling set Prof_SpatiModel, and then proceed to step C; wherein M represents the total number of survey lines in the target operation area, a represents the preset first ratio value, Indicates rounding up. Each profile in the independent validation set Prof_IndepValid and the spatial modeling set Prof_SpatiModel contains information such as the horizontal distribution of different stratum structure depths and the longitude and latitude of the starting and ending points of the survey line. Figure 6 shown.

[0100] Step C. For each survey line in the spatial modeling set Prof_SpatiModel, based on the length of the survey line, according to the preset horizontal sampling distance DisH, and combined with the two end points of the survey line, construct each sample point located on the survey line; and then obtain each sample point located on each survey line in the spatial modeling set Prof_SpatiModel, such as Figure 7 As shown, then go to step D. In practical applications, for all the survey lines in the spatial modeling set Prof_SpatiModel, the preset horizontal sampling distance DisH can be set according to the minimum number of all sample points required, thereby ensuring that the total number of all sample points finally obtained is the minimum number of all sample points required.

[0101] Step D. For each sample point on each survey line in the spatial modeling set Prof_SpatiModel, select the average value of the depth of each stratigraphic structure below the sample point along the survey line set for the sample point and within the distance range of b·DisH on both sides of the sample point as the depth of each stratigraphic structure directly below the sample point; then obtain the depth of each stratigraphic structure directly below each sample point on each survey line in the spatial modeling set Prof_SpatiModel, and then enter step E; wherein b represents a preset second ratio value. In actual application, for example, b=0.5 is selected, that is, for each sample point, the average value of the depth of each stratigraphic structure below the sample point along the survey line set for the sample point and within the distance range of 0.5·DisH on both sides of the sample point is selected as the depth of each stratigraphic structure directly below the sample point.

[0102] Step E. For each sample point on each survey line in the spatial modeling set Prof_SpatiModel, obtain the data value of each preset primary discrete environmental variable corresponding to the location of each sample point, and then apply the preset first analysis method to screen and obtain the primary discrete environmental variables associated with each stratigraphic structure depth as the discrete environmental variables associated with each stratigraphic structure depth.

[0103] In specific applications, the initially selected discrete environmental variables include parent material, land use, and hydrogeology; the variance analysis method is specifically applied as the preset first analysis method, and the following steps Ea-1 to Ea-2 are executed for each stratigraphic structure to obtain the discrete environmental variables associated with each stratigraphic structure depth.

[0104] Step Ea-1. For each sample point on each survey line in the spatial modeling set Prof_SpatiModel, perform variance analysis according to the depth of the stratigraphic structure corresponding to each sample point location and the data values ​​of each preset preliminary discrete environmental variable corresponding to each sample point location, and obtain the difference analysis results between the stratigraphic structure depth and each preliminary discrete environmental variable, and then enter step Ea-2.

[0105] Step Ea-2. For each preliminarily selected discrete environmental variable, determine whether the difference analysis result between the preliminarily selected discrete environmental variable and the stratigraphic structure depth is less than a preset difference threshold; if so, determine that the stratigraphic structure depth is not affected by the preliminarily selected discrete environmental variable and do not perform any operation; otherwise, associate the preliminarily selected discrete environmental variable with the stratigraphic structure depth, that is, construct a discrete environmental variable associated with the stratigraphic structure depth; and then obtain each discrete environmental variable associated with the stratigraphic structure depth.

[0106] At the same time, for each sample point on each survey line in the spatial modeling set Prof_SpatiModel, the data values ​​of the preset preliminary continuous environmental variables corresponding to the location of each sample point are obtained, and then the preset second analysis method is applied to screen and obtain the preliminary continuous environmental variables associated with each stratigraphic structure depth as the continuous environmental variables associated with each stratigraphic structure depth.

[0107] In practical applications, the initial selection of continuous environmental variables includes elevation, annual average sunshine time, annual average evapotranspiration, surface roughness, annual average rainfall, plane curvature, profile curvature, slope, slope aspect, terrain humidity index, annual average temperature, and annual average wind speed. Specifically, the correlation analysis method as the preset second analysis method is used to perform the following steps Eb-1 to Eb-2 for each stratum structure to obtain each continuous environmental variable associated with each stratum structure depth.

[0108] Step Eb-1. For each sample point on each survey line in the spatial modeling set Prof_SpatiModel, according to the depth of the stratum structure corresponding to each sample point position, and the data value of each preset primary continuous environmental variable corresponding to each sample point position, perform the Pearson correlation coefficient analysis method, according to the following formula:

[0109]

[0110] Obtain the Pearson correlation coefficient Pearson between the stratigraphic structure depth and each preliminary continuous environmental variable, and then enter step Eb-2; wherein PProf represents the stratigraphic structure depth corresponding to the sample point location, VPVari represents the data value of the preliminary continuous environmental variable corresponding to the sample point location, and Nt represents the number of sample points in the spatial modeling set Prof_SpatiModel.

[0111] Step Eb-2. For each preliminary continuous environmental variable, determine whether the Pearson correlation coefficient Pearson between the preliminary continuous environmental variable and the stratigraphic structure depth is less than a preset difference threshold. If so, determine that the stratigraphic structure depth is not affected by the preliminary continuous environmental variable and do not perform any operation; otherwise, associate the preliminary continuous environmental variable with the stratigraphic structure depth to form a continuous environmental variable associated with the stratigraphic structure depth; and then obtain each continuous environmental variable associated with the stratigraphic structure depth.

[0112] Then, the discrete environmental variables and continuous environmental variables respectively associated with each stratigraphic structure depth are obtained to form a set of available environmental variables respectively associated with each stratigraphic structure depth, and then the process goes to step F.

[0113] Step F. All the sample points in the spatial modeling set Prof_SpatiModel are divided into a training sample point set CaliPProf and a verification sample point set ValiPProf, wherein the number of sample points in the training sample point set CaliPProf is Nc, and the number of sample points in the verification sample point set ValiPProf is Nv, Nc>Nv, and then enter step Fa.

[0114] Step Fa. For each stratigraphic structure respectively, according to the correlation between environmental variables, use a stepwise linear regression analysis method such as the forward method or the backward method to screen the environmental variables in the set of available environmental variables associated with the stratigraphic structure depth, and update the set of available environmental variables associated with the stratigraphic structure depth; then update the sets of available environmental variables associated with each stratigraphic structure depth, and then enter step G.

[0115] Step G. For each stratigraphic structure, based on each sample point in the training sample point set CaliPProf, with the data value of each environmental variable in the available environmental variable set associated with the stratigraphic structure depth corresponding to the sample point as input, and the stratigraphic structure depth corresponding to the sample point as output, training is performed for each preset model to be trained, and the prediction models corresponding to the stratigraphic structure depth are obtained, and then the prediction models corresponding to each stratigraphic structure depth are obtained, and then step H is entered.

[0116] In practical applications, the preset models to be trained include random forest model, multivariate linear regression model, geographically weighted regression model, support vector machine model, and artificial neural network. Several linear regression and machine learning models with good prediction performance and high portability are provided here. Users are not limited to these prediction methods and can customize prediction models according to their personal business processing capabilities and the availability of methods. However, the construction of different prediction models requires the screening of the optimal set of environmental variables to ensure the robustness of the prediction models.

[0117] Step H. For each stratigraphic structure, based on each sample point in the verification sample point set ValiPProf, apply the root mean square error as the accuracy test indicator, and perform accuracy test on each prediction model corresponding to the stratigraphic structure depth, and obtain the first temporary optimal prediction model corresponding to the stratigraphic structure depth, and then obtain the first temporary optimal prediction model corresponding to each stratigraphic structure depth, and then enter step I.

[0118] Step I: For the m1 measurement lines in the independent validation set Prof_IndepValid, perform equal-interval discretization and random sampling in turn to obtain a total of Ni sample points on each measurement line, and Ni = L m1 ×Nc / Nt, Nt represents the number of sample points in the spatial modeling set Prof_SpatiModel, and then proceed to step J.

[0119] Step J. For the Ni sample points in the independent validation set Prof_IndepValid, according to the method in step D, obtain the depth of each stratigraphic structure directly below each sample point, and obtain the data value of each environmental variable in the available environmental variable set associated with each stratigraphic structure depth corresponding to each sample point. Based on this, for each stratigraphic structure, apply the root mean square error as the accuracy test indicator, perform accuracy test on each prediction model corresponding to the stratigraphic structure depth, obtain the second temporary optimal prediction model corresponding to the stratigraphic structure depth, and then obtain the second temporary optimal prediction model corresponding to each stratigraphic structure depth, and then enter step K.

[0120] Step K. For each stratigraphic structure, determine whether the first temporary optimal prediction model and the second temporary optimal prediction model corresponding to the stratigraphic structure depth are the same. If so, the prediction model is the optimal prediction model corresponding to the stratigraphic structure depth; otherwise, the first temporary optimal prediction model and the second temporary optimal prediction model are respectively assigned a weight of 0.5, and the optimal prediction model corresponding to the stratigraphic structure depth is constructed through a weighted method; thereby obtaining the optimal prediction models corresponding to each stratigraphic structure depth.

[0121] Based on the optimal prediction models corresponding to the depths of the stratigraphic structures in the target operating area obtained above, step i is then performed to obtain the spatial distribution of the depths of the stratigraphic structures in the target operating area according to the optimal prediction models corresponding to the depths of the stratigraphic structures.

[0122] Step i. For each stratigraphic structure, obtain the data value of each environmental variable in the set of available environmental variables associated with the stratigraphic structure depth corresponding to each position in the target operating area, apply the optimal prediction model corresponding to the stratigraphic structure depth, and obtain the stratigraphic structure depth corresponding to each position in the target operating area; then obtain the stratigraphic structure depth corresponding to each position in the target operating area, and constitute the spatial distribution of the stratigraphic structure depths in the target operating area.

[0123] The method for identifying the stratigraphic structure in red soil areas based on the shallow seismic refraction wave method designed by the present invention is applied in practice. Taking the spatial prediction of the stratigraphic structure depth in the operating area of ​​Yingtan City, Jiangxi Province as an example, a small watershed in Yingtan City, Jiangxi Province is taken as the research area. The area of ​​the area is 15.4ha, the altitude is 42-52m, the slope is 0.5°-4.1°, the soil texture is heavy, and the soil parent material is Quaternary red clay and red sandstone. The small watershed is located in the subtropical monsoon climate zone, with abundant rainfall and an annual precipitation of about 1789mm. Due to the frequent rainfall in the rainy season, soil and water loss caused by rainfall is very serious. The stratigraphic structure mainly includes three layers, which are homogeneous red soil layer, reticulated red soil layer and red sandstone from top to bottom. The land use in this area is mainly agricultural land. In order to explore the impact of the terrain in the operating area on soil erosion and the depth of the reticulated red soil, on the basis of the present invention, the shallow seismic refraction wave method is used to analyze the depth of different stratigraphic structures in the operating area, and then a spatial distribution map of the stratigraphic structure depth in the operating area is compiled.

[0124] In the application, combined with the size of the working area and the undulation of the terrain, it is necessary to collect 6 survey lines in the working area to understand the spatial distribution of the stratigraphic structure depth in the area. Specifically, through online maps and field surveys, one survey line is set in the paddy field, two survey lines are set in the forest and dry land respectively, and one survey line is set in the abandoned land.

[0125] Due to the superposition of seismic signals, the interference of other signals and the inconsistent strength of artificial seismic sources, there may be some interweaving between seismic signals. The original field seismic signals are processed using standard procedures, such as Figure 5 As shown in the figure, the operation area belongs to the red soil area and the stratum structure is shallow, so the two-dimensional filtering technology is used in the seismic wave signal processing process.

[0126] Based on the fitting of the propagation velocity of longitudinal waves on the exposed rock surface, the seismic wave propagation velocity of the bedrock surface in the operating area and the propagation velocity of the interface between the overburden and semi-weathered layer are determined, and all the collected seismic data are interpreted.

[0127] According to the interpretation results of the survey line, the seismic profile velocity map can be divided into three layers: low-speed layer (350-1000m / s), medium-speed layer (1000-2400m / s) and high-speed layer (>2400m / s), which correspond to the Quaternary red clay layer (first layer), semi-weathered layer (second layer) and bedrock layer (third layer) in the operating area. From the ground downward, the greater the depth from the surface, the greater the seismic wave velocity. The seismic wave interpretation profile is as follows: Figure 8 shown.

[0128] Six survey lines were actually investigated in the field, and the interpreted seismic profile data set is Prof = {L1, L2, ... L6}, where Li contains the horizontal distribution of different stratigraphic structural depths in survey line i and the longitude and latitude of the starting and ending points of the survey line.

[0129] Then, the above-designed method for identifying the stratigraphic structure in red soil areas based on the shallow seismic refraction wave method is executed, wherein a spatial distribution map of the depth of the first stratigraphic structure is generated, such as Fig. 9 As shown in the figure, the spatial distribution maps of the second and third layer stratigraphic structure depths are generated, as shown in the figure. Fig.10 and Fig.11 shown.

[0130] The above-mentioned technical solution designs a method for identifying the stratigraphic structure in red soil areas based on the shallow seismic refraction wave method, which can ensure the collection of seismic wave signals in standardized red soil areas, and is conducive to improving the acquisition accuracy and resolution of seismic signals. The survey line layout technology based on typical terrain can greatly reduce the number of survey line layouts, thereby improving the investigation time and economic cost of the operating area.

[0131] In applications, the fluctuation or depth change of the stratigraphic structure is of great significance to soil erosion, agriculture and industrial management in local areas. The typical survey line arrangement and typical sample point screening method proposed in the design of the present invention can calculate the correlation between the stratigraphic structure depth and elevation in the target operation area, and can effectively predict the future changes at different levels, which is of great significance for engineering safety precautions; and the depth of the stratigraphic structure in red soil areas has been significantly affected by other environmental factors for a long time. The stratigraphic structure depth prediction method based on environmental variables proposed in the design of the present invention can make full use of the correlation between the stratigraphic structure depth and environmental variables, directly improve the compilation accuracy of the stratigraphic structure depth spatial distribution map, can directly serve the monitoring and investigation of regional stratigraphic structures, and can also provide technical references for ecological models and surface process simulations in different regions.

[0132] The stratigraphic structure identification method for red soil areas based on the shallow seismic refraction wave method designed in the present invention fully retains the advantages of existing seismic exploration technology, improves the shortcomings of the limited application of the shallow seismic refraction wave method in red soil areas, and organically combines the geographic information system with spatial prediction, realizing the universal shallow seismic refraction wave method technology of "traditional technology, improved operation; spatial prediction, improved efficiency", and has broad industrial application prospects in geological exploration, hydrogeological survey and other departments.

[0133] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge scope of ordinary technicians in this field without departing from the purpose of the present invention.

Claims

1. A method for identifying the stratigraphic structure of red soil areas based on shallow seismic refraction wave method, which is used to identify the stratigraphic structure of red soil areas in target operating areas, characterized in that: Through the following steps A to K, the optimal prediction models corresponding to the depths of the stratigraphic structures in the target operation area are obtained; then step i is performed to obtain the spatial distribution of the depths of the stratigraphic structures in the target operation area according to the optimal prediction models corresponding to the depths of the stratigraphic structures; Step A. For each preset survey line in the target operation area, based on the shallow seismic refraction wave method, the horizontal distribution of different stratum structure depths below each survey line is obtained to form the seismic profile data corresponding to each survey line, and then proceed to step B; Step B. For all the survey lines in the target operation area, press Randomly select m1 survey lines to form an independent validation set Prof_IndepValid, and the remaining survey lines to form a spatial modeling set Prof_SpatiModel, and then proceed to step C; wherein M represents the total number of survey lines in the target operation area, a represents the preset first ratio value, Indicates rounding up; Step C. For each survey line in the spatial modeling set Prof_SpatiModel, based on the length of the survey line, according to the preset horizontal sampling distance DisH, and combined with the two end points of the survey line, construct each sample point located on the survey line; then obtain each sample point located on each survey line in the spatial modeling set Prof_SpatiModel, and then enter step D; Step D. For each sample point on each survey line in the spatial modeling set Prof_SpatiModel, select the average value of the depth of each stratigraphic structure below the sample point within the distance range of b·DisH on both sides of the sample point along the survey line set for the sample point as the depth of each stratigraphic structure directly below the sample point; then obtain the depth of each stratigraphic structure directly below each sample point on each survey line in the spatial modeling set Prof_SpatiModel, and then enter step E; wherein b represents the preset second ratio value; Step E. For each sample point on each survey line in the spatial modeling set Prof_SpatiModel, obtain the data value of each preset primary discrete environmental variable corresponding to the position of each sample point, and then apply the preset first analysis method to screen and obtain the primary discrete environmental variables associated with each stratigraphic structure depth as the discrete environmental variables associated with each stratigraphic structure depth; At the same time, for each sample point on each survey line in the spatial modeling set Prof_SpatiModel, the data values ​​of each preset primary continuous environmental variable corresponding to each sample point position are obtained, and then the preset second analysis method is applied to screen and obtain the primary continuous environmental variables associated with each stratigraphic structure depth as the continuous environmental variables associated with each stratigraphic structure depth; Then, the discrete environmental variables and continuous environmental variables respectively associated with each stratigraphic structure depth are obtained to form a set of available environmental variables respectively associated with each stratigraphic structure depth, and then the process goes to step F; Step F. Divide all the sample points in the spatial modeling set Prof_SpatiModel into a training sample point set CaliPProf and a verification sample point set ValiPProf, wherein the number of sample points in the training sample point set CaliPProf is Nc, and the number of sample points in the verification sample point set ValiPProf is Nv, Nc>Nv, and then proceed to step G; Step G. For each stratigraphic structure, based on each sample point in the training sample point set CaliPProf, the data value of each environmental variable in the available environmental variable set associated with the stratigraphic structure depth corresponding to the sample point is used as input, and the stratigraphic structure depth corresponding to the sample point is used as output. The preset models to be trained are trained respectively to obtain the prediction models corresponding to the stratigraphic structure depth, and then the prediction models corresponding to each stratigraphic structure depth are obtained, and then the process goes to step H; Step H. For each stratum structure, based on each sample point in the verification sample point set ValiPProf, perform accuracy tests on each prediction model corresponding to the stratum structure depth, obtain the first temporary optimal prediction model corresponding to the stratum structure depth, and then obtain the first temporary optimal prediction model corresponding to each stratum structure depth, and then enter step I; Step I: For the m1 measurement lines in the independent validation set Prof_IndepValid, perform equal-interval discretization and random sampling in turn to obtain a total of Ni sample points on each measurement line, and Ni = L m1 ×Nc / Nt, Nt represents the number of sample points in the spatial modeling set Prof_SpatiModel, and then proceed to step J; Step J. For Ni sample points in the independent validation set Prof_IndepValid, according to the method in step D, the depth of each stratigraphic structure directly below each sample point is obtained, and the data value of each environmental variable in the available environmental variable set associated with each stratigraphic structure depth corresponding to each sample point is obtained. Based on this, the accuracy of each stratigraphic structure and each prediction model corresponding to the stratigraphic structure depth is tested, and the second temporary optimal prediction model corresponding to the stratigraphic structure depth is obtained, and then the second temporary optimal prediction model corresponding to each stratigraphic structure depth is obtained, and then the process goes to step K; Step K. for each stratum structure, determine whether the first temporary optimal prediction model and the second temporary optimal prediction model corresponding to the stratum structure depth are the same. If yes, the prediction model is the optimal prediction model corresponding to the stratum structure depth; otherwise, the first temporary optimal prediction model and the second temporary optimal prediction model are respectively assigned a weight of 0.5, and the optimal prediction model corresponding to the stratum structure depth is constructed by weighting; and then the optimal prediction models corresponding to each stratum structure depth are obtained; Step i. For each stratigraphic structure, obtain the data value of each environmental variable in the set of available environmental variables associated with the stratigraphic structure depth corresponding to each position in the target operating area, apply the optimal prediction model corresponding to the stratigraphic structure depth, and obtain the stratigraphic structure depth corresponding to each position in the target operating area; then obtain the stratigraphic structure depth corresponding to each position in the target operating area, and constitute the spatial distribution of the stratigraphic structure depths in the target operating area.

2. The method for identifying the stratigraphic structure in red soil areas based on shallow seismic refraction wave method according to claim 1 is characterized by: In the step E, the variance analysis method as the preset first analysis method is applied, and the following steps Ea-1 to Ea-2 are performed for each stratigraphic structure respectively to obtain each discrete environmental variable associated with each stratigraphic structure depth; Step Ea-1. For each sample point on each survey line in the spatial modeling set Prof_SpatiModel, according to the depth of the stratum structure corresponding to each sample point position, and the data value of each preset primary discrete environmental variable corresponding to each sample point position, perform variance analysis to obtain the difference analysis results between the stratum structure depth and each primary discrete environmental variable, and then enter step Ea-2; Step Ea-2. For each preliminarily selected discrete environmental variable, determine whether the difference analysis result between the preliminarily selected discrete environmental variable and the stratigraphic structure depth is less than a preset difference threshold; if so, determine that the stratigraphic structure depth is not affected by the preliminarily selected discrete environmental variable and do not perform any operation; otherwise, associate the preliminarily selected discrete environmental variable with the stratigraphic structure depth, that is, construct a discrete environmental variable associated with the stratigraphic structure depth; and then obtain each discrete environmental variable associated with the stratigraphic structure depth.

3. The method for identifying the stratigraphic structure in red soil areas based on shallow seismic refraction wave method according to claim 1 is characterized by: In the step E, the correlation analysis method as the preset second analysis method is applied, and the following steps Eb-1 to Eb-2 are performed for each stratigraphic structure respectively to obtain each continuous environmental variable associated with each stratigraphic structure depth; Step Eb-1. For each sample point on each survey line in the spatial modeling set Prof_SpatiModel, according to the depth of the stratum structure corresponding to each sample point position, and the data value of each preset primary continuous environmental variable corresponding to each sample point position, perform the Pearson correlation coefficient analysis method, according to the following formula: Obtain the Pearson correlation coefficient Pearson between the stratigraphic structure depth and each preliminary continuous environmental variable, and then enter step Eb-2; wherein, PProf represents the stratigraphic structure depth corresponding to the sample point location, VPVari represents the data value of the preliminary continuous environmental variable corresponding to the sample point location, and Nt represents the number of sample points in the spatial modeling set Prof_SpatiModel; Step Eb-2. For each preliminary continuous environmental variable, determine whether the Pearson correlation coefficient Pearson between the preliminary continuous environmental variable and the stratigraphic structure depth is less than a preset difference threshold. If so, determine that the stratigraphic structure depth is not affected by the preliminary continuous environmental variable and do not perform any operation; otherwise, associate the preliminary continuous environmental variable with the stratigraphic structure depth, that is, constitute a continuous environmental variable associated with the stratigraphic structure depth; and then obtain each continuous environmental variable associated with the stratigraphic structure depth.

4. A method for identifying stratum structure in red soil region based on shallow seismic refraction wave method according to any one of claims 1 to 3, characterized in that: The preliminary selected discrete environmental variables include soil parent material, land use, and hydrogeology; the preliminary selected continuous environmental variables include elevation, average annual sunshine time, average annual evapotranspiration, surface roughness, average annual rainfall, plane curvature, profile curvature, slope, slope aspect, terrain humidity index, average annual temperature, and average annual wind speed.

5. The method for identifying the stratigraphic structure in red soil regions based on shallow seismic refraction wave method according to claim 1, characterized in that: The step Fa is also included as follows: after executing the step F, the process proceeds to the step Fa; Step Fa. For each stratigraphic structure respectively, according to the correlation between environmental variables, use the stepwise linear regression analysis method to screen the environmental variables in the set of available environmental variables associated with the stratigraphic structure depth, and update the set of available environmental variables associated with the stratigraphic structure depth; then update the set of available environmental variables associated with each stratigraphic structure depth, and then enter step G.

6. The method for identifying the stratigraphic structure in red soil regions based on shallow seismic refraction wave method according to claim 1, characterized in that: In step G, the models to be trained are preset as a random forest model, a multivariate linear regression model, a geographically weighted regression model, a support vector machine model, and an artificial neural network.

7. The method for identifying the stratigraphic structure in red soil areas based on shallow seismic refraction wave method according to claim 1, characterized in that: In the steps H and J, the root mean square error is used as an accuracy test indicator to perform accuracy tests on the prediction models corresponding to the depth of the formation structure.

8. The method for identifying stratum structure in red soil region based on shallow seismic refraction wave method according to claim 1, characterized in that: The preset measuring lines in the target operation area in step A are constructed according to the following method 1, method 2, or a combination of method 1 and method 2: Method 1: Based on the preset preliminary survey lines in the target operation area, according to the accessibility of the roads in the target operation area to the survey lines, the preliminary survey lines that are not accessible by roads are deleted, and the preliminary survey lines that are accessible by roads are retained, that is, the preset survey lines in the target operation area are formed; Method 2: Based on the grid division of the target operation area and the road distribution in the target operation, survey lines are arranged at the position closest to the surrounding roads in each grid, that is, the preset survey lines in the target operation area are formed.

9. The method for identifying stratum structure in red soil region based on shallow seismic refraction wave method according to claim 1, characterized in that: In the step A, for each preset survey line in the target operation area, hardware operation is performed in the following manner to realize the application of shallow seismic refraction wave method; Based on the detectors arranged in sequence along the survey line, a hammer point is set at the middle detector position, a hammer point is set at the two end detector positions, and a hammer point is set at a preset distance outside the two end detector positions. An artificial hammer source is operated at each hammer point, and the signal is detected by each detector and uploaded to the seismograph equipment. The seismograph equipment analyzes and processes the collected signals to obtain the horizontal distribution of different stratum structure depths below the survey line, thereby forming the seismic profile data corresponding to each survey line.

10. The method for identifying stratum structure in red soil region based on shallow seismic refraction wave method according to claim 9, characterized in that: The seismograph equipment sequentially performs channel head placement, static correction, pre-stack noise suppression, amplitude compensation, pre-stack deconvolution, CMP gather extraction, velocity analysis, dynamic correction and initial stacking, residual static correction, DMO-pre-stack time migration, post-stack time migration processing, extraction of first arrival wave, creation of initial velocity model, and feedback on the collected signals from the detector, thereby obtaining the horizontal distribution of different stratum structure depths below the survey line, and then forming the seismic profile data corresponding to each survey line.

Images