Adaptive spatial constraint AHR inversion imaging method based on quasi-SAR

Abstract

The invention relates to the field of urban TEM data imaging, in particular to a quasi-SAR-based adaptive spatial constraint AHR inversion imaging method, which comprises the following steps of: 1, converting TEM data acquired in a measuring line into PSW data; 2, calculating the correlation of the measurement area based on PSW data, and generating adaptive constraint intensity to improve the imaging resolution; and 3, carrying out inversion iterative calculation by taking all the measuring points and the constraint terms as a whole, and solving a minimum value of an overall objective function by utilizing an optimization or gradient method. According to the method, efficient and high-resolution 2D imaging and 3D imaging can be carried out on urban TEM data acquired in a complex urban geological environment.

Description

Technical Field

[0001] The invention relates to the field of urban TEM data imaging, and in particular to an adaptive spatially constrained AHR inversion imaging method based on pseudo-SAR. Background Art

[0002] Conventional LCI inversion can effectively improve the imaging resolution and smoothness of single-point inversion. However, this method is suitable for detection areas with strong geological continuity. It is easy to cause distortion of inversion imaging results in towed TEM data processing in urban environments. TEM data can be converted into PSW data, thereby introducing relatively mature plane wave imaging technology (including pseudo-seismic imaging, synthetic aperture imaging, etc.) for geological interpretation. However, this TEM pseudo-plane wave imaging interpretation can usually only obtain geological structure information, and it is difficult to directly obtain underground resistivity information, which loses the unique advantage of TEM's sensitivity to conductivity. In addition, in urban underground space exploration, quickly grasping the spatial distribution information of underground resistivity is conducive to geologists to quantitatively evaluate disaster landforms such as underground cavities, collapse areas and water bursting layers. Therefore, it is also necessary to perform efficient and high-resolution 3D imaging of urban TEM data. At present, the quasi-3D imaging technology based on 1D forward modeling operators is still the mainstream TEM fast 3D imaging method, but it is also difficult to directly apply to complex urban geological environments. Therefore, it is necessary to propose an adaptive spatial constraint AHR inversion imaging method based on pseudo-SAR. Summary of the invention

[0003] The purpose of the present invention is to provide an adaptive spatially constrained AHR inversion imaging method based on pseudo-SAR, which can meet the needs of efficient and high-resolution imaging of complex urban geological environments, so as to solve the existing technical defects and unattainable technical requirements.

[0004] To achieve the above object, the present invention provides the following technical solution: an adaptive spatially constrained AHR inversion imaging method based on pseudo-SAR, comprising:

[0005] 1. Convert the TEM data collected in the survey line into PSW data;

[0006] Second, the correlation of the survey area is calculated based on PSW data, and adaptive constraint strength is generated to improve imaging resolution;

[0007] 3. Perform inversion iterative calculation on all measurement points and constraints as a whole, and use optimization or gradient method to find the minimum value of the overall objective function.

[0008] Preferably, when 2D imaging is performed on the TEM data in step 1), the specific content of step 2) is to obtain the lateral correlation of the entire survey line based on the PSW data and adaptively generate a constraint matrix along the survey line direction.

[0009] Preferably, the content of obtaining the lateral correlation of the entire survey line based on the PSW data in step 2) is:

[0010] The continuity of spatial geology is indirectly evaluated by calculating the correlation of PSW data of the survey line. The correlation coefficient between adjacent survey points is rel i The calculation is as follows:

[0011]

[0012] in,

[0013] ψ i =1 / ||U obsi -U obs(i+1) ||2

[0014] (i=1,2,...,N s -1)

[0015] Among them, U obsi is the PSW data of the i-th measuring point, Ψ i is the difference coefficient of PSW data of adjacent measuring points, Ψ max is the maximum difference coefficient calculated for the entire measurement area, Ns is the total number of measurement points, rel i It represents the correlation between the i-th measuring point and the i+1-th measuring point. The closer its value is to 1, the higher the correlation between the two adjacent measuring points.

[0016] Preferably, in step 2), the content of the constraint matrix adaptively generated along the survey line direction is:

[0017] According to the correlation of adjacent measuring points, corresponding constraints can be imposed on the overall inversion to calculate the resistivity and thickness cross-constraint adaptive weighted matrix:

[0018] R p ′Δm=Δr p +e rp

[0019] R′ p =β·W p *R p ,Δr p =-R′ p m0

[0020] Where R p is the resistivity and layer thickness lateral constraint matrix, R p' is the optimized adaptive constraint matrix, m is all inversion model parameters, m0 is the initial inversion model parameter, Δm = m-m0, is the model correction amount, e rp represents the difference of model parameters of the same layer between adjacent measuring points on the same measuring line, β is the scaling factor, and W p is the weighting factor related to resistivity and thickness constraint adaptation. The larger the weighting factor, the stronger the lateral constraint, the better the lateral continuity of the inverted parameters, and the smoother the inversion result.

[0021] Preferably, when performing 3D imaging on the TEM data in step 1), the specific content of step 2) is to improve the imaging resolution by calculating the spatial correlation of the 3D measurement area, thereby applying an adaptive spatial constraint strength.

[0022] Preferably, the specific content of calculating the spatial correlation of the 3D measurement area is:

[0023] Calculate the correlation between PSW data and obtain the correlation coefficient of adjacent points in space to indirectly evaluate the continuity of space geology.

[0024] The correlation coefficient R of spatially adjacent measuring points is calculated as follows:

[0025]

[0026] Among them, i is the difference coefficient of PSW data of adjacent measuring points, satisfying:

[0027] ψ i =1 / ||U obsi -U obs(i+1) ||2

[0028] (i=1,2,...,N s -1)

[0029] Among them, U obsi is the PSW data of the i-th measuring point, N s is the total number of measuring points, Ψ ma x is the maximum coefficient of difference calculated in the entire survey area, R i It represents the correlation between the i-th measuring point and the i+1-th measuring point.

[0030] Preferably, the content of applying adaptive spatial constraint strength to improve imaging resolution is: according to the calculated correlation coefficient of spatially adjacent measuring points, a lateral constraint factor R is applied to adjacent measuring points along the measuring line direction. x , a vertical constraint factor R is applied to the measuring points perpendicular to the measuring line direction y , and this constraint factor will directly act on the iteration of model inversion parameters.

[0031] Preferably, the constraint perpendicular to the survey line direction in the model parameter iteration is:

[0032] W′ V Δm=Δr v +e rv

[0033] in,

[0034] W′ V =β·W V , Δ r = -W′ V m0

[0035] Where β is the scaling factor of the equilibrium constraint strength (i.e., scaling coefficient), m is all the inverse model parameters, m0 is the initial inverse model parameter, △m = m-m0, is the model parameter correction, and e rv is the difference in model parameters of the same layer between adjacent measuring points on the same vertical measuring line, W V is the layer resistivity and layer thickness adaptive vertical constraint matrix, W′ V is the optimized layer resistivity and layer thickness adaptive vertical constraint matrix, W V satisfy:

[0036]

[0037] Preferably, the lateral constraints in the model parameter iteration satisfy:

[0038] W′ L Δm=Δr L +e rL

[0039] in,

[0040] W′ L =β·W L , Δr L = -W′ L m0

[0041] In the formula, e rL is the difference in model parameters between adjacent measuring points in the horizontal direction.

[0042] Preferably, the W L satisfy:

[0043]

[0044] in,

[0045]

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] 1. This application can perform efficient, high-resolution 2D and 3D imaging of urban TEM data collected in a complex urban geological environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 The adaptive lateral constraint away map in embodiment 1 of the present invention;

[0049] Figure 2 It is a comparative view of imaging results of three inversion methods of complex geological models in Example 1, Comparative Example 1 and Comparative Example 2 of the present invention;

[0050] Figure 3 Identification of PSW data correlation of adjacent measuring points in Embodiment 1 of the present invention;

[0051] Figure 4 This is a schematic diagram of the principle of adaptive space constraints in Example 2 of the present invention;

[0052] Figure 5 It is the 3D measurement area used in the simulation experiments in Example 2, Comparative Example 3 and Comparative Example 4 of the present invention;

[0053] Figure 6 The "field"-shaped abnormal geological model involved in the simulation experiments in Example 2, Comparative Example 3 and Comparative Example 4 of the present invention;

[0054] Figure 7 It is a schematic diagram of the calculation results of the correlation of the measuring points along the measuring line direction based on the PSA data in the simulation experiments in Example 2, Comparative Example 3 and Comparative Example 4 of the present invention;

[0055] Figure 8 It is a schematic diagram of the calculation results of the correlation of the measuring points in the vertical measuring line direction based on the PSA data in the simulation experiments in Example 2, Comparative Example 3 and Comparative Example 4 of the present invention;

[0056] Fig. 9 The imaging results of the three inversion methods in Example 2, Comparative Example 3 and Comparative Example 4 of the present invention are compared;

[0057] Fig.10 Graphs showing imaging effect parameters of three inversion strategies in Example 2, Comparative Example 3 and Comparative Example 4 of the present invention;

[0058] Fig.11 The average RMSE and correlation results of the three inversion methods in Example 2, Comparative Example 3 and Comparative Example 4 of the present invention after 50 repeated calculations. DETAILED DESCRIPTION

[0059] The following will be combined with the attached embodiment of the present invention Figure 1-11, the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] See also Figure 1-11 , embodiments of the present invention:

[0061] Example 1: This example is 2D imaging of urban TEM data

[0062] The AHR imaging method can be decomposed into three steps. The first step is to convert the TEM data collected within the survey line into PSW data. The second step is to obtain the lateral correlation of the entire survey line based on the PSW data, and adaptively generate the constraint matrix along the survey line. The third step is to perform inversion iterative calculations on all the measurement points and constraint items as a whole, and use the optimization or gradient method to obtain the minimum value of the overall objective function. The constraint items of the adaptive lateral constraint are calculated using the correlation between PSW data that is sensitive to the underground electrical structure. Regardless of the continuity of the underground geology, a reasonable constraint strength can be applied, making it geologically adaptive. Specifically:

[0063] The continuity of spatial geology is indirectly evaluated by calculating the correlation of PSW data of the survey line. The correlation coefficient between adjacent survey points is rel i The calculation is as follows:

[0064]

[0065] in,

[0066] ψ i =1 / ||U obsi -U obs(i+1) ||2

[0067] (i=1,2,...,N s -1)

[0068] Among them U obsi is the PSW data of the i-th measuring point, Ψ i is the difference coefficient of PSW data of adjacent measuring points, Ψ max is the maximum difference coefficient calculated for the entire measurement area, Ns is the total number of measurement points, rel i It represents the correlation between the i-th measuring point and the i+1-th measuring point. The closer its value is to 1, the higher the correlation between the two adjacent measuring points. According to the correlation of adjacent measuring points, corresponding constraints can be imposed on the overall inversion. Drawing on the LCI idea, the resistivity and thickness cross-constraint adaptive weighted matrix calculation is performed:

[0069] Rp ′Δm=Δr p +e rp

[0070] R′ p =β·W p *R p , Δr p =-R′ p m0

[0071] Where R p is the resistivity and layer thickness lateral constraint matrix, R p ' is the optimized adaptive constraint matrix; m is all inversion model parameters; m0 is the initial inversion model parameter, Δm = m-m0, is the model correction amount, e rp represents the difference of model parameters of the same layer between adjacent measuring points on the same measuring line, β is the scaling factor, and W p is the weighting factor related to the adaptive resistivity and thickness constraints. The larger the weighting factor, the stronger the lateral constraint, the better the lateral continuity of the inverted parameters, and the smoother the inversion result. Figure 1 As shown in the figure. After the constraint matrix is ​​calculated, all the measuring points are taken as a whole for inversion imaging iteration. By finding suitable model parameters to fit the acquired data, the underground resistivity distribution can be obtained. On this basis, high-resolution imaging is achieved through adaptive model constraints.

[0072] In order to verify the imaging effect of this application, we established Figure 2 (a) shows a model consisting of an 8m thick low-resistivity layer at a depth of 6m. The layer is thinner in the middle and disappears in the central part of the profile. The resistivity of the low-resistivity stratum and the surrounding environment are 5Ω·m and 100Ω·m, respectively. Random changes are added and superimposed on the stratum resistivity to be closer to the actual geological structure (Auken 2005). In addition, 5% random noise is added to the forward simulation data to prove that the application can reasonably and accurately map the missing features of the model layer thickness.

[0073] like Figure 3 As shown, the correlation of adjacent measuring points calculated based on PSW data is shown. It can be seen that the model correlation between points 10 and 11 and points 12 and 13 is the lowest, and the correlation on both sides of the measuring line increases with the decrease of layer thickness change, which is basically consistent with the theoretical model. Figure 2 (d) shows the imaging results of the AHR method.

[0074] It needs to be explained that Figure 2 It includes 4 parts, namely Figure 2 (a) Figure 2 (b) Figure 2 (c) and Figure 2 (d) are the true model, SVD single-point inversion imaging results, LCI inversion imaging results, and AHR inversion imaging results.

[0075] Comparative Example 1:

[0076] The difference between this comparative example and Example 1 is that this comparative example uses the SVD inversion method without constraints to invert Figure 2 The model shown in (a) was imaged, and the results were shown in Figure 2 (b) as shown.

[0077] from Figure 2 (b) shows that the unconstrained SVD inversion method can image the overall structure of the anomaly and basically map out the thin layer characteristics of low resistivity in the middle area. Figure 2 (b)】. However, due to the influence of noise, the lateral smoothness of the imaging results is relatively poor, and the missing layer thickness features in the central area are not well mapped.

[0078] Comparative Example 2:

[0079] The difference between this comparative example and Example 1 is that this comparative example uses the LCI inversion method to Figure 2 The model shown in (a) was imaged, and the results were shown in Figure 2 (c) as shown.

[0080] from Figure 2 It can be seen from (c) that the LCI method improves the lateral smoothness of the imaging, but weakens the mapping of the missing layer thickness features, resulting in distorted imaging in the central area. Therefore, the LCI method is not suitable for fault zone geology and is difficult to be directly applied to complex urban geological environments.

[0081] In summary: AHR inversion has overcome the shortcomings of the SVD and LCI methods to a certain extent, which is reflected in the improvement of the lateral smoothness of its imaging results, the good mapping of the missing details of the intermediate layer thickness, and the improvement of the overall imaging resolution, which further verifies the feasibility of the AHR inversion method in complex geological environments.

[0082] Example 2: This example is 3D imaging of urban TEM data

[0083] For 3D imaging of urban TEM data, the spatial correlation of the 3D survey area is also calculated through PSW data, so as to apply adaptive spatial constraint strength to improve imaging resolution. First, the correlation between PSW data is calculated to indirectly evaluate the continuity of spatial geology. The correlation coefficient R of spatial adjacent survey points is calculated as follows:

[0084]

[0085] Among them, Ψi is the difference coefficient of PSW data of adjacent measuring points, which satisfies:

[0086] ψ i =1 / ||U obsi -U obs(i+1) ||2

[0087] (i=1,2,...,N s -1)

[0088] Among them, U obsi is the PSW data of the i-th measuring point, Ns is the total number of measuring points, Ψ max The maximum coefficient of variation calculated for the entire survey area.

[0089] like Figure 4 As shown in the figure, a set of spatial measurement areas consisting of K measurement lines (each measurement line contains M measurement points) is shown. In order to realize the spatial constraint of the measurement area, according to the correlation coefficient of the spatial adjacent measurement points calculated previously, a lateral constraint factor R is applied to the adjacent measurement points along the measurement line direction. x , a vertical constraint factor R is applied to the measuring points perpendicular to the measuring line direction y , and this constraint factor will directly act on the iteration of model inversion parameters.

[0090] First, the constraints perpendicular to the survey line in the iteration of model parameters (mainly including layer resistivity and layer thickness) are introduced:

[0091] W′ V Δm=Δr v +e rv

[0092] in,

[0093] W′ V =β·W V , Δ r =-W' V m0

[0094] Where β is the scaling factor of the equilibrium constraint strength, m, m0 and Δm are the model parameters, initial model parameters and model parameter corrections, respectively, and e rv is the model parameter difference between adjacent vertical measuring points. v is the adaptive vertical constraint matrix, satisfying:

[0095]

[0096] Similar to the vertical model parameter constraints, the lateral constraints of the model parameters satisfy:

[0097] W′ L Δm=Δr L +erL

[0098] in,

[0099] W′ L =β·W L , Δr L = -W′ L m0

[0100] In the formula, e rL is the model parameter difference between adjacent measuring points in the horizontal direction. L is the adaptive lateral constraint matrix, satisfying:

[0101]

[0102] in,

[0103]

[0104] After completing the calculation of the spatial constraint matrix, all the measurement points are taken as a whole for inversion iteration.

[0105] In order to verify the imaging quality of this embodiment, a simulation experiment is performed. Figure 5 The 3D survey area shown in the figure consists of 121 survey points, with 11 points collected on each survey line, for a total of 11 survey lines. The background geological resistivity of the survey area is set to 100Ω·m, and the abnormal geological resistivity is set to 5Ωm. Figure 6 As shown in (b), the low-resistance anomaly is set in the 10-15m underground interval, and the overall shape is crisscrossed to briefly simulate the distribution of underground domestic water pipes. In order to more intuitively reflect the geological conditions of the survey area, Figure 6 (c) shows the 2D geological map after all the survey points in the survey area are stitched together in sequence. It should be noted that the color bar ranges of the two maps are different. Figure 6 In order to highlight the underground anomaly in (b), the colorbar range is reduced to remove the background geology. Figure 6 The anomaly in (c) [i.e. Figure 6 (c) The rectangular parts of #3-#9 with colors different from the background and all (three) strip parts] and Figure 6 (b) Abnormality [i.e. Figure 6 The TEM data of the entire survey area are obtained by numerical forward simulation. When the signal-to-noise ratio of the TEM data reaches 30 dB or more, it has good stability. Therefore, in this embodiment, 5% Gaussian random noise is added to the TEM data of the forward simulation to simulate the field environment.

[0106] It needs to be explained that Figure 5The content is a schematic diagram of the 3D survey area of TEM. The "field" area corresponds to the situation where there are low-resistivity bodies underground, and the remaining area corresponds to the homogeneous half-space geology. And Figure 6 has two parts, namely Figure 6 (c) and Figure 6 (b), which respectively represent the 3D view of the underground low-resistivity anomaly body and the 2D survey area view stitched according to the measuring point order.

[0107] After converting the TEM data into PSW data, the correlation of the geology of the entire survey area is obtained. Figure 7 Shows the calculation results of the correlation degree along the survey line direction. The geological mutation area (i.e., the boundary between the homogeneous half-space geology and the low-resistivity anomaly) is defined as the low-correlation area, and the theoretically low-correlation area is circled by a dotted line in the figure. It can be seen that the low-correlation area obtained using PSW data is basically consistent with the theory. Figure 8 Shows the calculation results of the correlation degree in the direction perpendicular to the survey line, which is also basically consistent with the theory in the low-correlation area. Although affected by noise, the difference degree of the theoretically high-correlation area in the survey area is relatively high (the correlation degree generally varies between 0.2 and 1), but this is sufficient to ensure the information transfer between adjacent continuous strata during the inversion iteration process. Because in conventional lateral or spatial constrained inversion, the good range of constraint factor values is usually only between 0.02 and 0.2.

[0108] Through the method of this embodiment, 3D imaging calculation is performed on the Figure 7 geological model, as shown in Fig. 9 (c): The left side is the 3D view of the imaging result of the anomaly body, and the right side is the 2D view.

[0109] It should be explained that Figure 7 represents the correlation of the measuring points along the survey line direction, Figure 8 represents the correlation of the measuring points in the direction perpendicular to the survey line direction. The dotted lines in the two figures give the theoretically low-correlation areas, and Fig. 9 includes a small part, namely Fig. 9 (a), Fig. 9 (b) and Fig. 9 (c), which are the SVD inversion result, the SCI inversion result, and the AHR-3D inversion result respectively. Fig. 9 On the left side of the whole is the 3D view of the imaging result, and on the right side is the 2D view after all the measuring points are stitched in order.

[0110] The AHR-3D inversion method ensures the imaging resolution while increasing the imaging smoothness, and the imaging result is relatively closer to the real model.

[0111] The root mean square error (RMSE), correlation (Corr2 function in MATLAB) and calculation time between the imaging result obtained by this embodiment and the real model are calculated. The results are as follows: Fig.10 , it needs to be explained that Fig.10 and Fig. 9 The data in corresponds.

[0112] Comparative Example 3:

[0113] The difference from Example 2 is that this comparative example uses the SVD inversion method to Figure 7 The geological model is used for 3D imaging calculation, such as Fig. 9 (a) shows: the left side shows the 3D perspective of the abnormal body imaging result, and the right side shows the Fig. 9 (c) Consistent 2D view on the right.

[0114] The SVD inversion method can basically present the underground low-resistance structure, but from the displayed 2D imaging details, it can be seen that the overall imaging smoothness is poor and the local details are slightly distorted.

[0115] The root mean square error (RMSE), correlation (Corr2 function in MATLAB) and calculation time between the imaging results obtained by this comparative example and the real model are calculated. The results are as follows Fig.10 As shown, the RMSE of the SVD method is 16.184 and the imaging correlation is 0.752.

[0116] Comparative Example 4:

[0117] The difference from Example 2 is that this comparative example uses the SCI inverse training method to Figure 7 The geological model is used for 3D imaging calculation, such as Fig. 9 (b) shows: the left side shows the 3D perspective of the abnormal body imaging result, and the right side shows the Fig. 9 (c) Consistent 2D view on the right.

[0118] The SCI inversion method increases the continuity of the overall imaging, but the spatial constraint expands the burial depth of the anomaly, resulting in serious local distortion of the imaging results.

[0119] The root mean square error (RMSE), correlation (Corr2 function in MATLAB) and calculation time between the imaging results obtained by this comparative example and the real model are calculated. The results are as follows Fig.10 , the SCI method produces serious local distortion, the RMSE increases to 31.643, and the correlation decreases to 0.573. This also shows that under complex geological conditions, the imaging results of the SCI method are not as good as the SVD inversion results of single-point splicing.

[0120] In summary: Fig.10 The RMSE value (10.753) of the AHR-3D ​​method is 33.56% lower than that of the SVD method, and the correlation value (0.887) is 17.96% higher than that of the SVD method, which shows a significant improvement in imaging. All three imaging methods maintain a high computational efficiency. The AHR-3D ​​method is about 2s slower than the other two methods because the AHR-3D ​​method has an additional calculation time for wavefield transformation. Considering that the AHR-3D ​​algorithm improves the overall imaging resolution, the extra calculation time is acceptable.

[0121] Furthermore, repeated experiments were designed to evaluate the stability of the three inversion methods. Under 5% noise interference, the three inversion methods were repeated 50 times to calculate the average RMSE and correlation. Fig.11 The calculation results of repeated experiments are shown. The repeated experimental results also show that compared with SVD and SCI inversion, the AHR-3D ​​inversion method always has a smaller average RMSE (12.09) and a higher average correlation value (0.87), indicating that the AHR-3D ​​method has better stability.

[0122] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic features of the present invention. Therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the attached claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any figure mark in the claims should not be regarded as limiting the claims involved.

[0123] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation modes that those skilled in the art can understand.

Claims

1. An adaptive spatially constrained AHR inversion imaging method based on pseudo-SAR, characterized in that: include:

1. Convert the TEM data collected in the survey line into PSW data; Second, the correlation of the survey area is calculated based on PSW data, and adaptive constraint strength is generated to improve imaging resolution; 3. Perform inversion iterative calculation on all measurement points and constraints as a whole, and use optimization or gradient method to find the minimum value of the overall objective function.

2. The method for adaptive spatially constrained AHR inversion imaging based on pseudo-SAR according to claim 1, characterized in that: When 2D imaging is performed on the TEM data in step 1), the specific content of step 2) is to obtain the lateral correlation of the entire survey line based on the PSW data and adaptively generate a constraint matrix along the survey line direction.

3. The method for adaptive spatially constrained AHR inversion imaging based on pseudo-SAR according to claim 2, characterized in that: The content of obtaining the lateral correlation of the entire survey line based on PSW data in step 2) is: The continuity of spatial geology is indirectly evaluated by calculating the correlation of PSW data of the survey line. The correlation coefficient between adjacent survey points is rel i The calculation is as follows: in, ψ i =1 / ||U obsi -IN obs(i+1) ||2 (i=1,2,...,N s -1) Among them, U obsi is the PSW data of the i-th measuring point, Ψ i is the difference coefficient of PSW data of adjacent measuring points, Ψ max is the maximum difference coefficient calculated for the entire measurement area, Ns is the total number of measurement points, rel i It represents the correlation between the i-th measuring point and the i+1-th measuring point.

4. The method for adaptive spatially constrained AHR inversion imaging based on pseudo-SAR according to claim 3, characterized in that: In the step 2), the content of the constraint matrix adaptively generated along the survey line direction is: According to the correlation of adjacent measuring points, corresponding constraints can be imposed on the overall inversion to calculate the resistivity and thickness cross-constraint adaptive weighted matrix: R′ p Δm=Δr p +e rp R′ p =β·W p *R p ,Δr p =-R′ p m0 Where R p is the resistivity and layer thickness lateral constraint matrix, R p ′ is the optimized adaptive constraint matrix, m is all inversion model parameters, m0 is the initial inversion model parameter, Δm=m-m0, is the model correction amount, e rp represents the difference of model parameters of the same layer between adjacent measuring points on the same measuring line, β is the scaling factor, and W p is the weighting factor related to resistivity and thickness constraint adaptation.

5. The method for adaptive spatially constrained AHR inversion imaging based on pseudo-SAR according to claim 1, characterized in that: When 3D imaging is performed on the TEM data in step 1), the specific content of step 2) is to improve the imaging resolution by calculating the spatial correlation of the 3D measurement area, thereby applying an adaptive spatial constraint strength.

6. The method of adaptive spatially constrained AHR inversion imaging based on pseudo-SAR according to claim 5, characterized in that: The specific content of calculating the spatial correlation of the 3D measurement area is: Calculate the correlation between PSW data and obtain the correlation coefficient of adjacent points in space to indirectly evaluate the continuity of space geology. The correlation coefficient R of spatially adjacent measuring points is calculated as follows: Among them, Ψi is the difference coefficient of PSW data of adjacent measuring points, which satisfies: ψ i =1 / ||U obsi -IN obs(i+1 )||2 (i=1,2,...,N s -1) Among them, U obsi is the PSW data of the i-th measuring point, Ns is the total number of measuring points, Ψmax is the maximum difference coefficient calculated in the entire measuring area, and Ri represents the correlation between the i-th measuring point and the i+1-th measuring point.

7. The method for adaptive spatially constrained AHR inversion imaging based on pseudo-SAR according to claim 6, characterized in that: The content of applying adaptive spatial constraint strength to improve imaging resolution is: according to the calculated correlation coefficient of spatially adjacent measuring points, a lateral constraint factor Rx is applied to the adjacent measuring points along the measuring line direction, and a vertical constraint factor Ry is applied to the measuring points perpendicular to the measuring line direction, and this constraint factor will directly act on the iteration of the model inversion parameters.

8. The method of adaptive spatially constrained AHR inversion imaging based on pseudo-SAR according to claim 7, characterized in that: The constraint perpendicular to the survey line direction in the model parameter iteration is: W′ V Δm=Δr v +e rv in, IN' V =β·W V ,Δr V =-W′ V m0 Where β is the scaling factor of the equilibrium constraint strength, m is the total inversion model parameter, m0 is the initial inversion model parameter, Δm = m-m0, is the model parameter correction, and e rv is the difference in model parameters of the same layer between adjacent measuring points on the same vertical measuring line, W V is the layer resistivity and layer thickness adaptive vertical constraint matrix, W′ V is the optimized layer resistivity and layer thickness adaptive vertical constraint matrix, W V satisfy:

9. The method for adaptive spatially constrained AHR inversion imaging based on pseudo-SAR according to claim 8, characterized in that: The lateral constraints in the model parameter iteration satisfy: W′ L Δm=Δr L +e rL in, IN' L =β·W L ,Δr L =-W′ L m0 In the formula, e rL is the difference in model parameters between adjacent measuring points in the horizontal direction.

10. The method for adaptive spatially constrained AHR inversion imaging based on pseudo-SAR according to claim 9, characterized in that: The W L satisfy: in,

Images