Underground medium density inversion method and system based on high-precision microgravity

Abstract

The invention relates to the technical field of underground medium density inversion, in particular to an underground medium density inversion method and system based on high-precision microgravity, and the method comprises the steps: obtaining the gravity data of each observation point in an underground medium density inversion ground surface measurement region; calculating a superposition abnormal value of each gravity subsequence of each observation point, extracting a superposition characteristic value of each modal component, obtaining the superposition interference confidence of each gravity subsequence of each observation point, obtaining the superposition influence persistence of each observation point, adjusting the wavelet threshold of the wavelet transform of the gravity data on each observation point, and calculating the superposition abnormal value of each gravity subsequence of each observation point. And correcting the gravity data of each observation point after denoising processing to extract Bouguer gravity anomaly of each observation point, and obtaining an inversion result of the underground medium density through a Parker-Oldenburg density interface inversion algorithm. According to the method, the underground medium density can be effectively and accurately inverted through the high-precision microgravity data.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of underground medium density inversion, in particular to an underground medium density inversion method and system based on high-precision microgravity. BACKGROUND

[0002] With the rapid development of gravity measurement technology, satellite gravity detection technology, satellite altimetry technology and the like, people's understanding of the gravity field is deepening. By using a gravity anomaly monitoring instrument with a precision of microgal level, the abnormal monitoring demand of microgal level gravity can be met, so that the underground medium density can be more accurately inverted, which is helpful for fine monitoring of underground oil reservoirs.

[0003] In the prior art, the underground medium density is inverted by using gravity data. However, due to the interference of factors such as height error and zero point drift, the prior art usually adopts dual-tree complex wavelet transform (DTCWT) to denoise the gravity data, so as to improve the accuracy of the gravity data and more accurately invert the underground medium density. However, since the gravity data is interfered by multiple factors, the prior art cannot sufficiently consider the multiple factors to accurately denoise the gravity data, which may result in poor denoising effect of the gravity data and affect the accuracy of the inversion of the underground medium density. SUMMARY

[0004] In order to solve the above technical problems, the purpose of the present application is to provide an underground medium density inversion method and system based on high-precision microgravity, and the technical solution adopted is as follows: The embodiment of the present application provides an underground medium density inversion method based on high-precision microgravity, comprising the following steps: Obtaining gravity data of each observation point in the underground medium density inversion surface measurement area; Extracting gravity subsequences of each observation point and performing frequency domain analysis, obtaining superposition anomaly values of each gravity subsequence of each observation point according to the difference degree and difference discrete anomaly of the instantaneous amplitude of the gravity subsequence in the frequency domain and the corresponding amplitude of other frequency domains; Modal decomposition is performed on each gravity subsequence, and superposition eigenvalues of each modal component are extracted through random changes of each modal component. According to the difference between the superposition eigenvalues of different observation points with respect to each gravity subsequence and the deviation of the superposition anomaly values, the superposition interference confidence of each gravity subsequence of each observation point is obtained, and the superposition influence duration of each observation point is obtained through the change of the superposition interference confidence of all gravity subsequences of each observation point, which is used to adjust the wavelet threshold of the wavelet transform of the gravity data at each observation point to denoise the gravity data. The gravity data of each observation point after the denoising processing is corrected to extract the Bouguer gravity anomaly of each observation point, and the inversion result of the underground medium density is obtained through the Parker-Oldenburg density interface inversion algorithm.

[0005] Preferably, the gravity data of each observation point is arranged in time sequence to form a gravity data sequence of each observation point, and is uniformly divided into a plurality of subsequences as a gravity subsequence of each observation point.

[0006] Preferably, the method for obtaining the stacking anomaly value of each gravity subsequence of each observation point is: ; in the formula, is the stacking anomaly value of the kth gravity subsequence of the current observation point, is the first mean value of the kth gravity subsequence of the current observation point, is the first sum value of the kth gravity subsequence of the current observation point.

[0007] Preferably, for the current observation point, the frequency domain transformation is performed on each gravity subsequence to obtain the amplitude corresponding to all frequencies in the gravity subsequence, and the Hilbert Huang transformation is performed on each gravity subsequence to extract the instantaneous amplitude of the gravity subsequence, and the mean value of the absolute difference between the instantaneous amplitude of each gravity subsequence and the amplitude corresponding to all frequencies is taken as the first mean value of each gravity subsequence.

[0008] Preferably, the absolute difference corresponding to each gravity subsequence is detected respectively, and the sum value of the local outlier factors of all absolute differences is taken as the first sum value of each gravity subsequence.

[0009] Preferably, the method for extracting the stacking eigenvalue of each modal component is: The first difference vector of each modal component is calculated, and the permutation entropy of each modal component is calculated, and the normalized result of the product of the mean value of the absolute values of all elements in the first difference vector and the permutation entropy is taken as the stacking eigenvalue of each modal component.

[0010] Preferably, the method for obtaining the stacking interference confidence of each gravity subsequence of each observation point is: ; in the formula, is the stacking interference confidence of the kth gravity subsequence of the current observation point, is the difference degree between the stacking eigenvalue of the kth gravity subsequence of the current observation point and the jth adjacent observation point thereof, is the absolute difference between the stacking anomaly value of the kth gravity subsequence of the current observation point and the jth adjacent observation point thereof, In order to avoid the denominator being 0, the plurality of observation points closest to the current observation point are recorded as the adjacent observation points of the current observation point.

[0011] Preferably, the method for obtaining the superposition influence persistence degree of each observation point is as follows: The superposition confidence degree sequence of each observation point is obtained by arranging the superposition interference confidence degrees of all the gravity subsequences at each observation point in time sequence, and the ratio between the mean value of all elements in the superposition confidence degree sequence and the mean value of the absolute values of all elements in the first-order difference sequence of the superposition confidence degree sequence is taken as the superposition influence persistence degree of each observation point.

[0012] Preferably, the method for adjusting the wavelet threshold value of the wavelet transform of the gravity data at each observation point is as follows: , wherein, is the wavelet adjustment threshold value of the gravity data at the current observation point, is the standard deviation of the estimated noise in the wavelet denoising algorithm, is the length of the gravity data sequence at the current observation point, is the adjustment coefficient, is the hyperbolic tangent function, is the superposition influence persistence degree of the current observation point.

[0013] The embodiments of the present application also provide a high-precision microgravity-based underground medium density inversion system, which comprises a memory, a processor, and a computer program stored in the memory and running on the processor, and the processor implements the steps of the high-precision microgravity-based underground medium density inversion method according to any one of the above embodiments when executing the computer program.

[0014] As can be seen from the above, the high-precision microgravity-based underground medium density inversion method and system provided by the present application at least have the following beneficial effects: The present application adopts the frequency domain analysis method and combines the outlier detection algorithm to more accurately measure the abnormal characteristics of the instantaneous amplitude of the gravity subsequence under the superposition of multiple frequencies, thereby clearly showing the severity of the gravity data in the gravity subsequence disturbed by the external factor superposition noise, which is conducive to more accurately determining the wavelet threshold value and more accurately performing the denoising processing on the gravity data sequence. Further, the present application extracts the superposition feature vector by adopting the modal decomposition method and considers the consistency of the external noise superposition influence between the adjacent observation points to accurately measure the confidence degree characteristics of the gravity subsequence at the observation point disturbed by the superposition of multiple noise factors, thereby improving the accuracy of the analysis on the multiple factor superposition interference. The application accurately measures the persistence characteristics of the influence of noise factors superimposed interference on gravity data by superimposing the changes of interference confidence in multiple time periods, and accurately adjusts the wavelet threshold in the wavelet denoising algorithm through the superimposed influence persistence, thereby improving the effect of denoising of gravity data, and further more effectively and accurately inverting the underground medium density through high-precision micro-gravity data. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0016] Figure 1 The step flow chart of the underground medium density inversion method based on high-precision micro-gravity provided by the present application. DETAILED DESCRIPTION

[0017] In order to further illustrate the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the specific embodiments, structure, features and effects of the underground medium density inversion method and system based on high-precision micro-gravity according to the present application are described in detail as follows. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0018] Unless otherwise defined and limited, such as the terms "comprise", "include" or any other variants thereof, are intended to cover non-exclusive inclusion, so that the circuit structure, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or includes elements inherent to such article or device. Without more limitation, the element limited by the statement "including one" does not exclude the presence of another identical element in the article or device including the element. In addition, the term "and / or" used herein includes any and all combinations of one or more related listed items. All technical and scientific terms used herein have the same meaning as understood by those skilled in the art of the technology to which the present application belongs.

[0019] The specific scheme of the underground medium density inversion method and system based on high-precision micro-gravity provided by the present application is described in detail below with reference to the drawings.

[0020] Please refer to Figure 1Fig. 1 shows a flow chart of the underground medium density inversion method based on high-precision microgravity according to an embodiment of the present application, which comprises the following steps: Step 1: Obtain gravity data of each observation point in the surface measurement area of the underground medium density inversion.

[0021] In order to improve the accuracy of the inversion of the underground medium density, it is necessary to more accurately denoise the gravity data collected by the quantum intelligent sensor, eliminate the noise interference of the gravity data caused by the height error and zero drift and other external factors, and thus more effectively and accurately invert the underground medium density through the high-precision microgravity signal.

[0022] Firstly, in the present embodiment, preferably, the surface measurement area of the underground medium density inversion is selected, and 10 observation points are set in the measurement area. The implementer can set the position of the measurement area and the number of observation points according to the actual situation.

[0023] Further, the micro-g level gravity data of each observation point is collected by the quantum intelligent sensor. In the present embodiment, the gravity collection time interval is 1s, and the collection time of a single observation point is 30min. The implementer can adaptively set it according to the actual demand. In the present embodiment, the quantum intelligent sensor used is a quantum gravimeter. The quantum gravimeter has the core functions of an intelligent sensor, including gravity perception, data collection, analysis and processing, storage and communication functions. The quantum gravimeter uses sensor technology to perceive the difference in gravity of atoms or particles in the gravity field, realizing the collection function of gravity data in the gravity field.

[0024] At the same time, in order to facilitate the subsequent denoising analysis of the micro-g level gravity data, through the analysis and processing function of the intelligent sensor, preferably, in the present embodiment, the micro-g level gravity data collected at a single observation point is arranged in time sequence to obtain the gravity data sequence of each observation point in the measurement area, so as to facilitate the subsequent storage and communication of the gravity data at different observation points.

[0025] Step 2: Extract the gravity subsequence of each observation point and perform frequency domain analysis. According to the difference degree and difference discrete anomaly of the instantaneous amplitude of the gravity subsequence in the frequency domain and other corresponding amplitudes, the superposition anomaly value of each gravity subsequence of each observation point is obtained.

[0026] The prior art usually uses dual-tree complex wavelet transform (DTCWT) to denoise the gravity data. In the wavelet denoising algorithm, the wavelet threshold is a general threshold, and the multi-factor superimposed interference is not fully considered for accurately denoising the gravity data, which is easy to cause the wavelet threshold to be too large or too small.

[0027] However, if the wavelet threshold is too large, the denoised signal will be too smooth, and if the wavelet threshold is too small, the noise component in the gravity data cannot be effectively suppressed, which will ultimately affect the accuracy of the inversion of the density of the underground medium. Therefore, when collecting gravity data by a quantum intelligent sensor in an actual process, it will be affected by external factors, so it is necessary to fully consider that the gravity field signal is disturbed by the superposition of multiple factors, improve the wavelet threshold in the wavelet denoising algorithm, and improve the denoising effect of the gravity data and the accuracy of the inversion of the density of the underground medium.

[0028] In order to extract the disturbed features of the gravity data at different time periods, the gravity data sequence of each observation point is uniformly divided into K subsequences, denoted as K gravity subsequences of each observation point. In this embodiment, K is 30, so that the time length of each gravity subsequence is 1 min.

[0029] Further, for each gravity subsequence of each observation point, each gravity subsequence is taken as the input of DFT discrete Fourier transform, and the amplitudes corresponding to all frequencies in each gravity subsequence are obtained by DFT discrete Fourier transform; at the same time, each gravity subsequence of each observation point is taken as the input of Hilbert Huang transform, and the instantaneous amplitude of each gravity subsequence is obtained by using Hilbert Huang transform, wherein DFT discrete Fourier transform and Hilbert Huang transform are both known technologies, and will not be described in detail.

[0030] Generally, the instantaneous amplitude is the amplitude under the superposition of different frequencies. If the difference between the instantaneous amplitude and the amplitudes corresponding to all frequencies is larger, and the abnormal degree of the difference between the amplitudes corresponding to all frequencies is more prominent, the abnormal characteristics of the instantaneous amplitude under the superposition of multiple frequencies are more reflected, the gravity data is more seriously disturbed by the superposition of multiple factors, the gravity data collected by the quantum intelligent sensor is more complexly affected by the disturbance, and then a larger wavelet threshold needs to be set to more effectively suppress the superposition noise in the gravity data.

[0031] Therefore, for the kth gravity subsequence of the current observation point, the absolute difference between the instantaneous amplitude of the kth gravity subsequence and the amplitudes corresponding to all frequencies in the gravity subsequence is calculated, and the mean of all absolute differences is denoted as the first mean of the kth gravity subsequence of the current observation point. The larger the first mean is, the deeper the instantaneous amplitude is affected by the superposition of different frequencies, and the gravity data is more easily affected by the superposition of external noise components.

[0032] Meanwhile, all absolute difference values corresponding to the kth gravity sub-sequence of the current observation point are taken as inputs of the LOF anomaly value detection algorithm, the preset neighborhood parameter value in the algorithm is 10, the local outlier factor of all absolute difference values is obtained through the LOF anomaly value detection algorithm, and the sum of the local outlier factors of all absolute difference values is denoted as the first sum value of the kth gravity sub-sequence of the current observation point. The greater the first sum value, the more prominent the abnormal degree of the difference of the amplitude corresponding to all frequencies as a whole, and the more it can reflect the abnormal characteristics of the instantaneous amplitude under the superposition of multiple frequencies at this time.

[0033] Based on the above analysis, in the embodiment, the superposition abnormal value of the kth gravity sub-sequence of the current observation point is calculated as an example: In the formula, is the superposition abnormal value of the kth gravity sub-sequence of the current observation point, is the first mean value of the kth gravity sub-sequence of the current observation point, is the first sum value of the kth gravity sub-sequence of the current observation point.

[0034] The superposition abnormal value reflects the abnormal characteristics of the instantaneous amplitude under the superposition of multiple frequencies, the greater the superposition abnormal value, the more prominent the abnormal characteristics of the instantaneous amplitude of the gravity sub-sequence under the superposition of multiple frequencies, the more serious the superposition noise interference of the gravity data by external factors such as height error and zero drift at this time, the more inaccurate the gravity data sensed by the quantum intelligent sensor, and the greater the wavelet threshold value needs to be set, so as to more effectively suppress the noise superposition in the gravity data.

[0035] Step 3: Each gravity sub-sequence is modally decomposed, and the superposition eigenvalue of each modal component is extracted through the random change of each modal component. According to the difference between the superposition eigenvalues of each gravity sub-sequence of different observation points and the deviation of the superposition abnormal value, the superposition interference confidence of each gravity sub-sequence of each observation point is obtained, and the superposition influence duration of each observation point is obtained through the change of the superposition interference confidence of all gravity sub-sequences of each observation point, which is used to adjust the wavelet threshold value of the wavelet transform of the gravity data at each observation point to denoise the gravity data.

[0036] Further, the kth gravity sub-sequence at each observation point is taken as the input of the EMD empirical mode decomposition algorithm. The EMD empirical mode decomposition algorithm decomposes the gravity sub-sequence into each modal component through an iterative process. The higher the volatility of the element change in the modal component, and the greater the uncertainty of the element change in the modal component, the more likely the modal component is to produce superposition interference to the gravity data, thereby affecting the authenticity of the gravity data.

[0037] Therefore, the first-order difference vector of each modal component is calculated, reflecting the fluctuation change between adjacent elements in the modal component, and the permutation entropy of each modal component is calculated, reflecting the uncertainty of element change in the modal component; further, the product of the mean value of the absolute value of all elements in the first-order difference vector and the permutation entropy is normalized, and the result is recorded as the superposition characteristic value of each modal component. The larger the superposition characteristic value is, the more likely the modal component is to cause superimposed interference to the gravity data.

[0038] Further, for the kth gravity sub-sequence of the current observation point, the superposition characteristic values of all modal components corresponding to the kth gravity sub-sequence are arranged in the order of modal decomposition to obtain the superposition characteristic vector of the kth gravity sub-sequence at each observation point. If the superposition characteristic vectors between each observation point and its adjacent observation point in the measurement area are more similar, and the difference between the superposition abnormal values of each observation point and its adjacent observation point is smaller, it means that the superimposed influence of multiple noise factors on the observation points in the local area is more consistent, and it can more fully and accurately indicate that the central observation point is currently interfered by multiple factors. At this time, the credibility of the gravity data collected by the quantum intelligent sensor is smaller.

[0039] Therefore, the M observation points closest to each observation point in the Euclidean distance are recorded as the M adjacent observation points of each observation point, where M is 4, representing the nearest adjacent observation points in four directions of the central observation point. According to the adjacent observation points of each observation point, the superimposed interference confidence of the kth gravity sub-sequence of each observation point is calculated: ; In the formula, is the superimposed interference confidence of the kth gravity sub-sequence of the current observation point, is the difference degree between the superposition characteristic vectors of the kth gravity sub-sequence of the current observation point and its jth adjacent observation point, and the DTW dynamic programming distance is used to measure the difference degree in this embodiment, is the absolute difference value between the superposition abnormal values of the kth gravity sub-sequence of the current observation point and its jth adjacent observation point, is to avoid the denominator being a constant of 0, and the value range is (0.01, 0.05), and the value is 0.02 in this embodiment.

[0040] The superimposed interference confidence reflects the confidence characteristics of the observation point position being superimposed by multiple noise factors from the outside world. The greater the superimposed interference confidence, the higher the credibility of the gravity subsequence corresponding to the time period being superimposed by noise factors. If the superimposed interference confidence of multiple time periods is at a high level, the duration of the influence of the superimposed noise on the gravity data is longer, and therefore a larger wavelet threshold should be set to more effectively suppress the noise superposition in the gravity data. Conversely, the shorter the duration of the influence of the superimposed noise on the gravity data, the smaller the wavelet threshold should be set to avoid the denoised signal being too smooth, thereby improving the accuracy of the gravity data collected by the quantum intelligent sensor.

[0041] Therefore, the superimposed interference confidence of all gravity subsequences at each observation point position is arranged in time sequence to obtain a superimposed confidence sequence of each observation point position. The smaller the mean value of the absolute values of all elements in the first-order difference sequence of the superimposed confidence sequence, the better the continuity of the superimposed interference confidence in the superimposed confidence sequence. Further, the ratio between the mean value of all elements in the superimposed confidence sequence and the mean value of the absolute values of all elements in the first-order difference sequence of the superimposed confidence sequence is taken as the superimposed influence duration of each observation point position. The greater the superimposed influence duration, the longer the duration of the influence of the superimposed noise on the gravity data, and therefore a larger wavelet threshold should be set.

[0042] Therefore, according to the superimposed influence duration of each observation point position, the wavelet adjustment threshold of the gravity data sequence at each observation point position is calculated as follows: ; In the formula, is the wavelet adjustment threshold of the gravity data at the current observation point position, is the standard deviation of the estimated noise in the wavelet denoising algorithm, is the length of the gravity data sequence at the current observation point position, is an adjustment coefficient to avoid the wavelet threshold being adjusted too large or too small. In this embodiment, the adjustment coefficient is 0.5, is a hyperbolic tangent function, is the superimposed influence duration of the current observation point position, wherein The calculation of is a known technology and is not described in detail.

[0043] The longer the duration of the influence of the noise factor superimposed interference on the gravity data, the larger the wavelet threshold is set, so as to more effectively suppress the noise superimposed in the gravity data. Conversely, the shorter the duration of the influence of the noise factor superimposed interference on the gravity data, the smaller the wavelet threshold is set, so as to improve the denoising effect of the gravity data collected by the quantum intelligent sensor and avoid affecting the inversion accuracy of the underground medium density.

[0044] Further, preferably, in the embodiment, the gravity data sequence of each observation point in the measurement area is taken as the input of the DTCWT dual-tree complex wavelet transform, to obtain the wavelet bottom layer low-frequency coefficient and the wavelet high-frequency coefficient, and the wavelet high-frequency coefficient is taken as the input of the wavelet denoising algorithm. The wavelet threshold is selected as the wavelet threshold of the wavelet denoising algorithm, and the soft threshold processing mode is adopted to obtain the wavelet high-frequency coefficient after threshold processing. The wavelet high-frequency coefficient after threshold processing and the wavelet bottom layer low-frequency coefficient are taken as the input of the dual-tree complex wavelet inverse transform, and the signal is reconstructed through the dual-tree complex wavelet inverse transform to obtain the gravity data sequence of each observation point after denoising processing. The DTCWT dual-tree complex wavelet transform and the wavelet denoising algorithm are all known technologies, and will not be described in detail.

[0045] Step 4: Correcting the gravity data of each observation point after denoising processing to extract the Bouguer gravity anomaly of each observation point, and obtaining the inversion result of the underground medium density through the Parker-Oldenburg density interface inversion algorithm.

[0046] Further, the underground medium density of the measurement area is inverted through the gravity data sequence of each observation point after denoising processing. Preferably, in the embodiment, the specific process is as follows: The Bouguer gravity anomaly of all observation points in the surface measurement area is obtained by performing dimension correction, terrain correction and Bouguer correction on the gravity data in the gravity data sequence, and the Bouguer gravity anomaly of all observation points in the surface measurement area is taken as the input of the Parker-Oldenburg density interface inversion algorithm. The Parker-Oldenburg formula in the Parker-Oldenburg density interface inversion algorithm is used to invert the underground medium density of all observation points. The Parker-Oldenburg density interface inversion algorithm outputs the inversion result of the underground medium density in the surface measurement area. The dimension correction, terrain correction, Bouguer correction and Parker-Oldenburg density interface inversion algorithm are all known technologies, and will not be described in detail.

[0047] Based on the same inventive concept as the above method, the underground medium density inversion system based on high-precision microgravity provided in the embodiments of the present application also includes a memory, a processor, and a computer program stored in the memory and running on the processor, and the processor implements the steps of the underground medium density inversion method based on high-precision microgravity in any of the above embodiments when executing the computer program.

[0048] Preferably, in the embodiments, the underground medium density inversion system includes a sensor acquisition module, a stacking anomaly analysis module, a wavelet denoising module, and a density inversion module. The sensor acquisition module acquires gravity data of a measurement area through a quantum intelligent sensor and pre-processes the acquired gravity data through the processing function of the intelligent sensor to obtain a gravity data sequence. The stacking anomaly analysis module is configured to extract stacking anomaly values of gravity subsequences of each observation point at different time periods in a frequency domain analysis manner. The wavelet denoising module is configured to improve a wavelet threshold to obtain a wavelet adjusted threshold and perform wavelet denoising processing on the gravity data through DTCWT dual-tree complex wavelet transform. The density inversion module is configured to perform inversion on the underground medium density of the measurement area through a Parker-Oldenburg density interface inversion algorithm. It should be noted that each module in the underground medium density inversion system is implemented through the steps and processes in the underground medium density inversion method based on high-precision microgravity.

[0049] It can be understood that the above-mentioned sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. The above describes specific embodiments of the present application. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or can be advantageous.

[0050] Each of the embodiments in the present specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. Each embodiment mainly describes the differences from other embodiments.

[0051] The above is only the implementation of the present application, and is not used to limit the scope of the present application. Any equivalent structure or equivalent process conversion using the content of the present application specification and drawings, or direct or indirect application in other related technical fields, is also included in the protection scope of the present application.

Claims

1. A method for density inversion of subsurface media based on high-precision microgravity, characterized in that, The method comprises the following steps: Obtaining gravity data of each observation point in the surface measurement area of the underground medium density inversion; Extracting gravity sub-sequences of each observation point and performing frequency domain analysis, obtaining superimposed anomaly values of each gravity sub-sequence of each observation point according to the difference degree and difference discrete anomaly of the instantaneous amplitude of the gravity sub-sequence in the frequency domain and the corresponding amplitude in the frequency domain; Modal decomposition is performed on each gravity sub-sequence, and superimposed eigenvalues of each modal component are extracted through random changes of each modal component, and superimposed interference confidence of each gravity sub-sequence of each observation point is obtained according to the difference between the superimposed eigenvalues of each gravity sub-sequence of different observation points and the deviation of the superimposed anomaly values, and the superimposed influence duration of each observation point is obtained through the change of the superimposed interference confidence of all gravity sub-sequences of each observation point, which is used to adjust the wavelet threshold of the wavelet transform of the gravity data of each observation point to denoise the gravity data; The gravity data of each observation point after denoising is corrected to extract the Bouguer gravity anomaly of each observation point, and the inversion result of the underground medium density is obtained through the Parker-Oldenburg density interface inversion algorithm.

2. The high-precision microgravity-based inversion method for subsurface medium density according to claim 1, wherein, The gravity data of each observation point is arranged in time sequence to form a gravity data sequence of each observation point, and is evenly divided into multiple sub-sequences as gravity sub-sequences of each observation point.

3. The high-precision microgravity-based inversion method for subsurface medium density according to claim 1, wherein, The superimposed anomaly value of each gravity sub-sequence of each observation point is obtained by: wherein, is the stacking anomaly value of the kth gravity sub-sequence of the current observation point, is the first mean value of the kth gravity sub-sequence of the current observation point, is the first sum value of the kth gravity sub-sequence of the current observation point.

4. The high-precision microgravity-based inversion method for subsurface medium density according to claim 3, wherein, For the current observation point, the frequency domain transform is performed on each gravity sub-sequence to obtain the amplitude corresponding to all frequencies in the gravity sub-sequence, and the Hilbert Huang transform is performed on each gravity sub-sequence to extract the instantaneous amplitude of the gravity sub-sequence, and the mean value of the absolute difference between the instantaneous amplitude of each gravity sub-sequence and the amplitude corresponding to all frequencies is taken as the first mean value of each gravity sub-sequence.

5. The high-precision microgravity-based inversion method for subsurface medium density according to claim 4, wherein, The absolute difference corresponding to each gravity sub-sequence is detected respectively, and the sum of the local outlier factors of all absolute differences is taken as the first sum value of each gravity sub-sequence.

6. The high-precision microgravity based inversion method for subsurface medium density of claim 1, wherein, The superimposed eigenvalue of each modal component is obtained by: The first difference vector of each modal component is calculated, and the permutation entropy of each modal component is calculated, and the normalized result of the product of the mean value of the absolute value of all elements in the first difference vector and the permutation entropy is taken as the superimposed eigenvalue of each modal component.

7. The high-precision microgravity based inversion method for subsurface medium density of claim 1, wherein, The superimposed interference confidence of each gravity sub-sequence of each observation point is obtained by: wherein, is the superimposed interference confidence of the kth gravity sub-sequence of the current observation point, is the difference degree between the superimposed feature vectors of the kth gravity sub-sequence of the current observation point and its jth adjacent observation point, is the absolute difference between the superimposed anomaly values of the kth gravity sub-sequence of the current observation point and its jth adjacent observation point, is a constant to avoid the denominator being 0, wherein, the multiple observation points closest to the current observation point are all recorded as the adjacent observation points of the current observation point.

8. The high-precision microgravity based inversion method for subsurface medium density of claim 1, wherein, The superimposed influence duration of each observation point is obtained by: The superimposed confidence sequence of each observation point is obtained by arranging the superimposed interference confidence of all gravity sub-sequences of each observation point in time sequence, and the ratio between the mean value of all elements in the superimposed confidence sequence and the mean value of the absolute value of all elements in the first difference sequence of the superimposed confidence sequence is taken as the superimposed influence duration of each observation point.

9. The high-precision microgravity based inversion method for subsurface medium density of claim 2, wherein, The wavelet threshold of the wavelet transform of the gravity data of each observation point is adjusted by: wherein, is the wavelet adjusted threshold of the current observation point gravity data, is the standard deviation of the estimated noise in the wavelet denoising algorithm, is the length of the gravity data sequence of the current observation point, is the adjustment coefficient, is the hyperbolic tangent function, is the superposition influence persistence of the current observation point.

10. A system for density inversion of subsurface media based on high-precision microgravity, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, The processor executes the computer program to realize the steps of the underground medium density inversion method based on high-precision microgravity in any one of claims 1-9.

Images