JP7361255B2 - Vertical magnetic survey analysis system, vertical magnetic survey analysis device, and computer readable program - Google Patents

Description

Article 30, Paragraph 2 of the Patent Act applies Article 30, Paragraph 2 applies, posted on the website on July 29, 2019 Website address: https://geonews. zenchiren. or. jp/e-Forum/2019/PDF/2019_109. pdf

Application of Article 30, Paragraph 2 of the Patent Act Application of Article 30, Paragraph 2, announced at the All-Junior Federation “Technology Forum 2019” held at the Okayama Convention Center on September 13, 2019, Okayama

The present invention relates to vertical magnetic exploration in which a borehole (observation hole) is excavated vertically downward from the surface of land where unexploded ordnance is buried, and the surrounding magnetic field is explored by scanning the observation hole with a detector. More specifically, the present invention relates to a method for generating an estimation model for vertical magnetic survey analysis used in vertical magnetic survey analysis, a vertical magnetic survey analysis system, a vertical magnetic survey analysis device, and a computer-readable program.

Vertical magnetic exploration involves inserting a detector into a borehole (observation hole) and detecting magnetic anomalies in the surrounding area. This is a method of investigating the distance and magnetic quantity of the magnetic field. This method is effective when the magnetic material is buried too deep to be detected by horizontal exploration, or when it is desired to accurately determine the buried position of the magnetic material.

In this vertical magnetic survey, a cylindrical detector attached to the tip of a cable (see, for example, Patent Document 2) is inserted into an observation hole, and the signal waveform output from the detector is recorded. This signal waveform can be recorded in real time on measuring paper using, for example, a pen recorder, so on-site engineers can proceed with drilling an observation hole while checking safety based on the signal waveform and drill to a deep location. can be explored.

In recent years, it has become possible to record the signal waveform output from the detector as digital data on a mobile PC (PC: Personal Computer), display the waveform with special software, or save it as a CAD diagram. became. After taking the mobile PC back to the office, the on-site engineer reads the characteristic waveform dimensions (maximum amplitude ΔH _max , period 2x ₀ ) on the CAD diagram, applies it to a predetermined calculation formula, and calculates the size of the observation hole. The separation distance r from the magnetic body to the magnetic body and the magnetic quantity M of the magnetic body can be calculated accurately.

Japanese Patent Application Publication No. 11-183633 Japanese Patent Application Publication No. 2014-126440

However, in geophysical exploration such as magnetic exploration, analysis based on the judgment of engineers is the mainstream, and the credibility and accuracy of the analysis results are currently left to the abilities of the engineers. In the above-mentioned vertical magnetic survey, although the signal waveform can be saved as a CAD diagram, it is necessary for the engineer to manually measure the dimensions of the signal waveform (maximum amplitude ΔH _max , period 2x ₀ , etc.) on the PC screen. The credibility and accuracy of analysis results were largely influenced by manual precision.

Therefore, the present invention provides a method for generating an estimation model for vertical magnetic survey analysis, a vertical magnetic survey analysis system, a vertical magnetic survey analysis device, and a method for generating an estimation model for vertical magnetic survey analysis that can always obtain highly accurate analysis results regardless of the ability of an engineer. and a computer-readable program.

In order to solve the above problems, the present invention provides vertical magnetic exploration that generates an estimation model for estimating the buried state of magnetic material based on the signal waveform of a vertical magnetic exploration detector inserted into an underground observation hole. A method for generating an estimation model for analysis, comprising an acquisition procedure of acquiring as training data a correspondence between a buried state of a magnetic material underground and an image of a signal waveform generated by a detector for vertical magnetic exploration; Provided is a method for generating an estimation model for vertical magnetic survey analysis, comprising the steps of: training an estimation model using a convolutional neural network using acquired teacher data, and storing the learned estimation model.

In order to solve the above problems, in the method for generating an estimation model for vertical magnetic survey analysis according to the present invention, the buried state of the magnetic body may include a separation distance r from the observation hole to the magnetic body. .

In order to solve the above problems , in the method for generating an estimation model for vertical magnetic survey analysis according to the present invention , the buried state of the magnetic body further includes a magnetic quantity M of the magnetic body, and a magnetic quantity M as seen from the observation hole. At least one of the direction θ of the magnetic pole of the magnetic material may be included.

In order to solve the above problems , in the method for generating an estimation model for vertical magnetic survey analysis according to the present invention, in the acquisition step, images of a plurality of signal waveforms are acquired while changing the buried state of the magnetic sample, The teacher data may be expanded by performing at least one of inversion processing, scaling processing, and noise processing on the images of the plurality of signal waveforms.

In order to solve the above problems , in a method for generating an estimation model for vertical magnetic survey analysis according to the present invention, the detector is a uniaxial differential fluxgate type magnetic sensor.

In order to solve the above problems , in the method for generating an estimation model for vertical magnetic survey analysis according to the present invention , the magnetic body may be an unexploded ordnance.

In order to solve the above problems, a management means for managing an estimation model generated by a method for generating an estimation model for vertical magnetic exploration analysis according to the present invention , and an image of a signal waveform generated by a detector for vertical magnetic exploration are provided. Provided is a vertical magnetic survey analysis system comprising an estimation means for estimating the buried state of a magnetic body underground by inputting it to an estimation model, and a presentation means for presenting the buried state of the magnetic body to a user. do.

In order to solve the above problems, in the vertical magnetic exploration analysis system according to the present invention , the estimation means estimates the buried state of the magnetic material by applying the dimensions of the signal waveform to a predetermined calculation formula that does not use the estimation model. and the presenting means may present the estimated buried state and the calculated buried state to the user in a state where they can be compared with each other.

In order to solve the above problems, the vertical magnetic survey analysis system according to the present invention may further include a detector for vertical magnetic survey.

In order to solve the above problems, there is provided a vertical magnetic survey analysis device applied to the vertical magnetic survey analysis system according to the present invention , which is characterized by comprising the estimation means and the presentation means. do.

In order to solve the above problems, a computer-readable program is provided that causes a computer to function as a vertical magnetic survey analysis device according to the present invention .

According to the method for generating an estimation model for vertical magnetic exploration analysis according to the present invention, the association between the buried state of a magnetic material underground and an image of a signal waveform generated by a detector for vertical magnetic exploration is used as training data. Since the estimation model is trained using a convolutional neural network (CNN), the user (engineer) does not have to manually measure the dimensions of the signal waveform (amplitude ΔH, period of change 2x _0, etc.) on the PC screen. Not needed.

Here, the buried state of the magnetic material includes, for example, the separation distance r from the observation hole to the magnetic material, the magnetic quantity M of the magnetic material, and the direction θ (elevation/depression angle) of the magnetic pole of the magnetic material as seen from the observation hole. It is a combination of

There is a strong correlation between the shape (pattern) of the signal waveform and the buried state of the magnetic material. For example, the shorter the separation distance r between the magnetic bodies, the shorter the period _2x0 of change in the signal waveform, and the shorter the separation distance r between the magnetic bodies and the greater the magnetic amount M of the magnetic bodies, the greater the amplitude ΔH of the signal waveform. This is already known (see calculation formulas (1) and (2) described later). Furthermore, according to experiments conducted by the present inventors, the types of signal waveform shapes (patterns) are MV type, M type, V type, S type, and reverse S type depending on the direction θ of the magnetic pole of the magnetic material as seen from the observation hole. It has also been found that it changes among the following types: type, inverted V type, inverted M type, inverted MV type, etc.

In addition, convolutional neural networks are good at capturing feature points in images, so they can recognize the features of the shape (pattern) of the signal waveform, and detect magnetic material that has a strong correlation with the shape (pattern) of the signal waveform. This is suitable for estimating the buried state.

Therefore, according to the method for generating an estimation model for vertical magnetic survey analysis according to the present invention, a high-performance estimation for estimating the buried state of a magnetic body (for example, a combination of separation distance r, magnetic quantity M, and magnetic pole direction θ) is provided. The effect is that the model can be generated reliably.

Further, according to the vertical magnetic survey analysis system, vertical magnetic survey analysis device, or computer readable program according to the present invention, an estimation model generated by the method for generating an estimation model for vertical magnetic survey analysis according to the present invention is used. Since the buried state of the magnetic material is estimated, the effect is that highly accurate analysis results can always be obtained regardless of the ability of the user (engineer).

FIG. 1 is a schematic configuration diagram illustrating a vertical magnetic exploration system to which an example of the vertical magnetic exploration analysis system according to the present invention is applied. FIG. 2 is a flowchart illustrating the procedure of the vertical magnetic survey method. FIG. 3 is an explanatory diagram illustrating input data and output data of the estimation model for vertical magnetic survey analysis according to the present invention. FIG. 4 is an explanatory diagram illustrating a method for generating an estimation model for vertical magnetic survey analysis according to the present invention. FIG. 5 is a flowchart illustrating the analysis procedure by the vertical magnetic survey analysis apparatus according to the present invention. FIG. 6 is an explanatory diagram illustrating a display screen of the vertical magnetic survey analysis device according to the present invention. FIG. 7 is an explanatory diagram illustrating the display area of the differential signal waveform. FIG. 8 is an explanatory diagram illustrating a display area for analysis results. FIG. 9 is a graph showing the relationship between the separation distance estimated by the estimation model for vertical magnetic survey analysis according to the present invention and the true value (actually measured value). FIG. 10 is a graph showing the relationship between the separation distance estimated by the estimation model for vertical magnetic survey analysis according to the present invention and the separation distance calculated by the calculation formula. FIG. 11 is a configuration diagram illustrating a system using a desktop PC instead of a server. FIG. 12 is a configuration diagram illustrating a system in which AI is installed in a mobile PC. FIG. 13 is a flowchart illustrating the procedure of a vertical magnetic exploration method using AI determination in the excavation process. FIG. 14 is an explanatory diagram illustrating AI determination in the excavation process.

Hereinafter, one embodiment of the method for generating an estimation model for vertical magnetic survey analysis, the vertical magnetic survey analysis system, the vertical magnetic survey analysis device, and the vertical magnetic survey system to which a computer-readable program according to the present invention is applied will be described. .

1. Vertical Magnetic Survey System FIG. 1 is a schematic configuration diagram illustrating a vertical magnetic survey system to which an example of the vertical magnetic survey analysis system according to the present invention is applied.

As shown in FIG. 1, the vertical magnetic exploration system 1 includes a boring machine 10 that excavates vertically downward underground to form a cylindrical observation hole 11, and a cylindrical detector 21 inserted into the observation hole 11. and a vertical magnetic exploration analysis system 2 that performs magnetic exploration.

The vertical magnetic survey analysis system 2 includes a detector 21 inserted into an observation hole 11, a cable 22 connected to the detector 21, a control device 23 connected to the detector 21 via the cable 22, and a control device. An AD converter 24 connected to the output terminal of 23, a mobile PC 25 (an example of a vertical magnetic survey analysis device) connected to the control device 23 via the AD converter 24, and a mobile PC 25 connected to a network 26 such as the Internet. and a server 27.

The detector 21 is a uniaxial differential fluxgate type magnetic sensor in which a pair of magnetic detection coils are arranged coaxially at an interval of, for example, about 0.5 to 2 m, and the differential signals of the pair of magnetic detection coils are It is output from the detector 21. In this differential signal, noise (noise caused by environmental factors such as the earth's magnetic field, vibrations of the detector 21, etc.) of the signals individually output by the pair of magnetic coils cancel each other out. Note that, in addition to the differential signal, the detector 21 can also output a geomagnetic signal that is output by the coil on the tip side alone. Further, the interval between the pair of magnetic detection coils in the detector 21 is appropriately set according to the required width of the search range, the required resolution, and the like.

The control device 23 can associate the differential signal output from the detector 21 scanning the observation hole 21 with the depth of the detector 21 and output it to the AD converter 24 . The AD converter 24 AD converts the differential signal with depth output from the detector 21 and outputs it to the mobile PC 25, and the mobile PC converts the differential signal with depth according to a predetermined program stored in its internal memory. The differential signal is saved in the internal memory in a predetermined format (CSV format).

The control device 23 includes a switch that allows the user (engineer) to turn on/off the power supplied from the battery to the detector 21, and a differential signal output from the detector 21 (and a differential signal output from the tip side coil). a display unit that displays the level of the geomagnetic signal); and an adjustment unit that allows the user (engineer) to switch the sensitivity of the differential signal output from the detector 21 (and the geomagnetic signal output from the tip side coil); It includes an adjustment section for a user (engineer) to adjust the offset of the differential signal output from the detector 21 (and the geomagnetic signal output from the tip side coil). Note that the mobile PC 25 operates according to a predetermined program, monitors the differential signal given from the control device 23 side (or the geomagnetic signal output from the tip side coil), and detects a change in the level above a certain level. (If a ``magnetic anomaly'' appears that indicates the possibility that a magnetic material is buried directly under the detector).

2. Vertical Magnetic Exploration Method FIG. 2 is a flowchart illustrating the procedure of the vertical magnetic exploration method. As shown in Fig. 2, in the vertical magnetic survey method using the vertical magnetic survey analysis system 1, the user (engineer) first installs the boring machine 10 at the survey point (S11), and observes vertically downward to a depth of 1 m. A hole 11 is excavated (S12). Then, the user (engineer) inserts the detector 21 into the observation hole 11, and performs magnetic survey 1 m directly below the bottom of the observation hole 11 (S13). At this time, the user (engineer) checks whether there is a magnetic anomaly based on the differential signal or geomagnetic signal (S14), and if there is a magnetic anomaly (S14YES), the user (engineer) moves to the continuous measurement procedure (S16). However, if there is no magnetic anomaly (S14 NO), the excavation is further carried out by 1 m and magnetic exploration and confirmation are performed (S12 to S14). A series of steps (S12 to S14) from excavation to confirmation are repeated until there is a magnetic anomaly (S14 YES) or until the bottom of the observation hole 11 reaches the planned depth (for example, 11 m) (S15 YES).

After that, continuous measurement is performed by scanning the detector 21 at a constant speed from the bottom (lower end) to the hole entrance (upper end) of the observation hole 11, and the signal waveform of the differential signal output by the detector 21 during that period ( and the signal waveform of the geomagnetic signal) are stored in the internal memory of the mobile PC 25 (S16). Thereafter, an analysis (described later) of the signal waveform of the differential signal is performed (S17).

In addition, if there is a risk of contact with the magnetic material 30 (magnetic anomaly) during excavation (S14YES), take measures such as shifting the position of the borehole (position of observation hole 11) and continue the exploration. You can.

Further, regarding the depth where a magnetic anomaly is observed, the reproducibility of the data may be confirmed by scanning the detector 21 in the vertical direction multiple times.

3. Server The server 27 shown in FIG. 1 includes a storage unit that stores programs and data, a processor that executes the programs, a communication interface that accepts requests from the mobile PC 25 via the network 26, and a user interface that accepts input from the server administrator. It also includes a trained estimation model 28 trained by a predetermined algorithm (here, a convolutional neural network). The trained estimation model 28 is a program corresponding to a recognizer of a convolutional neural network, and stores the connection state of the recognizer in advance as a result of learning. In this embodiment, the learned estimation model 28 is used when estimating the buried state of the magnetic body 30 underground. Hereinafter, the learned estimation model 28 will be simply referred to as the "estimation model 28" or "AI" as necessary.

The processor of the server 27 manages the estimation model 28, and upon receiving differential signal waveform data (image Im) from the mobile PC 25 via the network 26 and the communication interface, inputs the image Im into the estimation model 28. The buried state of the magnetic material 30 underground is estimated.

The processor of the server 27 then provides the estimated buried state of the magnetic body 30 to the mobile PC 25 via the network 26. The buried state of the magnetic body 30 provided by the server 27 is presented to the user (engineer) via the screen of the mobile PC 25 or the like.

4. Estimation Model FIG. 3 is an explanatory diagram illustrating input data and output data of the estimation model for vertical magnetic survey analysis according to the present invention. As shown in FIG. 3, the estimation model 28 of this embodiment is a trained estimation model learned by a convolutional neural network, and is a differential signal waveform of the detector 21 for vertical magnetic survey inserted into the observation hole 11. This is an estimation model (an estimation model for vertical magnetic survey analysis) that estimates the buried state of the magnetic material 30 underground from the data (image Im). Below, as the buried state of the magnetic body 30, the separation distance r from the observation hole 11 to the magnetic body 30, the amount of magnetism M that the magnetic body 30 has, and the direction of the magnetic pole of the magnetic body 30 (elevation/depression angle) as seen from the observation hole 11. ) and θ (see Figure 1).

5. Method for Generating Estimated Model FIG. 4 is an explanatory diagram illustrating a method for generating an estimated model for vertical magnetic survey analysis according to the present invention. This estimation model generation method is mainly executed by the server 27.

The processor of the server 27 supervises the correspondence between the buried state of the magnetic body 30 underground (combination of r, M, and θ) and the image Im of the differential signal waveform generated by the detector 21 for vertical magnetic exploration. An acquisition procedure for acquiring data and a learning procedure for learning the estimated model 28 by a convolutional neural network using the acquired teacher data and storing the learned estimated model 28 are executed.

As shown in FIG. 4, the teacher data is images Im of a plurality of differential signal waveforms with different embedded states of the magnetic body 30 (combinations of r, M, and θ). The images Im of these differential signal waveforms can be acquired using the vertical magnetic exploration system 1 described above or a system with the same specifications. For example, an observation hole 11 with a depth of 11 m is drilled using a boring machine 10, a sample of magnetic material 30 (simulated bomb, iron pipe, etc.) is buried in a position horizontally remote from the observation hole 11, and a sample of the magnetic material 30 is buried from the bottom of the hole to the hole. By scanning the detector 21 at a predetermined speed toward the target, differential signal waveform data (image Im) is acquired. Furthermore, about 400 images Im are acquired while changing the embedded state of the sample of the magnetic material 30 (combinations of r, M, and θ) in about 400 ways. According to experiments conducted by the inventors of the present application, these 400 images Im can be roughly classified into four types depending on the pattern type (S type, V type, M type, MW type) of the differential signal waveform. was completed. Furthermore, when images Im belonging to the same pattern type were compared, differences were found such as being reversed vertically or horizontally, having different vertical or horizontal sizes, or having different noise frequencies or amounts of noise.
Therefore, in this embodiment, approximately 400 images Im that can be classified into four types are subjected to a reversal process in which they are each inverted vertically and horizontally, a scaling process in which the image Im is scaled vertically and horizontally, and an artificial noise reduction process. By applying additional noise processing, the number of training data was increased to approximately 60,000. This makes it possible to increase (expand) the amount of training data without actually measuring it. As a result, it is considered that the estimation model 28 can be easily and reliably improved in performance.

6. Vertical Magnetic Survey Analysis Apparatus FIG. 5 is a flowchart illustrating the analysis procedure by the vertical magnetic survey analysis apparatus according to the present invention. This analysis procedure is executed by the mobile PC 25 according to a program stored in the internal memory of the mobile PC 25.

As shown in FIG. 5, the processor of the mobile PC 25 internally stores CSV data (CSV file) of the differential signal waveform specified by the user (engineer) and CSV data (CSV file) of the corresponding geomagnetic signal waveform. The differential signal waveform and the geomagnetic signal waveform are read from the memory (S171) and displayed on the screen as graphs (S172, symbols A2 and A3 in FIG. 6).

Next, the processor of the mobile PC 25 allows the user (engineer) to specify each point (FIG. 7) of the waveform appearing on the graph (reference symbols A2 and A3 in FIG. 6) (S173, FIG. 7). By this designation, the reference depth d, analysis range AA, period 2x ₀ of the differential signal waveform, maximum amplitude ΔH _max , etc. are specified. Note that the method of specification by the user (engineer) will be described later.

Next, the processor of the mobile PC 25 reads the dimensions such as the period 2x ₀ and the maximum amplitude ΔH _max of the differential signal waveform based on each point specified by the user (engineer), and sends it to the server 27 via the network 26. Send the dimensions (period 2x ₀ , maximum amplitude ΔH _max ). The server 27 calculates the separation distance r' of the magnetic body 30 and the magnetism of the magnetic body 30 by applying the values of the period 2x ₀ and the maximum amplitude ΔH _max to the following calculation formulas (1) and (2) that do not use the estimation model 28. The amount M' is calculated, and the calculated values of the separation distance r' and the magnetic amount M' are transmitted to the mobile PC 25 via the network 25. Then, the mobile PC 25 receives the separation distance r' and the magnetic amount M' via the network 25 (S174).

r' ² = {(x ₀ +0.5) (x ₀ -0.5)} ^2/3 ×{(x ₀ +0.5) ^2/3 +(x ₀ -0.5) ^2/3 }... (1)

M'=ΔH _max (0.5 ² + r' ² ) ^3/2 ×10...(2)

Next, the processor of the mobile PC 25 displays the separation distance r' and the magnetic quantity M' received from the server 27 on the screen (S175, reference numeral A4 in FIG. 6, FIG. 8).

On the other hand, the processor of the mobile PC 25 generates differential signal data belonging to the analysis range AA (FIG. 7) as an image Im, and transmits the generated image Im to the server 27 via the network 26. The server 27 that has received the image Im inputs the image Im into the estimation model 28 to estimate the separation distance r, the amount of magnetism M, and the direction θ of the magnetic poles of the magnetic body 30, and then calculates the separation distance r, via the network 26. The magnetic quantity M and the magnetic pole direction θ are transmitted to the mobile PC 25. Thereby, the mobile PC 25 can acquire the separation distance r of the magnetic body 30, the magnetic amount M, and the direction θ of the magnetic pole (S176).

Next, the processor of the mobile PC 25 displays on the screen the separation distance r, the magnetic quantity M, and the magnetic pole direction θ received from the server 27 (S177, reference numeral A4 in FIG. 6, FIG. 8).
As a result of the above, the user (engineer) can compare the separation distance r' and magnetic quantity M' calculated by the calculation formula with the separation distance r, magnetic quantity M, and magnetic pole direction θ estimated by the estimation model 28. (Figure 8). Note that it is desirable that the depth d is also displayed at this time.

7. Display Screen FIG. 6 is an explanatory diagram illustrating the display screen of the vertical magnetic survey analysis device according to the present invention. This display screen is displayed on the screen of the mobile PC by the processor of the mobile PC 25 according to a program stored in the internal memory of the mobile PC.

At the top of the display screen in FIG. 6, an area A1 is provided for the user (engineer) to specify a file of differential signal waveform data.

To the left of the center of the display screen in FIG. 6, there is an area A2 where the geomagnetic signal waveform generated independently by the tip side coil of the detector 21 is displayed as a graph, and an area A2 where the differential signal waveform generated by the detector 21 is displayed as a graph. area A3 (see FIG. 7) is provided side by side.

On the center right side of the display screen in FIG. 6, it is possible to view the separation distance r, magnetic quantity M, and magnetic pole direction θ estimated by the estimation model 28, and the separation distance r' and magnetic quantity M' calculated by the calculation formula. Area A4 (see FIG. 8) is provided for displaying the status.
At the bottom of the display screen in FIG. 6, an area A5 is provided in which execution logs related to analyzes performed via the server 27 are displayed in chronological order.

8. Differential Signal Waveform Display Area FIG. 7 is an explanatory diagram illustrating the differential signal waveform display area A3. The user (engineer) can specify the upper end depth and lower end depth of the half cycle x ₀ of the differential signal waveform by pointing with a mouse or inputting numerical values while visually observing the display area A3 of the differential signal waveform. can. This allows the mobile PC 25 to specify the period _2x0 of the differential signal waveform.

Additionally, the user (engineer) can specify the tangent line of the peak with the largest amplitude in the differential signal waveform by pointing with a mouse or inputting numerical values while visually viewing the display area A3 of the differential signal waveform. can. This allows the mobile PC 25 to specify the maximum amplitude ΔH _max of the differential signal waveform.

Further, the user (engineer) can specify the reference depth d by pointing with a mouse or inputting a numerical value while visually observing the display area A3 of the differential signal waveform. Note that the reference depth d may be specified by the user (engineer) based on the graph of the geomagnetic signal waveform (symbol A2 in FIG. 6).

Further, the user (engineer) can specify the upper end depth and lower end depth of the analysis range AA by pointing with a mouse or inputting numerical values while visually observing the display area A3 of the differential signal waveform. Note that the mobile PC 25 may automatically define a predetermined range centered on the depth d as the analysis range AA.

6. Display Area for Analysis Results FIG. 8 is an explanatory diagram illustrating the display area A4 for analysis results. As shown in FIG. 8, the analysis result display area A4 includes an area A41 that displays the magnetic pole direction θ, an area A42 that displays the separation distance r, an area A43 that displays the depth d, and the magnetic quantity M. area A44 to display the maximum amplitude ΔH _max , area A51 to display the period 2x0, area A53 to display the separation distance r', area A54 to display the magnetic quantity M', and the magnetic quantity M An area A61 is provided for displaying ".The magnetic quantity M" is calculated by applying the separation distance r estimated by the estimation model 28 and the maximum amplitude ΔH _max mentioned above to the calculation formula (2). This is the magnetic quantity calculated by the processor of the server 27.

9. Evaluation of Estimation Model FIG. 9 is a graph showing the relationship between the separation distance estimated by the estimation model for vertical magnetic survey analysis according to the present invention and the true value (actually measured value). As shown in FIG. 9, although there is some error in the separation distance r estimated by the estimation model 28 when compared with the true value of the separation distance, this error is within a range of approximately ±0.5 m.

FIG. 10 is a graph showing the relationship between the separation distance estimated by the estimation model for vertical magnetic survey analysis according to the present invention and the separation distance calculated by the calculation formula. As shown in FIG. 10, the separation distance r estimated by the estimation model 28 tended to have a slightly larger value than the separation distance r' calculated by the formula. In addition, the separation distance r shows a characteristic distribution depending on the combination of separation distance and weight (i.e. magnetic quantity). For example, a 150mm bomb size has high accuracy at a separation distance of 2m or less, and a 500kg bomb size has high accuracy at a separation distance of 3m or more. The results showed that slightly smaller values were likely to be obtained depending on the distance.

Although only the estimation error of the separation distance r has been verified here, it is considered that the estimation model 28 can also estimate the magnetic quantity M and the direction θ of the magnetic poles with a certain degree of accuracy. This is because it is well known that there is a correlation between the maximum amplitude ΔH _max of the differential signal waveform and the magnetic quantity M of the magnetic body 30, and the pattern type of the differential signal waveform (MV type, M type, V type, This is because it has been found through experiments by the inventors that there is a correlation between the S type, inverted S type, V type, inverted M type, and inverted MV type) and the direction θ of the magnetic pole. (See Figure 4).

10. Effects of Embodiment According to this embodiment, the association between the buried state of the magnetic body 30 (combination of r, M, and θ) and the image Im of the differential signal waveform generated by the detector 21 is convolved as training data. Since the estimation model 28 is trained by a neural network (CNN: Convolution Neural Network), the user (engineer) does not have to manually measure the dimensions of the differential signal waveform (amplitude ΔH, period of change 2x ₀ , etc.) on the screen. Not needed.

There is a strong correlation between the shape (pattern) of the differential signal waveform and the buried state of the magnetic body 30 (combination of r, M, and θ). For example, the shorter the separation distance r between the magnetic bodies 30, the shorter the period _2x0 of change in the differential signal waveform. It is already known that the maximum amplitude ΔH _max also increases (see calculation formulas (1) and (2)). Furthermore, according to the experiments conducted by the inventors of the present application, the pattern type of the differential signal waveform is MV type, M type, V type, S type, and reverse S type depending on the direction θ of the magnetic pole of the magnetic body 30 as seen from the observation hole 11. It has also been found that it varies among the following types: type, inverted V type, inverted M type, and inverted MV type (Figure 4).

Furthermore, convolutional neural networks are good at capturing feature points in images, so they recognize the features of the shape (pattern) of the differential signal waveform as the image Im, and there is a strong correlation with the shape (pattern) of the differential signal waveform. This is suitable for estimating the buried state (combination of r, M, and θ) of the magnetic body 30 having the following properties.

Therefore, according to this embodiment using the learned estimation model 28 using the convolutional neural network, the buried state (combination of r, M, and θ) of the magnetic body 30 can be estimated with high accuracy.

11. Others As described above, the preferred embodiments of the present invention have been described in detail, but the present invention is not limited to such specific embodiments, and within the scope of the gist of the present invention as described in the claims. , it goes without saying that various modifications and changes are possible.

11-1. Additional Learning of AI For example, in the system described above, the estimated model 28 may be updated by additionally learning the estimated model 28 with new training data. By repeating such updates, the performance of the estimation model 28 can be improved over time.

11-2. Assignment of Functions For example, the assignment of functions in the system described above is not limited to that described above, and some or all of the functions installed on the side of the server 27 may be installed on the side of the mobile PC 25. However, some of the functions installed on the mobile PC 25 side may be installed on the server 27 side. Further, some of the functions installed in the mobile PC 25 (such as a function to display estimation results, a function to access a server, etc.) may be installed in a desktop PC (not shown). Further, a part of the functions of the control device 23 and the AD converter 24 may be installed on the mobile PC 25 side, or a part or all of the functions of the AD converter 24 and the mobile PC 25 may be installed on the side of the control device 23. You may. Further, it is also possible to use a PC instead of the server 27, and for example, a desktop PC 127 installed in an office or the like may be used instead of the server 27 (see FIG. 11).

11-3. Mobile PC equipped with AI
Further, as shown in FIG. 12, if the estimation model 28 is installed in the mobile PC 25 in advance, it becomes possible to use the estimation model 28 in the field where the mobile PC 25 cannot be connected to the network 26. In that case, the process of learning the estimation model 28 by a convolutional neural network using training data (the process of generating the estimation model 28) may be executed by the server 27 (FIG. 1) or by an office or the like. It may be executed by the installed desktop PC 127 (FIG. 11). This is because the learned estimation model 28 can be installed into the mobile PC 25 via the network 26. In addition, when the server 27 (FIG. 1) or the desktop PC 127 (FIG. 11) updates the estimation model 28 by additional learning, the updated estimation model 28 is reinstalled on the mobile PC 25 via the network 26 (the program is updated). )do it. Further, the server 27 (FIG. 1) or the desktop PC 127 (FIG. 11) may collect additional teacher data necessary for additional learning via the network 26.

11-4. Judgment directly below using AI In addition, direct below judgment using AI may be performed during the excavation process at the site. FIG. 13 is a flowchart illustrating the procedure of a vertical magnetic exploration method using AI determination in the excavation process. The difference between the flow in FIG. 13 and the flow in FIG. 2 is that steps S101, S102, and S103 are inserted between steps S14 to S16.

In the flow of FIG. 13, if a waveform that seems to be the initial stage of an abnormal waveform is observed during observation every 1 m during excavation (S14YES), it is determined whether or not the magnetic material 30 exists directly under the observation hole 11. Estimation is performed using the estimation model 28 (AI) (S101). If the estimation model 28 estimates that the magnetic material 30 exists directly under the observation hole 11 (S101 YES), the process immediately ends the excavation (S103) and moves to step S16 (S103). If not (S101 NO), the process moves to step S15 to perform further excavation.

In such learning of the estimation model 28, as the buried state of the magnetic material 30 underground, the classification of the area to which the separation distance r and the depth d belong (classifications 1 to 9 in FIG. 14), and the classification of the area for vertical magnetic survey. The correspondence between the differential signal waveform generated by the detector 21 and the image Im may be used as the teacher data. In that case, the learned estimation model 28 can estimate which of classifications 1 to 9 the area where the magnetic body 30 is present is. Therefore, in step S102 described above, for example, if the estimation model 28 estimates whether or not the classification of the area to which the magnetic body 30 belongs is "1", and the estimation result that the classification is "1" is obtained. If so, it may be determined that the magnetic body 30 exists directly under the observation hole 11 (step S102 YES), and if not, it may be determined that the magnetic body 30 does not exist directly under the observation hole 11 (step S102 NO). .

11-4. System update In addition, for example, unexploded ordnance detection involves measuring the presence or absence of unexploded ordnance at every 1 meter depth while drilling, so if the estimation accuracy of the estimation model is low, there is a risk that it will lead to a major accident. For this reason, the ability to quickly estimate without human error using an estimation model is of great value in terms of preventing accidents. Regarding the system of this embodiment, the user interface can also be modified in various ways. Furthermore, in this embodiment, it is possible to devise a tool that can increase its versatility as a tool that can compare results using a calculation formula and results using an estimation model. Furthermore, the estimated model can be sequentially trained (additional learning) using new training data prepared based on comparisons (updating the estimated model), or the estimated model can be further trained using nearby training data with a separation distance of 50 cm or less. By updating the estimation model (updating the estimation model), it is possible to improve the estimation accuracy and expand the scope of application. It is also possible to construct a system that can learn new training data each time it is analyzed.

1 Vertical magnetic survey system 10 Boring machine 2 Vertical magnetic survey analysis system 21 Detector 22 Cable 23 Control device 24 AD converter 25 Mobile PC
27 Server 28 Estimated model 20 Estimated model 26 Network 30 Magnetic body r Separation distance M Magnetic amount θ Direction of magnetic pole

Claims (10)

An acquisition procedure for acquiring as training data the correspondence between the buried state of magnetic materials underground and the image of the signal waveform generated by a detector for vertical magnetic survey inserted into an observation hole in the ground, and a learning procedure for learning an estimation model by a convolutional neural network using training data and storing the learned estimation model; and management for managing the estimation model generated by the estimation model generation method for vertical magnetic exploration analysis. means and
Estimating means for estimating the buried state of the magnetic material underground by inputting an image of a signal waveform generated by a detector for vertical magnetic exploration inserted into an underground observation hole into the estimation model;
A presentation means for presenting the buried state of the magnetic material to the user;
Equipped with
The estimation means calculates the buried state of the magnetic body by applying the dimensions of the signal waveform to a predetermined calculation formula that does not use the estimation model,
The presenting means presents the estimated buried state and the calculated buried state to the user in a state where they can be compared with each other.
A vertical magnetic survey analysis system characterized by: The vertical magnetic survey analysis system according to claim 1,
The buried state of the magnetic material includes:
A vertical magnetic survey analysis system characterized by including a separation distance r from the observation hole to the magnetic body. The vertical magnetic survey analysis system according to claim 2,
The buried state of the magnetic material further includes:
a magnetic quantity M of the magnetic substance;
A vertical magnetic exploration and analysis system comprising at least one of the following: a direction θ of a magnetic pole of the magnetic body as seen from the observation hole. The vertical magnetic survey analysis system according to claim 3,
In the acquisition procedure,
By acquiring images of a plurality of signal waveforms while changing the embedded state of the magnetic sample, and performing at least one of inversion processing, scaling processing, and noise processing on the images of the plurality of signal waveforms, A vertical magnetic survey analysis system characterized by expanded training data. The vertical magnetic survey analysis system according to any one of claims 1 to 4,
The detector includes:
A vertical magnetic survey analysis system characterized by a single-axis differential fluxgate type magnetic sensor. The vertical magnetic survey analysis system according to any one of claims 1 to 5,
A vertical magnetic exploration and analysis system, characterized in that the magnetic material is an unexploded ordnance. The vertical magnetic survey analysis system according to any one of claims 1 to 6 ,
A vertical magnetic exploration analysis system further comprising a detector for vertical magnetic exploration. An acquisition procedure for acquiring as training data the correspondence between the buried state of magnetic materials underground and the image of the signal waveform generated by a detector for vertical magnetic survey inserted into an observation hole in the ground, and a learning procedure for learning an estimation model by a convolutional neural network using training data and storing the learned estimation model; and management for managing the estimation model generated by the estimation model generation method for vertical magnetic exploration analysis. means and
By inputting an image of a signal waveform generated by a vertical magnetic exploration detector inserted into an observation hole being drilled into the estimation model, it is possible to determine whether or not there is a magnetic material directly under the observation hole. an estimation means for estimating the buried state during excavation ;
Presentation means for presenting to a user during excavation whether or not a magnetic substance exists directly under the observation hole being excavated as a buried state of the magnetic substance;
A vertical magnetic survey analysis system characterized by comprising : A vertical magnetic exploration analysis device applied to the vertical magnetic exploration analysis system according to any one of claims 1 to 8 ,
A vertical magnetic survey analysis device comprising the estimating means and the presenting means. A computer-readable program for causing a computer to function as the vertical magnetic survey analysis device according to claim 9 .

Images