Buried object display device, buried object detection device, and program

Abstract

To use an AI detection function and a MG detection function in combination in an underground radar survey to improve accuracy of survey.SOLUTION: An AI detection unit 12 performs an AI detection function that represents processing for detecting an underground range in which a buried object exists by using survey data acquired from a radar survey device 10 and using machine learning related to a shape of the buried object in the survey data. A MG detection unit 13 performs a MG detection function that represents processing for performing migration processing of a reflection signal from the buried object by using the survey data acquired from the radar survey device 10 to detect a position of the buried object. An integration processing unit 14 superimposes the detection result by the AI detection unit 12 and the detection result by the MG detection unit 13 together on one screen so as to display the position of the buried object as the respective detection results.SELECTED DRAWING: Figure 9

Description

[Technical field]

[0001] The present invention relates to a technique for detecting buried objects underground. [Background technology]

[0002] Known technologies for detecting buried objects underground include a function (hereinafter referred to as the MG detection function) that performs a migration process on exploration data obtained by underground radar exploration and detects the position of the buried object based on the strength of the reflected signal from the buried object, as described in Patent Document 1, and a function (hereinafter referred to as the AI ​​detection function) that uses a machine learning algorithm for the shape of the buried object in the above exploration data to detect the underground area in which the buried object exists. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2021-124290 Summary of the Invention [Problem to be solved by the invention]

[0004] While the MG detection function can identify the location of buried objects as point coordinates, it has problems such as time-consuming analysis and false positives. Also, the AI ​​detection function has the advantage that workers can refer to the detection results and easily select objects that should not be judged as buried objects, but since it expresses the underground range where buried objects exist as a rectangular area, it is less accurate than the detection of point coordinates by MG detection.

[0005] The present invention has been made in response to such problems, and aims to improve the exploration accuracy by combining the AI ​​detection function and the MG detection function in underground radar exploration. [Means for solving the problem]

[0006] In order to solve the above problems, the present invention provides a buried object display device that displays the location of a buried object by superimposing the result of detecting the underground range where the buried object exists using machine learning on the shape of the buried object in the exploration data obtained by underground radar exploration and the result of performing migration processing on the buried object reflection signal by the radar exploration. With this configuration, it is possible to superimpose the result of detection using machine learning and the result of migration processing, thereby supporting the identification of the location of the buried object by the worker who views it.

[0007] The present invention also provides a buried object detection device that detects an underground area where a buried object exists by using machine learning on the shape of the buried object in the exploration data obtained by underground radar exploration, and detects point coordinates of the buried object's location by migration processing of the buried object reflection signal in the detected underground area. With this configuration, it is possible to improve the exploration accuracy of the buried object by specifying the location of the buried object by combining detection using machine learning and detection by migration processing.

[0008] In the buried object detection device, migration processing of the buried object reflection signal may be performed only for the underground range detected using the machine learning to detect point coordinates that are the positions of the buried objects. According to this configuration, by performing detection by migration processing only for the underground range detected by detection using machine learning, the time required for processing can be shortened compared to the case where detection by migration processing is performed for the entire search range.

[0009] In the buried object detection device, when a plurality of point coordinates are detected by the migration process in the detected underground range, the point coordinates corresponding to the buried object reflection signal having the largest absolute value may be detected as the position of the buried object. With this configuration, even if a plurality of point coordinates are detected by the migration process in the underground range detected by the detection using machine learning, it is possible to detect any one of the point coordinates as the position of the buried object.

[0010] The buried object detection device may detect, as the position of the buried object, point coordinates that are included in a range near the feature point of the buried object in the exploration data and that correspond to a buried object reflection signal with the largest absolute value. With this configuration, even if a plurality of point coordinates are detected by detection using migration processing in an underground range detected by detection using machine learning, it is possible to detect the most probable point coordinates as the position of the buried object.

[0011] In the buried object detection device, the point coordinates may include coordinates in a direction of movement of a radar exploration device and coordinates in a depth direction underground.

[0012] The present invention also provides a program for causing a computer to detect an underground range in which a buried object exists by using machine learning on the shape of the buried object in the exploration data obtained by underground radar exploration, and to detect point coordinates that are the position of the buried object by migration processing of the buried object reflection signal in the detected underground range. With this configuration, it is possible to improve the accuracy of the exploration of buried objects by specifying the position of the buried object by combining detection using machine learning and detection by migration processing. Effect of the Invention

[0013] According to the present invention, it is possible to improve the exploration accuracy by using both the AI ​​detection function and the MG detection function in underground radar exploration. [Brief description of the drawings]

[0014] [Figure 1] 1 is a diagram showing an example of a device configuration of an underground radar system according to a first embodiment of the present invention; [Diagram 2] 1 is a block diagram showing an example of a hardware configuration of a radar exploration device according to a first embodiment. [Diagram 3] 1 is a block diagram showing an example of a hardware configuration of a data processing device according to a first embodiment. [Figure 4] FIG. 1 is a diagram illustrating a conventional concept of buried object detection. [Diagram 5] FIG. 1 is a diagram showing an example of a buried object detection result using a conventional AI detection function. [Figure 6] FIG. 13 is a diagram showing an example of a buried object detection result using a conventional MG detection function. [Figure 7] FIG. 2 is a diagram illustrating the concept of buried object detection in the first embodiment. [Figure 8] FIG. 4 is a diagram showing an example of a detection result of a buried object in the first embodiment. [Figure 9] 1 is a block diagram showing an example of a functional configuration of a data processing device according to a first embodiment. [Figure 10] 4 is a flowchart showing an example of the operation of the data processing device according to the first embodiment. [Figure 11] 10 is a flowchart showing an example of the operation of a data processing device according to a second embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0015] (First embodiment) FIG. 1 is a diagram showing an example of the configuration of an underground radar system according to a first embodiment of the present invention. As shown in FIG. 1, the underground radar system 1 includes a radar exploration device 10 that irradiates a radar into the ground to generate exploration data, and a data processing device 100 that performs processing to detect buried objects in the ground using the exploration data generated by the radar exploration device 10. The radar exploration device 10 is mounted on a moving object such as an automobile, train, or dolly equipped with wheels Wh, and buried objects such as buried pipes buried in the ground of the exploration target area are explored by the moving object moving on the ground surface of the exploration target area. In this first embodiment, the data processing device 100 functions as a buried object display device of the present invention.

[0016] Fig. 2 is a block diagram showing an example of a hardware configuration of the radar exploration device 10. As shown in Fig. 2, the radar exploration device 10 includes a processor 101, a memory 102, a storage 103, a communication device 104, a transmitter 105, a receiver 106, and a drive mechanism 107.

[0017] The processor 101 is, for example, a CPU (Central Processing Unit).

[0018] The memory 102 is a computer-readable recording medium, and may be composed of at least one of, for example, a Read Only Memory (ROM), an Erasable Programmable ROM (EPROM), an Electrically Erasable Programmable ROM (EEPROM), a Random Access Memory (RAM), etc. The memory 102 may also be called a register, a cache, a main memory (primary storage device), etc. The memory 102 can store a program (program code), software modules, etc. that are executable to implement the method according to the present embodiment.

[0019] The storage 103 is a computer-readable recording medium, and may be configured by, for example, at least one of an optical disk, a hard disk drive, a flexible disk, a magneto-optical disk, a smart card, a flash memory, and the like.

[0020] The communication device 104 is hardware (transmission / reception device) for performing communication between computers via at least one of a wired network and a wireless network, and communicates with the data processing device 100.

[0021] The transmitter 105 includes, for example, a D / A converter, an amplifier, and a transmission antenna, and transmits radar signals toward the ground.

[0022] The receiver 106 includes, for example, an A / D converter, an amplifier, various filters, a receiving antenna, and the like, and receives a reflected signal of the radar transmitted by the transmitter 105 and reflected underground.

[0023] The drive mechanism 107 is a mechanism for moving the mobile body on which the radar exploration device 10 is mounted on the ground surface, and includes various gears, drive sources, etc. in addition to the wheels Wh described above.

[0024] The processor 101 generates exploration data showing the underground state based on information on the radar transmitted by the transmitter 105 and the reflected signal received by the receiver 106, information on the number of revolutions of the wheels Wh, etc. This exploration data includes data for generating image data called a B-scope image, for example, in which the moving direction of the mobile object on which the radar exploration device 10 is mounted is defined as the x direction and the depth direction underground is defined as the y direction.

[0025] 3 is a block diagram showing an example of a hardware configuration of the data processing device 100. The data processing device 100 is configured as a computer device including a processor 1001, a memory 1002, a storage 1003, a communication device 1004, an input / output device 1005, and a bus connecting these devices, as hardware. Each function of the data processing device 100 is realized by loading a predetermined software (program) onto hardware such as the processor 1001 and the memory 1002, causing the processor 1001 to perform calculations, control communication by the communication device 1004, and control at least one of reading and writing data from and to the memory 1002 and the storage 1003. The processor 1001, the memory 1002, the storage 1003, and the communication device 1004 are similar in hardware to the processor 101, the memory 102, the storage 1003, and the communication device 1004 described for the radar exploration device 10, and therefore their description will be omitted.

[0026] The input / output device 1005 includes input devices (eg, keys, microphones, switches, buttons, etc.) that accept input from the outside, and output devices (eg, displays, speakers, LED lamps, etc.) that perform output to the outside.

[0027] The processor 1001 uses the exploration data generated by the radar exploration device 10 to perform a process (hereinafter referred to as AI detection) of detecting the underground area where buried objects are located using machine learning on the shape of buried objects in the exploration data, and also performs a process (hereinafter referred to as MG detection) of detecting the position of the buried object by performing a migration process on the reflected signal from the buried object.

[0028] Here, Fig. 4 is a diagram explaining the concept of conventional buried object detection. Conventionally, a method of using an AI detection function for exploration data and a method of using an MG detection function for exploration data have been implemented separately and independently.

[0029] According to the AI ​​detection function, the underground area where the buried object exists is represented by a rectangular area P, as shown in the B-scope image in Fig. 5. The AI ​​detection function learns the shape of the underground area where the buried object exists on the B-scope image based on a large amount of training data by machine learning in a broad sense including deep learning, and identifies the underground area where the buried object exists using the learned algorithm. In Fig. 5 and Fig. 6 described later, the exploration distance is the distance in the movement direction (x direction) of the mobile object on which the radar exploration device 10 is mounted, and the depth is the distance in the depth direction underground (y direction).

[0030] According to the MG detection function, the position of a buried object is expressed by point coordinates R, as shown in the B-scope image in Fig. 6. The MG detection function performs migration processing of the reflected signal from the buried object, and identifies, in the B-scope image, for example, a portion Q that corresponds to a circular arc near the top end of a buried pipe extending horizontally underground, a position R where the reflected signal intensity is high as the point coordinates indicating the position of the buried object.

[0031] Fig. 7 is a diagram for explaining the concept of buried object detection in this embodiment, in contrast to the conventional concept of buried object detection explained in Fig. 4. In this embodiment, a process is carried out to integrate the detection result of buried object detection performed on exploration data using an AI detection function and the detection result of buried object detection performed on the same exploration data using an MG detection function (hereinafter referred to as an integration process function), and the processing result is output.

[0032] Fig. 8 is a diagram showing an example of the detection result of a buried object in the first embodiment. As shown in Fig. 8, the integrated processing function superimposes the detection result of the buried object detection using the AI ​​detection function and the detection result of the buried object detection using the MG detection function to display the position of the buried object. By simultaneously referring to the detection results of the AI ​​detection function and the MG detection function on the same screen, the worker can easily determine the presence of a buried object.

[0033] For example, in FIG. 8, of the rectangular regions P1 and P2 identified by the AI ​​detection function, the rectangular region P1 includes point coordinates R1, R2, and R3 identified by the MG detection function as the position of the upper end of the portion Q1 corresponding to a part of a circular arc near the upper end of the buried pipe. In contrast, the rectangular region P2 does not include point coordinates identified by the MG detection function. Also, point coordinates R4 and R5 are detected by the MG detection function as the position of the upper end of the portion Q2 corresponding to a part of a circular arc near the upper end of the buried pipe, but are not detected by the AI ​​detection function. In this case, the worker determines that the portion where the rectangular region identified by the AI ​​detection function and the point coordinates identified by the MG detection function overlap is the buried object, and further determines that the point coordinate (point coordinate R2 in FIG. 8) that is considered to be the uppermost point in the portion Q1 corresponding to a part of a circular arc near the upper end of the buried pipe, etc., among the point coordinates detected for the buried object, is the position of the buried object.

[0034] Fig. 9 is a block diagram showing an example of a functional configuration of the data processing device 100 according to the first embodiment. As shown in Fig. 9, the data processing device 100 realizes each function such as an exploration data acquisition unit 11, an AI detection unit 12, an MG detection unit 13, and an integration processing unit 14.

[0035] The exploration data acquisition unit 11 acquires exploration data from the radar exploration device 10.

[0036] The AI ​​detection unit 12 uses its AI detection function, i.e., the exploration data acquired from the radar exploration device 10, to perform a process of detecting the underground area where buried objects exist as a rectangular area using machine learning on the shape of buried objects in the exploration data.

[0037] The MG detection unit 13 has an MG detection function, that is, performs a process of performing migration processing of a reflected signal from a buried object using the exploration data acquired from the radar exploration device 10 to detect the position of the buried object as point coordinates.

[0038] The integrated processing unit 14 displays the positions of buried objects, which are the detection results of the AI ​​detection unit 12 and the detection results of the MG detection unit 13, by overlaying them on one screen.

[0039] Fig. 10 is a flowchart showing an example of the operation of the data processing device 100. In Fig. 10, the processor 1001 (prospecting data acquiring unit 11) acquires the prospecting data from the radar prospecting device 10 (step S1).

[0040] Next, the processor 1001 (AI detection unit 12) performs AI detection using the acquired exploration data (step S2).

[0041] Next, the processor 1001 (MG detection unit 13) performs MG detection using the acquired exploration data (step S3).

[0042] Then, the processor 1001 (integration processing unit 14) displays the detection result by the AI ​​detection unit 12 and the detection result by the MG detection unit 13 in an overlapping manner on one screen (step S4).

[0043] According to the first embodiment described above, the results of AI detection and the results of MG processing are displayed in an overlapping manner, thereby assisting workers viewing the results in identifying the location of buried objects, thereby making it possible to improve the exploration accuracy.

[0044] Second embodiment Next, a second embodiment of the present invention will be described. In the second embodiment, a data processing device 100 functions as a buried object detection device of the present invention. The hardware configuration and functional configuration of the entire system and each device in the second embodiment are similar to those in the first embodiment. The second embodiment differs from the first embodiment in that the data processing device 100 judges the position of a buried object, whereas in the first embodiment, the worker judges the position of the buried object.

[0045] Fig. 11 is a flowchart showing an example of the operation of the data processing device. In Fig. 11, the processor 1001 (prospecting data acquiring unit 11) acquires the prospecting data from the radar prospecting device 10 (step S11).

[0046] Next, the processor 1001 (AI detection unit 12) performs AI detection using the acquired exploration data (step S12).

[0047] Next, the processor 1001 (MG detection unit 13) performs MG detection using the acquired exploration data (step S13). .

[0048] The processor 1001 (integration processing unit 14) determines whether or not one or more point coordinates detected by MG detection are included in a certain rectangular area detected by AI detection (step S14).

[0049] When the processor 1001 (integration processing unit 14) determines that one or more point coordinates detected by MG detection are included in a rectangular area detected by AI detection (step S14; YES), it refers to the absolute values ​​of the intensity of the reflected signal corresponding to each of those point coordinates in the exploration data (step S15).

[0050] Next, the processor 1001 (integration processing unit 14) judges whether the position of the point coordinate that is the maximum absolute value of the intensity of each reflected signal is within a predetermined distance from a characteristic point in the shape of the buried object (for example, in the case of a buried pipe, the upper end of a portion Q that corresponds to a partial arc near the upper end of the buried pipe) (step S16). This judgment can be realized by machine learning using teacher data prepared in advance to determine, for example, which position corresponds to the uppermost end of a portion Q that corresponds to a partial arc near the upper end of a buried pipe in a B-scope image.

[0051] When the processor 1001 (integration processing unit 14) determines that the point coordinate that is the maximum absolute value of the intensity of the reflected signal is within a predetermined distance range from the characteristic point of the buried object (step S16; YES), it identifies the position of the point coordinate as the position of the buried object (step S17).

[0052] The processor 1001 (integration processing unit 14) continues the above-mentioned processing of steps S14 to S17 until all detection results of AI detection and MG detection are inspected (step S18; NO). Note that, if the determination is negative in each of steps S14 and S17 (step S14; NO, step S17; NO), the processor 1001 (integration processing unit 14) excludes the detection result to be processed and proceeds to processing the next detection result (step S19).

[0053] According to the second embodiment, the position of a buried object is detected using the results of AI detection and MG processing, so that it is possible to improve the exploration accuracy.

[0054] <Modification> The above is a description of the embodiment, but the contents of this embodiment may be modified as follows. In addition, the following modifications may be combined.

[0055] <Variation 1> In the above embodiment, the radar exploration device 10 and the data processing device 100 have been described as being separate entities, but they may be integrated, for example, the data processing device 100 may be mounted on the radar exploration device 10. Also, in the above embodiment, the radar exploration device 10 may have a drive mechanism 107 to be self-propelled, or may be moved manually by an operator.

[0056] <Variation 2> In the second embodiment, the processor 1001 (AI detection unit 12) detects an underground range in which buried objects exist by using machine learning on the shape of buried objects in the exploration data obtained by underground radar exploration, and the processor 1001 (MG detection unit 13) detects point coordinates of the buried object's location by migration processing of the buried object reflection signal in the detected underground range. In other words, the processor 1001 (AI detection unit 12 and MG detection unit 13) performed AI detection and MG detection for the entire range of the acquired exploration data. Instead of this, the processor 1001 (MG detection unit 13) may perform migration processing of the buried object reflection signal only for a rectangular area of ​​the underground range detected by the processor 1001 (AI detection unit 12) to detect point coordinates of the buried object's location. In this way, the processing time and processing load of the processor 1001 (MG detection unit 13) can be reduced.

[0057] <Modification 3> The feature point of the buried object is not limited to the upper end of the buried pipe described in the second embodiment, but may be determined appropriately depending on the shape of the buried object.

[0058] <Modification 4> The present invention may be a program written to cause a computer to execute each step shown in Fig. 11. That is, this program causes a computer to detect an underground range in which a buried object exists by using machine learning on the shape of the buried object in exploration data obtained by underground radar exploration, and to detect point coordinates that are the position of the buried object by performing migration processing on the buried object reflection signal in the detected underground range. [Explanation of symbols]

[0059] 10...radar exploration device, 11...exploration data acquisition unit, 12...AI detection unit, 13...MG detection unit, 14...integrated processing unit, 100...data processing device, 101...processor, 102...memory, 103...storage, 104...communication device, 105...transmitter, 106...receiver, 107...drive mechanism, 1001...processor, 1002...memory, 1003...storage, 1004...communication device, 1005...input / output device, P, P1, P2...rectangular area, Q, Q1, Q2...part, R, R1, R2, R3, R4, R5...point coordinates, Wh...wheel.

Claims

1. A buried object display device that displays the location of a buried object by overlaying the results of detecting the underground area where the buried object exists using machine learning on the shape of the buried object from exploration data obtained by underground radar exploration and the results of migration processing of the buried object reflection signal from the radar exploration.

2. A buried object detection device that uses machine learning on the shape of buried objects from exploration data obtained by underground radar exploration to detect the underground area in which the buried objects exist, and detects the point coordinates of the buried object's location by migration processing of the buried object reflection signals in the detected underground area.

3. The buried object detection device according to claim 2 , further comprising: a migration process for the buried object reflection signal only for the underground range detected using the machine learning; and detecting point coordinates that are the position of the buried object.

4. 3. The buried object detection device according to claim 2, wherein when a plurality of point coordinates are detected by the migration process in the detected underground range, the point coordinates corresponding to the buried object reflection signal having the largest absolute value are detected as the position of the buried object.

5. 5. The buried object detection device according to claim 4, wherein the coordinates of a point included in the search data in the vicinity of the characteristic point of the buried object and corresponding to a buried object reflection signal with the largest absolute value are detected as the position of the buried object.

6. 3. The buried object detection device according to claim 2, wherein the point coordinates include coordinates in a moving direction of the radar exploration device and coordinates in a depth direction underground.

7. On the computer, A program for detecting the underground area in which a buried object exists by using machine learning on the shape of the buried object from exploration data obtained by underground radar exploration, and detecting the point coordinates of the location of the buried object by performing migration processing on the buried object reflection signal in the detected underground area.

Images