Rock estimation system, learning device, rock estimation method, learning model generation method, and terminal device

Abstract

To provide a rock estimation system capable of estimating properties of a rock by a method different from conventional methods.SOLUTION: A rock estimation system (100, 100A) comprises an acquisition unit (50) configured to acquire tactile information of a rock, and an estimation unit (60) configured to estimate properties of the rock based on the tactile information.SELECTED DRAWING: Figure 4

Description

[Technical Field]

[0001] The present invention relates to a rock type estimation system, a learning device, a rock type estimation method, a learning model generation method, and a terminal device. [Background technology]

[0002] Conventionally, techniques for estimating rock properties based on visual information are known. For example, an application is known that estimates the type of rock from image data of the rock obtained by a smartphone camera. Patent Document 1 discloses a technique for evaluating the condition of a tunnel face through image capture or temperature detection. Also known is a technique for estimating rock properties based on auditory information (e.g., sound produced by tapping). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 7304781 Summary of the Invention [Problem to be solved by the invention]

[0004] However, visual and auditory information may not be sufficient to fully evaluate rock properties. Visual information is often affected by factors such as the environment at the time of photography, making it insufficient for evaluating rock properties. Furthermore, auditory information, such as sound, is transient, making it difficult to handle as data.

[0005] One aspect of the present invention has been made in consideration of the above problems, and its purpose is to realize a rock estimation system that can estimate rock properties using a method different from conventional methods. [Means for solving the problem]

[0006] In order to solve the above problem, the rock estimation system according to aspect 1 of the present invention comprises an acquisition unit that acquires tactile information of a rock, and an estimation unit that estimates the characteristics of the rock based on the tactile information.

[0007] In a rock estimation system according to aspect 2 of the present invention, in aspect 1 above, the acquisition unit may further acquire visual information of the rock, and the estimation unit may estimate the characteristics of the rock based on the tactile information and the visual information.

[0008] In the rock estimation system according to a third aspect of the present invention, in the first or second aspect, the estimation unit may estimate the type of the rock as the property of the rock.

[0009] The rock estimation system according to a fourth aspect of the present invention may be any of the first to third aspects described above, further comprising a display unit that displays the estimated rock type or the estimation rate of the rock type.

[0010] In the rock estimation system according to a fifth aspect of the present invention, in any of the first to fourth aspects, the estimation unit may estimate a state of the rock as the property of the rock.

[0011] In the rock estimation system according to aspect 6 of the present invention, in aspects 1 to 5 above, the tactile information may include information indicating vibration of the contact body while the contact body is tracing the surface of the rock.

[0012] In a rock estimation system according to aspect 7 of the present invention, in aspect 6 above, the acquisition unit may further acquire a pressing force that presses the contact body against the rock while tracing the surface of the rock, and the estimation unit may estimate the characteristics of the rock based on the tactile information corrected based on the pressing force.

[0013] In order to solve the above problem, a learning device according to aspect 8 of the present invention generates a learning model using training data in which tactile information about rocks is input and information indicating the characteristics of the rocks is output.

[0014] In order to solve the above problem, a rock estimation method according to aspect 9 of the present invention includes an acquisition step of acquiring tactile information of a rock, and an estimation step of estimating the characteristics of the rock based on the tactile information.

[0015] In a rock estimation method according to aspect 10 of the present invention, in the above aspect 9, the acquisition step may further acquire visual information of the rock, and the estimation step may estimate the characteristics of the rock based on the tactile information and the visual information.

[0016] In order to solve the above problem, the learning model generation method according to aspect 11 of the present invention includes an acquisition step of acquiring training data in which tactile information of a rock is input and information indicating the characteristics of the rock is output, and a learning step of generating a learning model using the training data.

[0017] In order to solve the above problem, a terminal device according to aspect 12 of the present invention comprises an acquisition unit that acquires tactile information of a rock, and a communication unit that transmits the tactile information to a server and receives information from the server indicating the characteristics of the rock estimated based on the tactile information. [Effects of the Invention]

[0018] According to one aspect of the present invention, rock properties can be estimated using a method different from conventional methods. [Brief explanation of the drawings]

[0019] [Figure 1] 10 is a graph showing an example of tactile information obtained by the tactile unit. [Figure 2] 10 is a graph showing another example of tactile information obtained by the tactile unit. [Figure 3] 10 is a graph showing yet another example of tactile information obtained by the tactile unit. [Figure 4] 1 is a block diagram showing an example of the configuration of a rock estimation system according to a first embodiment. [Figure 5]3 is a flowchart showing the flow of processing executed by the rock estimation system according to the first embodiment. [Figure 6] FIG. 10 is a diagram showing an example of a display screen displayed by a display unit according to the second embodiment. [Figure 7] FIG. 10 is a diagram showing another example of the display screen displayed by the display unit according to the second embodiment. [Figure 8] FIG. 10 is a diagram showing yet another example of the display screen displayed by the display unit according to the second embodiment. [Figure 9] FIG. 10 is a diagram showing yet another example of the display screen displayed by the display unit according to the second embodiment. [Figure 10] FIG. 10 is a diagram showing yet another example of the display screen displayed by the display unit according to the second embodiment. [Figure 11] FIG. 10 is a diagram showing yet another example of the display screen displayed by the display unit according to the second embodiment. [Figure 12] FIG. 10 is a block diagram showing an example of the configuration of a rock estimation system according to a modified example. DETAILED DESCRIPTION OF THE INVENTION

[0020] [Embodiment 1] (Features of one aspect of the present invention) In recent years, in order to address the decline in the number of skilled geological engineers, the digitalization of technology for estimating rock properties based on the intuition of skilled engineers has been raised as a challenge. In particular, understanding the condition of the tunnel face, such as the condition of cracks (roughness, smoothness, etc.) and the spacing of cracks, is important in terms of the safety of tunnel excavation work.

[0021] The inventors have focused on the fact that skilled engineers grasp the properties of rocks not only visually but also by touch, and have discovered that by accumulating tactile information about rocks as digital data, it is possible to achieve the same level of skill as a skilled engineer. The rock quality estimation system 100 described below was conceived based on this knowledge and is a system for estimating rock properties using a method that is different from conventional methods.

[0022] Here, tactile information refers to sensing data corresponding to information obtained by human touch. Tactile information is digital data representing the characteristics of Meissner's corpuscles and / or Pacinian corpuscles, which are human tactile sensory receptors. Tactile information includes the texture (feel, feel) of an object. Information indicating the texture of an object is obtained by tracing the object with a contact object (a human finger, a human hand, or a probe).

[0023] Information corresponding to the texture of an object is obtained as information indicating the vibration of the contact object while the contact object is being traced over the object (hereinafter referred to as vibration data). The vibration data of the contact object is detected, for example, by a skin vibration detection sensor as time-series data of voltage or current corresponding to the vibration of the contact object. Figures 1 to 3 are graphs showing tactile information of rocks detected by such a skin vibration detection sensor. Figures 1 to 3 show tactile information of andesite, gabbro, and diorite, respectively. In the graphs shown in Figures 1 to 3, the vertical axis represents the voltage value (V) corresponding to the vibration of the contact object, and the horizontal axis represents time (s). The tactile information of rocks shown in Figures 1 to 3 is data obtained when the surface of the rock is traced circularly multiple times at a predetermined speed.

[0024] As shown in Figures 1 to 3, the amplitude of the voltage waveform in andesite is smaller than that in gabbro or diorite. The amplitude of the voltage waveform in andesite changes less than that in gabbro or diorite. Furthermore, the amplitude of the voltage waveform in gabbro shows convex portions of different sizes (areas surrounded by dashed lines in Figure 2), while the amplitude of the voltage waveform in diorite shows convex portions of similar sizes (areas surrounded by dashed lines in Figure 3). As such, since the tactile information of rocks differs depending on the type of rock, it can be seen that the type of rock can be estimated by converting the tactile information of rocks into digital data. Similarly, the condition of the rock can also be estimated by converting the tactile information of rocks into digital data.

[0025] (Schematic configuration of rock estimation system 100) Fig. 4 is a block diagram showing an example of the configuration of the rock estimation system 100. As shown in Fig. 4, the rock estimation system 100 includes an acquisition unit 50 and an estimation unit 60 (learning device). The rock estimation system 100 may include a display unit 70. The acquisition unit 50 includes a tactile unit 1. The tactile unit 1 is a tactile sensor or the like. The acquisition unit 50 may include an imaging unit 2. The imaging unit 2 is a camera or the like.

[0026] The tactile unit 1 detects tactile information of the rock. The estimation unit 60 estimates the properties of the rock based on the tactile information (e.g., vibration data) of the rock acquired by the tactile unit 1. The imaging unit 2 detects visual information (e.g., image data) of the rock. The estimation unit 60 estimates the properties of the rock based on the visual information of the rock acquired by the imaging unit 2. The display unit 70 displays information indicating the properties of the rock estimated by the estimation unit 60 (hereinafter referred to as the estimation result).

[0027] In this embodiment, the tactile unit 1 obtains information indicating the rock texture, which is one type of tactile information about rocks. The tactile unit 1 obtains information indicating the rock texture by detecting vibrations of a contacting object. The tactile sensor of the tactile unit 1 is, for example, a skin vibration detection sensor that detects voltage or current corresponding to the vibrations of the contacting object. For example, the "Finger Recorder" by Tech Gihan Co., Ltd. may be used as the tactile sensor of the tactile unit 1.

[0028] The tactile unit 1 may be provided on a person's finger or hand, or on a probe. That is, the acquisition unit 50, estimation unit 60, and display unit 70 in the rock estimation system 100 may be separate and distinct. For example, when the rock estimation system 100 is used to evaluate a tunnel face, the tactile unit 1 is installed on a probe mounted on a mobile robot (e.g., a drone). Similarly, the imaging unit 2 is also mounted on the mobile robot.

[0029] The acquisition unit 50 may further include a pressing unit 5. The pressing unit 5 is a pressing force sensor or the like that measures the force (hereinafter referred to as pressing force) with which the contact body is pressed against the rock while tracing the surface of the rock. The pressing force is used for preprocessing in the preprocessing unit 12, which will be described later. Note that if the pressing force falls within a predetermined range, the sensor may be omitted. For example, if the pressing force of the contact body is outside the predetermined range, the pressing force can be kept within the predetermined range by outputting an error without using the obtained tactile information of the rock.

[0030] The estimation unit 60 includes a preprocessing unit 12 , a processing unit 13 , and a learning unit 15 .

[0031] The pre-processing unit 12 performs pre-processing on the tactile information and / or visual information of rocks acquired by the acquisition unit 50. That is, the pre-processing unit 12 converts the tactile information and / or visual information of rocks acquired by the acquisition unit 50 into a suitable format for use in estimating rocks.

[0032] For example, the pre-processing unit 12 normalizes the vibration data based on the pressing force. Specifically, the pre-processing unit 12 corrects the vibration data acquired by the acquiring unit 50 to vibration data at a reference pressing force based on the pressing force acquired by the acquiring unit 50. The pre-processing unit 12 may, for example, store in advance a correlation between pressing force and vibration magnitude. Similarly, the pre-processing unit 12 may normalize the time axis of the vibration data based on the speed at which the contact body traces the surface of the rock.

[0033] The pre-processing unit 12 may also perform a Fourier transform on the vibration data in the time domain (time series data of voltage or current) to generate vibration data in the frequency domain. The pre-processing unit 12 may extract multiple peak frequencies and their intensities from the vibration data in the frequency domain as tactile information of the rock.

[0034] The processing unit 13 estimates rock properties based on the tactile information and / or visual information of the rock preprocessed by the preprocessing unit 12. The processing unit 13 may estimate the type of rock as the rock property. The processing unit 13 may estimate the condition of the rock as the rock property. The condition of the rock refers to the brittleness, ease of crumbling, etc. The condition of the rock may also be the texture of the rock, such as roughness or mirror-like surface. For example, when the rock estimation system 100 is used to evaluate a tunnel face, the processing unit 13 estimates the type of rock that constitutes the face, as well as the degree of cracks or deterioration. The preprocessing unit 12 and the processing unit 13 may be a single component.

[0035] The processing unit 13 acquires the trained learning model 16 from the learning unit 15. The learning model 16 is a learning model that receives tactile information and / or visual information about rocks as input and outputs information indicating the characteristics of the rocks. The information indicating the characteristics of the rocks includes the type of rock, an estimated value of the type of rock, a quantified degree of brittleness of the rock, etc. The processing unit 13 inputs the tactile information and / or visual information about the rocks that has been preprocessed by the preprocessing unit 12 into the learning model 16 and obtains information indicating the characteristics of the rocks as output.

[0036] The display unit 70 displays the estimation result estimated by the processing unit 13. A system for controlling the display of the estimation result may be included in the display unit 70 or the estimation unit 60.

[0037] The learning unit 15 generates a learning model 16 using training data that takes tactile and / or visual information about rocks as input and outputs information indicating the rock's characteristics. The tactile and / or visual information about rocks included in the training data is obtained by sensing multiple rocks with known characteristics using the tactile unit 1 and / or the imaging unit 2. Alternatively, such tactile and / or visual information may be stored as a database. A learning algorithm that can be used is, for example, a convolutional neural network (CNN), but is not limited to this. Any known learning algorithm can be used. The learning unit 15 learns the rock's texture, such as roughness or mirror-like texture, as an output data item in the training data, thereby improving the estimation accuracy of the rock's condition, such as its brittleness and fragility.

[0038] (Example of operation of rock estimation system 100) 5 is a flowchart showing the flow of processing executed by the rock type estimation system 100. An example of the operation of the rock type estimation system 100 will be described below with reference to FIG.

[0039] In S1, the rock estimation system 100 acquires tactile information about the rock from the tactile unit 1 (acquisition step). In S2, the rock estimation system 100 acquires visual information about the rock from the imaging unit 2. In S3, the rock estimation system 100 performs preprocessing on the tactile information and visual information about the rock. Here, S1 and S2 may be performed simultaneously, or S2 may be performed before S1. Furthermore, if only tactile information about the rock is used to estimate the rock, only S1 may be performed and S2 may be omitted. If only visual information about the rock is used to estimate the rock, only S2 may be performed and S1 may be omitted.

[0040] In S4, the rock estimation system 100 estimates the properties of the rock based on the preprocessed tactile information and / or visual information of the rock (estimation step). Specifically, the rock estimation system 100 inputs the preprocessed tactile information and / or visual information of the rock to the learning model 16, and obtains information indicating the properties of the rock as an output from the learning model 16.

[0041] In S5, the rock estimation system 100 displays the estimation result on the display unit 70.

[0042] (Action and effect) The above configuration allows us to acquire tactile information about rocks and estimate their properties based on that information. This allows us to replicate the rock property estimation techniques based on the intuition of skilled engineers, dramatically improving the accuracy of rock property estimation.

[0043] Furthermore, by estimating the characteristics of a rock based on a combination of tactile and visual information about the rock, the accuracy of estimating the characteristics of the rock can be further improved compared to estimating the identity of a rock based solely on tactile information about the rock, or estimating the identity of a rock based solely on visual information about the rock.

[0044] Furthermore, by acquiring vibration data from the contact body as tactile information about the rock, the characteristics of the rock can be estimated based on the texture of the rock's surface. This makes it possible to reproduce an estimation of the rock based on the tactile information (such as roughness or smoothness) that a skilled technician feels when running his or her hand over the rock. Furthermore, since the state of cracks in the rock is also reflected in the vibration data, it is also possible to estimate whether or not the rock has cracks.

[0045] Furthermore, by correcting the tactile information based on the pressure, the properties of the rock can be estimated appropriately even when there is variation in the pressure of the contacting body.

[0046] Furthermore, with the above configuration, the condition of the tunnel face at the tunnel excavation site can be estimated based on the texture of the rock mass, thereby improving the safety of tunnel excavation work. Furthermore, by estimating the condition of the bedrock in landslide areas, landslides can be prevented in advance. These effects also contribute to achieving, for example, Goal 11 of the United Nations' Sustainable Development Goals (SDGs), "Make cities and human settlements inclusive, safe, resilient, and sustainable."

[0047] [Embodiment 2] Other embodiments of the present invention will be described below. For ease of explanation, the same reference numerals will be used to designate components having the same functions as those described in the above embodiment, and the description thereof will not be repeated.

[0048] 6 to 11 are diagrams showing display screens displayed by the display unit 70. In this embodiment, various examples of the display screen will be described.

[0049] The estimation unit 60 may generate an estimation result (referred to as a first estimation result) of the rock properties based only on the tactile information of the rock. For example, the estimation unit 60 calculates an estimation rate for each type of rock included in the first estimation result. Then, as shown in FIG. 6, the display unit 70 displays the calculated estimation rate. The estimation rate may be displayed for one or more types of rock. Furthermore, as shown in FIG. 7, the display unit 70 may display the rock type with the highest estimation rate as the identification result. Alternatively, the display unit 70 may display only the identification result without displaying the estimation rate.

[0050] The estimation unit 60 may also generate an estimation result of rock properties based solely on tactile information about the rock and an estimation result of rock properties based solely on visual information about the rock (referred to as a second estimation result). For example, the estimation unit 60 calculates an estimation rate for each type of rock included in the first estimation result and an estimation rate for each type of rock included in the second estimation result. As shown in FIG. 8, the display unit 70 then displays the calculated estimation rates. As shown in FIG. 9, the display unit 70 may also display images captured by the image capture unit 2. As shown in FIG. 8, the first estimation result and the second estimation result may differ. This is because the second estimation result based on visual information is affected by factors such as the environment at the time of capture. As shown in FIG. 8, knowing the tactile estimation rate (the estimation rate included in the first estimation result) and the visual estimation rate (the estimation rate included in the second estimation result) allows the user to determine the type of rock based on both values.

[0051] The estimation unit 60 may also generate a third estimation result based on the first estimation result and the second estimation result. For example, the estimation unit 60 calculates the third estimation result as the sum of the estimation rate for each rock type contained in the first estimation result weighted by a first weighting coefficient and the estimation rate for each rock type contained in the second estimation result weighted by a second weighting coefficient. The display unit 70 then displays the rock type identified based on the third estimation result, as shown in FIG. 10 . The display unit 70 may also display an image captured by the image capture unit 2, as shown in FIG. 11 . Alternatively, the display unit 70 may display only the identification result without displaying the estimation rate and the image captured by the image capture unit 2. This configuration allows the user to appropriately determine the rock type even when the first estimation result and the second estimation result differ.

[0052] The display unit 70 may also display a graph showing vibration data of tactile information as shown in Figures 1 to 3. The display unit 70 may also display the vibration data and an image captured by the image capturing unit 2.

[0053] As described above, the display unit 70 may display tactile information and visual information of the rock that has been preprocessed by the preprocessing unit 12. The user can estimate the properties of the rock based on the tactile information and visual information of the rock displayed on the display unit 70.

[0054] (Variation 1) Fig. 12 is a block diagram showing an example of the configuration of a rock estimation system 100A according to a modified example. As shown in Fig. 12, the acquisition unit 50 and display unit 70 of the rock estimation system 100A are present on a terminal device 20 (user terminal), and the estimation unit 60 is present on a server 30. The terminal device 20 includes the acquisition unit 50 and a communication unit 22. The terminal device 20 may also include a display unit 70. The acquisition unit 50 includes a tactile unit 1. The acquisition unit 50 may also include an imaging unit 2 and a pressing unit 5.

[0055] The tactile unit 1 of the acquisition unit 50 acquires tactile information of the rock. The imaging unit 2 of the acquisition unit 50 acquires visual information of the rock. The communication unit 22 transmits the tactile information of the rock to a server and receives an estimation result of the rock's properties from the server. The communication unit 22 may transmit the visual information of the rock to the server and receive an estimation result of the rock's properties from the server.

[0056] That is, in the rock estimation system 100A, sensing data acquired on the terminal device 20 side is sent to the server 30, and rock estimation is performed on the server 30. Then, the server 30 sends the estimation results of the rock properties to the terminal device 20, and the estimation results are displayed on the display unit 70.

[0057] (Variation 2) The rock estimation system 100 may also estimate rock properties based only on tactile information about the rock. In this case, the learning model 16 is generated using training data that takes tactile information about the rock as input and outputs information indicating the rock properties.

[0058] The rock estimation system 100 may also estimate rock properties based on the auditory information of the rock in addition to the tactile information of the rock. In this case, the learning model 16 is generated using training data that takes the tactile information and auditory information of the rock as input and outputs information indicating the rock properties.

[0059] [Software implementation example] The functions of the rock estimation system 100 (hereinafter referred to as the "device") can be realized by a program that causes a computer to function as the device, and a program that causes a computer to function as each control block of the device (particularly the pre-processing unit 12, the processing unit 13, and the learning unit 15).

[0060] In this case, the device includes a computer having at least one control device (e.g., a processor) and at least one storage device (e.g., a memory) as hardware for executing the program. The control device and storage device execute the program, thereby realizing the functions described in each of the above embodiments.

[0061] The program may be non-transitory and may be recorded on one or more computer-readable recording media. The recording media may or may not be included in the device. In the latter case, the program may be supplied to the device via any wired or wireless transmission medium.

[0062] Furthermore, some or all of the functions of the control blocks can be realized by logic circuits. For example, integrated circuits in which logic circuits functioning as the control blocks are formed are also included in the scope of the present invention. In addition, the functions of the control blocks can also be realized by, for example, a quantum computer.

[0063] The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope of the claims. Embodiments obtained by appropriately combining the technical means disclosed in different embodiments are also included in the technical scope of the present invention. [Explanation of symbols]

[0064] 1. Tactile part 2. Filming Department 5 Pressing part 12 Pretreatment section 13 Processing section 15 Learning Department 16 Learning Model 20 Terminal equipment (user terminal) 22 Communications Department 30 servers 50 Acquisition Department 60 Estimation part 70 Display section 100, 100A Rock Estimation System

Claims

1. an acquisition unit for acquiring tactile information of the rock; and an estimation unit that estimates the properties of the rock based on the tactile information.

2. The acquisition unit further acquires visual information of the rock, The rock estimation system according to claim 1 , wherein the estimation unit estimates the properties of the rock based on the tactile information and the visual information.

3. The rock estimation system according to claim 1 or 2, wherein the estimation unit estimates the type of the rock as the property of the rock.

4. The rock estimation system according to claim 3 , further comprising a display unit that displays the estimated rock type or the estimation rate of the rock type.

5. The rock estimation system according to claim 1 or 2, wherein the estimation unit estimates a state of the rock as the property of the rock.

6. The rock estimation system according to claim 1 or 2, wherein the tactile information includes information indicating vibration of the contact body while the contact body is tracing the surface of the rock.

7. The acquisition unit further acquires a pressing force that presses the contact body against the rock while tracing the surface of the rock, The rock estimation system according to claim 6 , wherein the estimation unit estimates the rock characteristics based on the tactile information corrected based on the pressing force.

8. A learning device that generates a learning model using training data in which tactile information about rocks is input and information indicating the characteristics of the rocks is output.

9. an acquisition step of acquiring tactile information of the rock; and estimating a property of the rock based on the tactile information.

10. The acquiring step further acquires visual information of the rock; The rock estimation method according to claim 9 , wherein the estimation step estimates the properties of the rock based on the tactile information and the visual information.

11. an acquisition step of acquiring training data in which tactile information of a rock is input and information indicating the characteristics of the rock is output; A learning model generation method including a learning step of generating a learning model using the training data.

12. an acquisition unit for acquiring tactile information of the rock; a communication unit that transmits the tactile information to a server and receives information indicating the rock characteristics estimated based on the tactile information from the server.

Images