JP7292176B2 - SOIL PROPERTIES ESTIMATION METHOD, DEVICE, PROGRAM AND SYSTEM - Google Patents

Description

The present invention relates to technology for estimating properties of soil materials.

In order to reuse soil and rock materials generated by excavation work and the like, quality control is performed by grasping their particle size distribution. Particle size distribution is generally examined by sieving tests. In Patent Document 1, as a method for sorting generated soil with large variations in particle size for each particle size composition in a short time, generated soil and water are put into a container, and the container is well stirred to wash sand in the mud. A method is disclosed in which the volume of sedimented sand is measured after separation, and the mass of coarse-grained soil is determined from the void ratio to obtain the fine-grained content in the generated soil.

Further, Patent Document 2 describes determining the properties of sampled earth and sand using a particle size addition curve. Further, Patent Document 3 discloses a method for reducing the time required to specify the grain size of soil and the equipment cost. The method described in Patent Literature 3 classifies soil into a plurality of soil particle groups each having a determined particle size range, and obtains the weight of each soil particle group. Subsequently, based on the results of imaging each soil particle group, the particle size of each soil particle group is specified, and the particle size of the entire soil is specified based on the weight of each soil particle group and the particle size of each soil particle group. .

In addition, in Patent Document 4, in a method for measuring the particle size of a soil material, before photographing the soil material with an imaging device, the soil particles are reliably dispersed without taking much time, and the edge of each soil particle is clarified to improve the accuracy of the particle size distribution of the soil material measured by photographing the soil material with an imaging device and image processing the image of the soil material with an image processing device.

JP 2012-220229 A JP 2013-256842 A JP 2014-178281 A JP 2019-113429 A

By the way, there was a problem that the above-mentioned sieving test required labor and time. In addition, although the techniques described in Patent Documents 1 to 4 can also shorten the time required to specify the characteristics of soil materials such as particle size distribution, work such as classifying soil particle groups into a plurality of soil particle groups is required. There was a problem that it was necessary to perform the work, and that work took time.

One aspect of the present invention is made in view of the above problems. An object of one aspect of the present invention is to shorten the time required for estimating the properties of a soil material.

In order to solve the above-described problems, a soil property estimation method according to the present invention includes the steps of acquiring photographed data representing an image of a photographed soil material; and generating a trained model by machine-learning the correlation between the photographed data and the analysis result data using the photographed data and the analysis result data corresponding to the photographed data; wherein the analysis result data includes (i) a particle size accumulation curve representing the analysis result of the particle size distribution of the soil material approximated to a predetermined function. parameter values of the function in the above case, (ii) weight percentage of permeated mass for each range when the particle size of the soil material is divided into a plurality of ranges, and (iii) measurement results of the water content ratio of the soil material. and data .

According to the above configuration, photographed data representing a photographed image of the soil material and the parameter value of the function when the particle size accumulation curve indicating the analysis result of the particle size distribution of the soil material is approximated to a predetermined function. and analysis result data including at least one of the following: , data indicating the measurement result of the water content ratio of the soil material, and the weight percentage of the passing material for each range when the particle size of the soil material is divided into a plurality of ranges. , a trained model is generated by machine-learning the correlation between the two. By using this trained model , the time required to estimate at least one of the particle size accumulation curve indicating the particle size distribution of the soil material, the passing mass percentage for each particle size range, and the water content ratio of the soil material is shortened.

The method for estimating soil characteristics includes the steps of: acquiring as input data photographed data representing a photographed image of the soil material; and estimating the characteristics of the soil material by inputting the input data into the trained model. , may also be provided.

According to the above configuration, the characteristics of the soil material are estimated by inputting photographed data obtained by photographing the soil material into the learned model. This shortens the time required to estimate the properties of the soil material.

The method for estimating soil characteristics further includes a step of performing image processing on the photographed data, and in the step of generating the learned model, the machine learning is performed using the preprocessed photographed data. may be done.

According to the above configuration, machine learning is performed using photographed data on which image processing has been performed. As a result, it is possible to improve the accuracy of estimating the properties of the soil material.

Further, a method for estimating soil characteristics according to the present invention includes a step of obtaining photographed data representing a photographed image of a soil material as input data; a step of estimating the characteristics of the soil material by inputting the correlation with analysis result data indicating the analysis result of analyzing the characteristics of the material into a trained model that has undergone machine learning, and the soil characteristics performed by the information processing device. In the estimation method, the analysis result data includes (i) parameter values of a predetermined function when a particle size accumulation curve indicating the analysis result of the particle size distribution of the soil material is approximated to the function, (ii ) when the particle size of the soil material is divided into a plurality of ranges, and (iii) data showing the measurement result of the water content ratio of the soil material. characterized by

According to the above configuration, photographed data representing a photographed image of the soil material and the parameter value of the function when the particle size accumulation curve indicating the analysis result of the particle size distribution of the soil material is approximated to a predetermined function. , the weight percentage of passing material for each range when the particle size of the soil material is divided into a plurality of ranges, and the data showing the measurement result of the water content ratio of the soil material. Correlation with analysis result data including at least one is used to estimate the properties of the soil material from the photographed data of the soil material. This shortens the time required for estimating at least one of the particle size accumulation curve indicating the particle size distribution of the soil material, the passing mass percentage for each particle size range, and the water content ratio of the soil material.

Further, the apparatus according to the present invention includes an input data acquisition unit that acquires input data representing an image of a photographed soil material; and an output unit that outputs output data indicating the characteristics of the soil material, obtained by inputting the correlation with analysis result data indicating the analysis result of analyzing the The analysis result data includes (i) parameter values of a predetermined function when a particle size accumulation curve indicating the analysis result of the particle size distribution of the soil material is approximated to the function, (ii) the soil characterized by including at least one of the weight percentage of passing mass for each range when the particle size of the material is divided into a plurality of ranges, and (iii) data showing the measurement result of the water content ratio of the soil material. It is a device that

According to the above configuration, photographed data representing a photographed image of the soil material and the parameter value of the function when the particle size accumulation curve indicating the analysis result of the particle size distribution of the soil material is approximated to a predetermined function. , the weight percentage of passing material for each range when the particle size of the soil material is divided into a plurality of ranges, and the data showing the measurement result of the water content ratio of the soil material. Correlation with analysis result data including at least one is used to estimate the properties of the soil material from photographed data obtained by photographing the soil material. This shortens the time required for estimating at least one of the particle size accumulation curve indicating the particle size distribution of the soil material, the passing mass percentage for each particle size range, and the water content ratio of the soil material.

In the characteristic estimation device, the analysis result data includes parameter values of a predetermined function obtained by approximating a particle size accumulation curve indicating the analysis result of the particle size distribution of the soil material, and the output unit may output a particle size additive curve resulting from applying the parameter values obtained by the trained model to the function.

According to the above configuration, the correlation between the photographed data representing the photographed image of the soil material and the parameter values of the function approximating the particle size accumulation curve indicating the analysis result of the particle size distribution of the soil material is machine-learned. Using the trained model, a particle size accumulation curve representing the particle size distribution of the soil material is estimated from photographed data of the soil material. This shortens the time required for estimating the particle size accumulation curve that indicates the particle size distribution of the soil material.

Further, the present invention is a program for causing a computer to function as the device, and includes a configuration for causing the computer to function as the input data acquisition section and the output section.

According to the above configuration, photographed data representing a photographed image of the soil material and a parameter value of the function when the particle size accumulation curve indicating the analysis result of the particle size distribution of the soil material is approximated to a predetermined function, Correlation with analysis result data including at least one of data indicating the measurement results of the weight percentage of passing material for each range when the particle size of the soil material is divided into a plurality of ranges, and the water content ratio of the soil material. By operating a device having a machine-learned model, characteristics of the soil material can be estimated from image data of the soil material. This shortens the time required for estimating at least one of the particle size accumulation curve indicating the particle size distribution of the soil material, the passing mass percentage for each particle size range, and the water content ratio of the soil material.

Further, a system according to the present invention includes an estimating device and a terminal device, wherein the estimating device includes an input data receiving unit for receiving input data representing an image of a photographed soil material; The characteristics of the soil material are estimated by inputting the correlation between the photographed data representing the photographed image of the soil material and the analysis result data indicating the analysis result of analyzing the characteristics of the soil material into the trained model that has undergone machine learning. , an estimating unit that transmits output data indicating an estimation result, the terminal device includes: an input data acquiring unit that acquires photographed data representing a photographed image of the soil material as input data; a receiving unit that transmits the output data to a server and receives the output data transmitted from the server as a response to the input data; and an output unit that outputs data relating to characteristics represented by the output data, wherein the analysis result data (i) a parameter value of the function when the particle size accumulation curve indicating the analysis result of the particle size distribution of the soil material is approximated to a predetermined function; It is a configuration including at least one of the passing mass weight percentage for each range when divided into ranges, and (iii) data showing the measurement result of the water content ratio of the soil material.

According to the above configuration, photographed data representing a photographed image of the soil material and a parameter value of the function when the particle size accumulation curve indicating the analysis result of the particle size distribution of the soil material is approximated to a predetermined function, Correlation with analysis result data including at least one of data indicating the measurement results of the weight percentage of passing material for each range when the particle size of the soil material is divided into a plurality of ranges, and the water content ratio of the soil material. Using a system that has a machine-learned model, the properties of the soil material are estimated from the image data of the soil material. This shortens the time required for estimating at least one of the particle size accumulation curve indicating the particle size distribution of the soil material, the passing mass percentage for each particle size range, and the water content ratio of the soil material.

ADVANTAGE OF THE INVENTION According to one aspect of the present invention, it is possible to shorten the time required for estimating the properties of a soil material.

2 is a diagram illustrating a functional configuration of an estimation device according to Embodiment 1; FIG. 4 is a flow chart showing the flow of processing executed by an estimating device; FIG. 10 is a diagram showing an outline of a characteristic estimation system according to Embodiment 2; FIG. FIG. 2 is a block diagram showing the main configuration of various devices included in the characteristic estimation system; FIG. 10 is a diagram schematically showing an example of a trained model according to Embodiment 2; FIG. 10 is a flow chart showing the flow of processing related to generation of a trained model in an estimation device according to Embodiment 2. FIG. It is the figure which illustrated the image represented by imaging|photography data. FIG. 4 illustrates a function by which a particle size accumulation curve is approximated; FIG. 4 illustrates a function by which a particle size accumulation curve is approximated; FIG. 4 is a diagram showing an example of approximating a particle size accumulation curve to a function; FIG. 4 is a diagram showing an example of approximating a particle size accumulation curve to a function; 9 is a flow chart showing the flow of processing related to estimation of properties of soil materials in the terminal device according to Embodiment 2. FIG. 9 is a flow chart showing the flow of processing related to estimation of properties of soil materials in the estimation device according to Embodiment 2. FIG. 10 is a flow chart showing the flow of processing related to estimation of soil material characteristics in an estimation device according to Embodiment 3. FIG.

[Embodiment 1]
FIG. 1 is a diagram showing the functional configuration of an information processing device 200 according to one embodiment of the present invention. The information processing device 200 is a device that generates a learned model for estimating characteristics of soil materials. In the present embodiment, soil material refers to soil generated by excavation work or the like, and includes, for example, soil, sand, gravel materials, and the like. Properties of a soil material refer to characteristics and/or properties of the soil material. The characteristics of the soil material are, for example, the particle size distribution of the soil material, the particle size accumulation curve specified from the particle size distribution, the parameter value of the particle size accumulation curve, and the passage quality for each range when the particle size is divided into multiple ranges. Weight percentage or content. Further, the particle size refers to the distribution state of the particle size of the soil material expressed in weight percentage. The passing weight percentage is a percentage value obtained by dividing the weight of soil particles having a particle size within each range by the total weight when the particle size of the soil material is divided into a plurality of ranges.

The information processing apparatus 200 includes a captured data acquisition unit 101 , an analysis result acquisition unit 102 and a generation unit 103 . The photographed data acquisition unit 101 acquires photographed data representing a photographed image of a soil material.

The analysis result acquisition unit 102 acquires analysis result data indicating the analysis result of analyzing the properties of the soil material. The analysis result data includes, for example, the particle size distribution of the soil material, the particle size accumulation curve corresponding to the particle size distribution, the parameter value of the particle size accumulation curve, and the passing mass weight percentage for each range when the particle size is divided into multiple ranges. , or data indicating the content rate.

The analysis result data may be generated by the information processing device 200 , or the analysis result data generated by another device may be acquired by the information processing device 200 . The analysis result data is, for example, data indicating the analysis result of the soil particle size analysis performed by the "soil particle size test method" defined in JIS A 1204. In addition, the soil particle size analysis method is not limited to this, and may be performed, for example, by the method described in Japanese Patent Application Laid-Open No. 2014-178281.

The generating unit 103 generates a learned model 121 by machine-learning the correlation between the photographed data and the analysis result data using a set of the photographed data and the analysis result data corresponding to the photographed data. Various parameters that define the learned model 121 are stored in the storage unit 12 . The trained model 121 is a trained model obtained by machine-learning correlations between images of various soil materials and analysis results. The structure of the trained model 121 is not particularly limited. For example, the trained model 121 can be realized by a model having a CNN (Convolutional Neural Network) structure.

FIG. 2 is a flow chart showing the flow of processing (an example of a soil property estimation method) performed by the information processing device 200 . Note that some steps may be performed in parallel or out of order. The photographed data acquisition unit 101 acquires photographed data representing a photographed image of the soil material (S10). The analysis result acquisition unit 102 acquires analysis result data indicating the analysis result of analyzing the characteristics of the soil material (S11). The generation unit 103 generates a learned model 121 by machine-learning the correlation between the photographed data and the analysis result data using the photographed data and the analysis result data corresponding to the photographed data (S12).

The information processing apparatus 200 configured as described above generates the learned model 121 by machine-learning the correlation between the photographed data and the analysis result data. It is possible to shorten the time required for estimating characteristics.

[Embodiment 2]
≪System Overview≫
FIG. 3 is a diagram showing an overview of the characteristic estimation system 100 according to this embodiment. The property estimation system 100 is a system for estimating properties of soil materials using a learned model. A property estimation system 100 includes an estimation device 1 , a photographing device 2 , a terminal device 3 , a storage device 4 and a terminal device 5 . Note that the terminal device 3 is not an essential component. The imaging device 2, the terminal device 3 and the estimation device 1, the terminal device 3 and the storage device 4, the estimation device 1 and the terminal device 3, the estimation device 1 and the storage device 4, and the estimation device 1 and the terminal device 5 can communicate with each other. It is connected to the.

The photographing device 2 photographs the soil material and outputs photographing data representing the photographed image. The shooting data may be data representing a still image, or may be data representing a moving image. Also, the photographed data may be data including a plurality of image data. The photographed data is supplied to the terminal device 3 . For example, the imaging device 2 transmits imaging data to the terminal device 3 connected by wire or wirelessly. Alternatively, the photographing device 2 may record photographed data in an external recording medium such as a USB flash memory and an SD card. Then, the external recording medium may be connected to the terminal device 3 and read by the terminal device 3 . Thereby, the photographed data can be transferred from the photographing device 2 to the terminal device 3 . In this way, when photographing data is supplied via an external recording medium, communication connection between the photographing device 2 and the terminal device 3 is not essential.

The terminal device 3 is a device for experts to browse various data such as photographed data. In this embodiment, expert refers to a person skilled in analyzing the properties of soil materials, such as a geologist or geotechnical engineer. Also, the terminal device 3 is a device for the expert to input analysis results obtained by analyzing the characteristics of the soil material represented by the photographed data viewed by the expert. The terminal device 3 transmits data indicating the input analysis result to the estimation device 1 . The terminal device 3 transmits to the estimating device 1 data indicating the analysis results of the properties of the soil material. Henceforth, the analysis result of the characteristic of a soil material is also simply called an "analysis result." Further, data indicating the analysis result is referred to as "analysis result data".

The terminal device 3 also associates the received photographed data with the analysis result corresponding to the photographed data, and transmits them to the storage device 4 . In the storage device 4, a set of the photographed data and the analysis result is stored as teaching data. That is, the training data is a combination of photographed data of various soil materials and analysis results corresponding to each of the various soil materials.

The estimation device 1 is a device that generates a trained model 121 and uses the generated trained model 121 to perform a characteristic estimation process. The estimation device 1 generates a learned model 121 using the teacher data stored in the storage device 4 . Here, the learned model 121 is a learning model obtained by machine-learning the correlation between photographed data of the geological material and analysis results of the properties of the soil material.

The storage device 4 is a storage device that stores the aforementioned teacher data. The storage device 4 reads out the stored training data in response to a request from the estimation device 1 and transmits it to the estimation device 1 .

A terminal device 5 (an example of a property estimation device) is a device for a user to browse results of estimation of soil material properties by the estimation device 1 . The terminal device 5 is, for example, a tablet terminal or a smart phone. The terminal device 5 transmits image data representing a photographed image of the soil material to the estimating device 1 according to the user's operation. The image represented by the image data transmitted from the terminal device 5 to the estimation device 1 may be a still image or a moving image. The image data transmitted from the terminal device 5 to the estimation device 1 is data to be input to the trained model, and hereinafter this image data is referred to as "input data". The terminal device 5 receives the analysis result data transmitted from the estimation device 1 in response to the transmitted input data, and displays the analysis result indicated by the analysis result data on its liquid crystal display or the like.

In the characteristic estimation system 100, the terminal device 3 and the estimation device 1 may be integrated. That is, the photographing device 2 may transmit photographed data to the estimating device 1 , and the expert may view the photographed data on the estimating device 1 . Then, the expert may input the analysis result to the estimation device 1 .

Moreover, in the characteristic estimation system 100, the terminal device 5 and the estimation device 1 may be configured integrally. In other words, the user may browse the estimation results of the characteristics using the estimation device 1 .
In that case, instead of input data from the terminal device 5 , image data obtained by photographing the target soil material with the photographing device 2 is transmitted to the estimation device 1 .

≪Main part composition≫
FIG. 4 is a block diagram showing the main configuration of a device included in characteristic estimation system 100. As shown in FIG. Although the terminal device 3 is also illustrated in FIG. 2 , the terminal device 3 is not an essential component in the characteristic estimation system 100 .

(Photographing device 2)
The imaging device 2 includes at least a camera 22 . The camera 22 photographs the soil material and outputs photographed data representing the photographed image. Camera 22 may be a high resolution camera. Note that a high-resolution camera is a camera in which the number of pixels of the mounted image sensor is 100 million pixels or more. Furthermore, the photographing device 2 may include one or more of the control section 20 , the communication section 21 , the light source 24 and the storage section 25 . Light source 24 is a standard light source for photographing soil materials. The light source 24 can be implemented by, for example, a light using light emitting diodes. Further, the photographing device 2 may include an input section such as a button, a mouse, and a touch panel, and a display section such as a display.

The communication unit 21 is a communication interface that performs communication between the imaging device 2 and other devices. For example, the communication section 21 communicates with the terminal device 3 under the control of the control section 20 . For example, the communication unit 21 transmits photographed data input from the control unit 20 to the terminal device 3 . Further, the communication unit 21 may receive an instruction to start shooting with the shooting device 2 from another device (not shown). When receiving an instruction to start shooting, the communication unit 21 outputs the instruction to the control unit 20 .

The control unit 20 controls the imaging device 2 in a centralized manner. The control unit 20 controls the camera 22 to photograph the soil material. For example, the control unit 20 may cause the camera 22 to perform photographing in response to a photographing start instruction received from another device via the communication unit 21 . Further, for example, the control unit 20 may receive an input operation for instructing the start of imaging by the user via the input unit of the imaging device 2 . Then, when the input operation is accepted, the control unit 20 may cause the camera 22 to shoot. The photographed data is temporarily stored in the storage unit 25 until it is transmitted to the terminal device 3 .

The storage unit 25 is a storage device that stores various data necessary for processing of the imaging device 2 . In addition, the storage unit 25 stores shooting data. The control unit 20 reads the photographed data stored in the storage unit 25 and transmits it to the terminal device 3 . The timing at which the photographed data is transmitted may be, for example, the timing at which a request from the terminal device 3 is received, or, for example, the photographed data may be transmitted each time photographing is performed.

(Terminal device 3)
The terminal device 3 is realized by, for example, a personal computer (PC) or the like. The terminal device 3 includes a display section for displaying imaging data, and an input section such as a mouse, buttons, and keyboard for the expert to input analysis results. Also, the terminal device 3 has a communication unit for communicating with the estimation device 1 and the photographing device 2 . In addition, the terminal device 3 has an interface for connecting an external recording medium in which photographed data of the photographing device 2 is recorded.

The terminal device 3 transmits analysis result data to the estimation device 1 . The terminal device 3 also associates the imaging data acquired from the imaging device 2 with the analysis result data for the imaging data input by the expert, and transmits them to the storage device 4 .

(Estimation device 1)
The estimation device 1 includes a communication unit 11 , a storage unit 12 and a control unit (information processing device) 10 . The estimating device 1 may have an interface to which an external recording medium can be connected. For example, the estimating device 1 may have an interface to which a USB flash memory or SD card can be connected. Also, the estimation device 1 may include an input unit such as a button, mouse, and touch panel, and a display unit such as a display.

The communication unit 11 realizes communication between the estimation device 1 and other devices. The communication unit 11 communicates with the terminal device 3 , the storage device 4 and the terminal device 5 . For example, the communication unit 11 receives analysis result data from the terminal device 3 . Note that the communication unit 11 does not need to communicate with the terminal device 3 when acquiring analysis result data via an external recording medium. Also, for example, the communication unit 11 receives teacher data from the storage device 4 . For example, the communication unit 11 receives input data from the terminal device 5 .

The storage unit 12 is a storage device that stores data necessary for processing of the estimation device 1 . Storage unit 12 stores learned model 121 . The learned model 121 is the same as in the first embodiment described above, and is a learned model obtained by machine-learning correlations between images of various soil materials and analysis results. For example, the trained model 121 can be realized with a model having a CNN structure. As an example, the characteristic estimation system 100 according to the present embodiment uses a trained model 121 having a model structure of AlexNet. Note that the storage unit 12 may be an external device of the estimation device 1 . For example, the storage unit 12 may be a storage device such as a server communicably connected to the estimation device 1 .

The control unit 10 comprehensively controls the estimation device 1 . The control unit 10 includes a photographed data acquisition unit 101 , an analysis result acquisition unit 102 , a generation unit 103 , an input data acquisition unit 104 , an estimation unit 105 and an image processing unit 106 . The imaging data acquisition unit 101, the analysis result acquisition unit 102, and the generation unit 103 are the same as those of the information processing apparatus 200 according to the first embodiment described above. That is, the photographed data acquisition unit 101 acquires photographed data representing a photographed image of the soil material. The analysis result acquisition unit 102 acquires analysis result data indicating the analysis result of analyzing the properties of the soil material. The generating unit 103 generates a learned model 121 by machine-learning the correlation between the photographed data and the analysis result data using a set of the photographed data and the analysis result data corresponding to the photographed data.

The input data acquisition unit 104 acquires photographed data representing an image of a soil material as input data. In this embodiment, the input data acquisition unit 104 (an example of the input data reception unit) receives input data from the terminal device 5 . The estimation unit 105 inputs the input data acquired by the input data acquisition unit 104 to the learned model 121 to estimate the properties of the soil material. The image processing unit 106 executes image processing on the input data acquired by the captured data acquiring unit 101 or the input data acquiring unit 104 .

In the example of FIG. 4 , the control unit 10 of the estimation device 1 includes a captured data acquisition unit 101, an analysis result acquisition unit 102, a generation unit 103, an input data acquisition unit 104, an estimation unit 105, and an image processing unit 106. However, the functions realized by these components may be shared and realized by a plurality of devices. For example, the control unit 10 of the estimation device 1 does not include the captured data acquisition unit 101, the analysis result acquisition unit 102, the generation unit 103, and the image processing unit 106, but includes the input data acquisition unit 104 and the estimation unit 105. may be In this case, the trained model 121 is generated by an information processing device (not shown) separate from the estimating device 1, which includes the captured data acquisition unit 101, the analysis result acquisition unit 102, the generation unit 103, and the image processing unit 106, Using the generated trained model 121, the estimation device 1 executes the characteristic estimation process. Note that the control unit 10 of the estimation device 1 may be configured to include the image processing unit 106 in addition to the input data acquisition unit 104 and the estimation unit 105 .

The storage device 4 is a storage device that stores learning data sets 41 . The learning data set 41 is a group of data sets containing one or more pieces of teacher data. The storage device 4 reads a data set from the learning data set 41 in response to a request from the estimating device 1 and transmits the data set to the estimating device 1 . Note that the storage unit 12 and the storage device 4 described above may be configured integrally. That is, the trained model 121 and the learning data set 41 may be stored in the same storage device.

<<Details of trained model>>
FIG. 5 is a diagram schematically showing an example of the trained model 121 according to this embodiment. As shown, photographed data is input to the trained model 121 as input data.

The trained model 121 consists of, for example, a convolution layer, a pooling layer, and a connection layer. In the convolutional layer, input data is convolved with information by filtering. Data that has undergone convolution is subjected to pooling processing in the pooling layer. This improves the model's ability to recognize changes in the position of features in the data. The data that has undergone the pooling process is processed in the coupling layer to be converted into the output data of the learned model 121, ie, the results of estimating the properties of the soil material, and output.

That is, by passing the photographed data input to the learned model 121 through each layer shown in FIG. . Note that the output format of the estimation result is not particularly limited. For example, various characteristics may be indicated numerically as index values, or a graph representing a particle size accumulation curve may be displayed. Also, various characteristics may be indicated by text data.

Returning to the description of FIG. Terminal device 5 includes control unit 50 , communication unit 51 , storage unit 52 , input/output unit 53 , and camera 54 . The communication unit 51 realizes communication between the terminal device 5 and other devices. The communication unit 51 communicates with the estimation device 1 . For example, the communication unit 51 transmits input data to the estimation device 1 . For example, the communication unit 51 receives analysis result data from the estimation device 1 . The storage unit 52 is a storage device that stores data necessary for processing of the terminal device 5 . Input/output unit 53 includes an input unit such as a button, mouse, and touch panel, and a display unit such as a display. The camera 54 photographs the soil material and outputs photographed data representing the photographed image.

The control unit 50 comprehensively controls the terminal device 5 . Control unit 50 includes acquisition unit 501 , reception unit 502 , and output unit 503 . The acquisition unit 501 acquires, as input data, photographed data representing a photographed image of a soil material. The receiving unit 502 transmits input data to the estimation device 1 and receives output data transmitted from the estimation device 1 in response to the input data. The received output data is data indicating properties of the soil material obtained by inputting the input data into the trained model 121 . The output unit 503 outputs data regarding the characteristics indicated by the received output data.

<<Process flow>>
<<Process related to collection of learning data set>>
FIG. 6 is a sequence diagram showing the flow of processing related to generation of the trained model 121 according to this embodiment. The input data used in this explanation is the image data from the photographing device 2 . Note that some steps may be performed in parallel or out of order.

First, the photographing device 2 photographs the soil material to be analyzed (S20). The timing at which the photographing is performed by the photographing device 2 is, for example, the timing at which a predetermined condition is satisfied. The timing at which the predetermined condition is satisfied is, for example, the timing at which the user of the imaging device 2 performs an operation to instruct photographing, or the timing at which a predetermined unit time has elapsed. The photographed data generated by photographing is transmitted to the terminal device 3 (S21).

FIG. 7 is a diagram exemplifying an image represented by shooting data. The camera 22 photographs, for example, the soil material and the scale placed thereon. Note that the photographed data may be a photographed image of a soil material that does not include scale. When the photographed image does not include a scale, a numerical value indicating the scale is input by the user's operation on the photographing device 2 or the terminal device 3 .

An expert or the terminal device 3 analyzes the properties of the soil material corresponding to the photographed data transmitted to the terminal device 3 . This analysis may be performed by an expert using the terminal device 3, or may be performed by the terminal device 3 reading and executing a computer program stored in its own storage unit. The analysis result indicated by the analysis result data is, for example, the particle size distribution of the soil material, the particle size accumulation curve indicated by the particle size distribution, the parameter value of the particle size accumulation curve, or the water content ratio. When an expert performs the analysis, for example, the particle size analysis of the soil is performed according to the "soil particle size test method" specified in JIS A 1204.

In addition, the terminal device 3 performs image analysis processing on the photographed data, detects the edges of the soil particles contained in the soil material, measures the particle size of each soil particle (the major axis and the minor axis of the ellipse), and measures the The results may be used to determine particle size distribution. At this time, if the image represented by the photographed data does not include a scale, the terminal device 3 obtains the numerical value indicating the scale input to the photographing device 2 or the terminal device 3, and uses the obtained numerical value to estimate the particle size distribution. Perform specific processing. Further, for example, soil particle size analysis may be performed by the method described in JP-A-2014-178281.

In this embodiment, the analysis result data includes values of parameters of a function that approximates the particle size accumulation curve corresponding to the analysis result. The function to be used is preselected by an expert or the like. A function includes one or more parameters. The terminal device 3 approximates the particle size accumulation curve corresponding to the analysis result by adjusting the parameter values in the function selected in advance (S121).

As a function by which the particle size accumulation curve is approximated, in this embodiment, two types of functions of the following formula (1) and/or formula (2) are used. In the following formulas (1) and (2), D indicates the particle size (median diameter), and y indicates the passing mass percentage. a and b are parameters. That is, the shape of the graph represented by the function of formula (1) or formula (2) changes depending on the values of a and b.
y = 0.5*[tanh(b*LOG10(a*D)+1]*100 (1)
y = {2/[1+exp(-a*(Db))]-1}*100 (2)

8 and 9 are diagrams illustrating functions by which the particle size accumulation curve is approximated. 8 and 9 are graphs illustrating the above equations (1) and (2), respectively. 8 and 9, the horizontal axis indicates the grain size D (mm) of the soil material, and the vertical axis indicates the passing mass percentage y (%). In FIG. 8, three patterns of "Case 1", "Case 2" and "Case 3" are illustrated. The parameter values in each of "Case 1", "Case 2" and "Case 3" are as follows.
Case 1: a=b=1.0
Case 2: a=0.3, b=2.5
Case 3: a=3.0, b=5.0

10 and 11 are diagrams illustrating graphs in which the grain size accumulation curves showing the results of actually conducting grain size tests on soil materials are approximated by functions. A particle size test is a test for examining the particle size distribution of a soil material. The particle size test is performed by, for example, the "soil particle size test method" defined in JIS A 1204. FIG. 10 shows an example approximated by the function of equation (1). FIG. 11 shows an example approximated by the function of equation (2). 10 and 11, the horizontal axis indicates the grain size D (mm) of the soil material, and the vertical axis indicates the passing mass percentage y (%).

In FIG. 10, the particle size accumulation curves corresponding to the results of two particle size tests, "Test Result 1" and "Test Result 2", are indicated by solid lines. Also, the dotted line indicates a graph when each particle size accumulation curve is approximated to the function of the formula (1). The values of the parameters a and b in formula (1) when the particle size accumulation curves of "test result 1" and "test result 2" are approximated to the function of formula (1) are as follows.
Test result 1: a=0.65, b=1.45
Test result 2: a=0.35, b=2.70

In FIG. 11, particle size accumulation curves corresponding to the two test results of "Test Result 1" and "Test Result 2" are indicated by solid lines. Also, a dotted line indicates a graph obtained by approximating each particle size accumulation curve to the function of formula (2). The values of parameters a and b in formula (2) when the particle size accumulation curves of "test result 1" and "test result 2" are approximated to the function of formula (2) are as follows.
Test result 1: a = 0.55, b = -0.15
Test result 2: a=0.38, b=0.10

Returning to the description of FIG. The terminal device 3 generates teacher data in which the photographed data and the analysis result data are linked (S122), and transmits it to the storage device 4 as a learning data set for machine learning (S22). The storage device 4 stores the learning data set (S23). The learning data set is accumulated in the storage device 4 by repeating the processing of steps S20 to S23.

<<Process related to generation of learned model>>
The captured data acquisition unit 101 of the estimation device 1 acquires captured data by receiving the captured data from the storage device 4 (S24). Also, the analysis result acquisition unit 102 acquires the analysis result data by receiving the analysis result data from the storage device 4 (S25). The generating unit 103 generates a trained model 121 by machine-learning the correlation between the photographed data and the analysis result data using a set of the photographed data and the analysis result data corresponding to the photographed data (S26).

≪Processing related to estimation of properties of soil materials≫
Next, the process for estimating the properties of the soil material will be described with reference to the drawings. Image data from the terminal device 5 is used as input data. A user who wants to estimate the characteristics uses the terminal device 5 to perform an operation of photographing the soil material to be analyzed. The terminal device 5 photographs the soil material according to the user's operation. Photographed data representing the photographed image is stored in the storage unit 52 . Also, the user uses the terminal device 5 to perform an operation of instructing the estimation of the characteristics of the soil material. The terminal device 5 executes processing for estimating characteristics in accordance with the user's operation.

FIG. 12 is a flowchart showing the flow of processing related to estimation of soil material characteristics in the terminal device 5 . Note that some steps may be performed in parallel or out of order. The process shown in FIG. 12 is triggered by, for example, the user performing an operation instructing execution of the characteristic estimation process. The acquiring unit 501 of the terminal device 5 acquires the photographed data as input data by reading the photographed data representing the photographed image of the soil material from the storage unit 52 (S30). The receiving unit 502 transmits the input data acquired in step S30 to the estimation device 1 (S31). The receiving unit 502 waits until receiving a response from the estimating device 1 (S32; NO).

FIG. 13 is a flow chart showing the flow of processing related to estimation of soil material characteristics in the estimation device 1 . Note that some steps may be performed in parallel or out of order. The input data acquisition unit 104 acquires input data transmitted from the terminal device 5 (S40). The input data acquisition unit 104 supplies the acquired input data to the image processing unit 106 . The image processing unit 106 performs image processing as preprocessing on the input data (S41). Image processing as preprocessing is, for example, brightness adjustment and contrast adjustment processing. The image processing unit 106 outputs the image-processed input data to the estimation unit 105 . The estimation unit 105 inputs the input data to the learned model 121 (S42), and acquires the characteristic estimation result output from the learned model 121 (S43). This estimation result is, for example, the particle size distribution of the soil material (particle size accumulation curve, parameter values of a function approximating the particle size accumulation curve, etc.) or water content. The estimation unit 105 transmits output data indicating the obtained estimation result to the terminal device 5 via the communication unit 11 (S44). Note that the input data acquired by the input data acquiring unit 104 may be directly supplied to the estimating unit 105 (without image processing) without image processing being performed on the input data.

Returning to the description of FIG. When the receiving unit 502 of the terminal device 5 receives the output data from the estimating device 1 (S32; YES), the output unit 503 displays on the input/output unit 53 the estimation result indicated by the received output data. When the output data are parameter values of a function, the output unit 503 displays on the input/output unit 53 a graph obtained by applying the parameter values to the function. For example, if a function contains two parameters, parameters a and b, the graph is specified by applying the values of these parameters to the function. As a result, the graph illustrated in FIG. 8 or 9, for example, is displayed. At this time, an image represented by photographing data corresponding to this estimation result may be displayed together.

Note that the mode of outputting the output data is not limited to the one described above. The output unit 503 may output the output data, for example, by causing the input/output unit 53 to display a numerical value indicating the water content. Further, when the output data is a particle size accumulation curve, the output unit 503 may cause the input/output unit 53 to display the particle size accumulation curve indicated by the output data.

By checking the graph displayed on the input/output unit 53, the user of the terminal device 5 can grasp the characteristics (for example, the grain size accumulation curve) of the soil material to be analyzed. As described above, in this embodiment, the user takes an image of the soil material to be analyzed using the terminal device 5 , and simply performs a specific operation using the terminal device 5 . It is possible to obtain the estimation result of the characteristics of That is, in this embodiment, the user does not need to perform complicated work or operations to estimate the properties of the soil material, and the time required for estimation is shortened.

Conventionally, a method of measuring the particle size distribution of soil materials by image processing (for example, the method described in Patent Document 4) has been used. It becomes difficult to detect the edge of a particle of . Therefore, there is a problem that it is necessary to disperse the soil particles so that they do not overlap. On the other hand, in this embodiment, when estimating the particle size distribution, the particle size can be measured without dispersing the soil particles and even if the soil particles overlap each other.

Also, in this embodiment, a method is used in which the particle size accumulation curve is approximated by a function and the values of the parameters in the function are machine-learned. Since the combination of parameter values is the object of machine learning, machine learning can be performed by a simpler method than learning the particle size addition curve itself.

(Modification 1)
In the above-described embodiment, the values of the parameters of the function approximating the particle size accumulation curve were used as the analysis result data. The analysis result data are not limited to those shown in the above embodiments. The analysis result data may be data indicating the analysis result of analyzing the characteristics of the soil material, and the analysis result data indicates, for example, the passage percentage for each range when the particle size of the soil material is divided into a plurality of ranges. may contain data. Specifically, for example, "0.01 mm or less: 5%", "0.05 mm or less: 11%", "0.075 mm or less: 15%", or "0 to 0.01 mm: 5%", " 0 to 0.05 mm: 11%”, “0 to 0.075 mm: 15%”, and so on. good.

Moreover, the analysis result data may include the water content ratio of the soil material. In this case, the water content ratio is measured by a measuring device (not shown), and data indicating the measurement result is stored in the storage device 4 . In this case, machine learning is performed on the correlation between the data indicating the measurement results and the photographed data to generate a learned model.

According to this aspect, the user can obtain the result of estimating the water content of the soil material simply by photographing the soil material to be analyzed using the terminal device 5 and performing a specific operation using the terminal device 5. can. That is, in this embodiment, there is no need to provide a measuring device for measuring the water content ratio of the soil material, and the user does not need to perform complicated work or operations for measuring the water content ratio.

(Modification 2)
In the above-described embodiment, the estimation device 1 (or the terminal device 3) may perform image processing as preprocessing on the captured data, and machine learning may be performed using the captured data subjected to this image processing. Image processing as preprocessing is, for example, brightness adjustment and contrast adjustment processing.

(Modification 3)
In the above-described embodiment, the estimating device 1 performed the process of estimating the properties of the soil material (Fig. 13). In the property estimation system 100 , the terminal device 5 may perform the process of estimating the property of the soil material. In this case, a software program for executing the process of estimating the characteristics of the soil material is downloaded to the storage unit 52 of the terminal device 5, and the terminal device 5 reads and executes this program, thereby enabling the estimation device 1 to be used. Characteristic estimation processing is executed in The terminal device 5 wirelessly communicates with the storage device 4 and uses the generated learned model 121 to execute the characteristic estimation process.

(Modification 4)
In each of the above-described embodiments, a user such as an expert may correct the estimation result of the trained model and re-learn the trained model. In this case, the estimation result is corrected by the user's operation on the terminal device 3 or the terminal device 5 . When the terminal device 3 corrects the estimation result, the estimation result by the estimation device 1 is transmitted to the terminal device 3 .

The terminal device 3 (or terminal device 5) accepts an operation by the user to instruct correction of the estimation result. The terminal device 3 (or terminal device 5) corrects the estimation result according to the user's operation to create the analysis result of the properties of the soil material. The terminal device 3 (or terminal device 5) transmits the analysis result to the estimation device 1, and transmits the corrected analysis result of the properties of the soil material and the target photographing data to the storage device 4 in association with each other.

The analysis result acquisition unit 102 of the estimation device 1 acquires analysis result data from the terminal device 3 (or terminal device 5) via the communication unit 11, and combines the acquired analysis result data with the photographed data corresponding to the analysis result data. to retrain the learning model. The storage device 4 adds data to the learning data set 41 .

(Modification 5)
Note that the estimation device 1 may be configured not to transmit the estimation result to the terminal device 3 . Then, similarly to the terminal device 3 according to the first embodiment, the terminal device 3 may create expert analysis results from the photographed data and transmit the analysis results to the estimation device 1 . In this case, the estimation device 1 may compare the estimation result by the estimation unit 105 and the analysis result by the expert, and re-learn the trained model 121 when at least part of them are different.

In this case, the estimation device 1 compares the estimation result by the estimation unit 105 with the analysis result acquired from the terminal device 3 (or the terminal device 5), and if at least a part of them are different, the learning data set , and the trained model 121 is retrained using the extracted data set.

(Modification 6)
In the above-described embodiment, the terminal device 3 transmits a set of the photographed data and the analysis result data to the storage device 4, and the storage device 4 associates and stores the photographed data and the analysis result data (step in FIG. 6). S22-S23). The imaging data and the analysis result data may be transmitted to the storage device 4 separately. For example, photographed data may be transmitted from the imaging device 2 to the storage device 4 and analysis result data may be transmitted from the terminal device 3 to the storage device 4 . In this case, for example, identification information for identifying the photographed data may be added to the analysis result data to link the two.

(Modification 7)
In the above-described embodiment, the terminal device 3 approximates the particle size accumulation curve by adjusting the parameter values in the function selected in advance (S121 in FIG. 6). The process of identifying the parameter values of the function may be performed by a device other than the terminal device 3 (for example, the estimation device 1). For example, the estimating device 1 may receive data indicating the particle size accumulation curve from the terminal device 3, approximate the particle size accumulation curve indicated by the received data to a predetermined function, and identify the parameter value.

(Modification 8)
The trained model 121 used in each of the above-described embodiments is not limited to those shown in the above-described embodiments. The learned model 121 may be, for example, an MTRNN (Multi Timescale RNN), an RNN (Recurrent Neural Network) such as an LSTM (Long Short Term Memory), an ARIMA (AutoRegressive, Integrated and Moving Average) model, or the like.

[Embodiment 3]
Further, in each of the above-described embodiments, even if the estimation is performed using both the analysis result by the existing software for particle size analysis (hereinafter referred to as "particle size analysis software") and the estimation result using the trained model 121, good. The particle size analysis software, for example, performs image analysis processing on the image data of the soil material, detects the edges of the soil particles contained in the soil material, measures the particle size of each soil particle, and uses the measurement results. to identify the particle size distribution. Note that the particle size analysis processing executed by the particle size analysis software is not limited to this, and the particle size analysis may be performed by other methods. The particle size analysis software is installed in the estimation device 1, and implemented by the controller 10 of the estimation device 1 reading and executing the computer program of the particle size analysis software.

FIG. 14 is a flowchart showing a flow of processing related to estimation of soil material characteristics in the estimation device 1 . The processing shown in FIG. 14 differs from the processing in FIG. 13 in that the processing of step S51 is executed after the processing of step S41, and the processing of step S52 is executed after the processing of step S43. . Note that some steps may be performed in parallel or out of order.

In step S40, the particle size analysis software implemented by the control unit 10 of the estimation device 1 performs image analysis processing on the photographed data, for example, detects the edges of soil particles contained in the soil material, and determines the particle size of each soil particle. The particle size distribution is specified by measuring (the major axis and the minor axis of the ellipse) (S51). After that, the generation unit 103 of the estimation device 1 weights the analysis result by the granularity analysis software and the estimation result using the trained model 121 according to the machine learning reliability calculated by the trained model 121, The estimation result of the particle size distribution is specified (S52). For example, computation is performed such that the higher the reliability, the heavier the weighting of the estimation result by machine learning.

According to this embodiment, it is possible to obtain an estimation result that takes into consideration both the analysis result of particle size analysis by image analysis and the estimation result using the trained model, thereby improving the accuracy of estimation of the characteristics of the soil material. can be higher.

[Embodiment 4]
Further, in each of the above-described embodiments, the analysis result by the existing particle size analysis software and the estimation result using the trained model 121 may be used together. In this case, the particle size analysis software implemented by the control unit 10 of the estimation device 1 performs image analysis processing on the photographed data, detects the edges of soil particles contained in the soil material, and measures the particle size of each soil particle. do. The particle size analysis software then identifies the particle size distribution for the large particle size range over which the particles can be counted. On the other hand, for the small particle size range that cannot be counted, the estimation result using the learned model 121 is adopted. That is, as the estimation result of the particle size distribution by the estimation device 1, the analysis result by the particle size analysis software is adopted for the soil particles having a particle size equal to or larger than the predetermined threshold, while for the soil particles having a particle size less than the threshold, , an estimation result using the trained model 121 is adopted. That is, in this embodiment, the control unit 10 learns soil particles whose properties satisfy predetermined conditions (for example, the range of particle diameters of soil particles) among the soil particles contained in the soil material to be analyzed. A particle size analysis result estimated by the model 121 is obtained. On the other hand, for soil particles that do not satisfy the above conditions, the control unit 10 acquires the analysis result specified by the particle size analysis processing that does not use the learned model 121 .

By using a combination of analysis results by existing software and estimation results using the trained model 121, the accuracy of estimation of soil material properties can be increased.

[Example of realization by software]
Estimation device 1, terminal device 3, terminal device 5, and control blocks of information processing device 200 (captured data acquisition unit 101, analysis result acquisition unit 102, generation unit 103, input data acquisition unit 104, estimation unit 105, image processing unit 106, acquisition unit 501, reception unit 502, and output unit 503) may be realized by a logic circuit (hardware) formed in an integrated circuit (IC chip) or the like, or by software. You may

In the latter case, the estimating device 1, the terminal device 3, the terminal device 5, and the information processing device 200 are provided with computers that execute instructions of programs, which are software for realizing each function. This computer includes, for example, one or more processors, and a computer-readable recording medium storing the program. In the computer, the processor reads the program from the recording medium and executes it, thereby achieving the object of the present invention. As the processor, for example, a CPU (Central Processing Unit) can be used. As the recording medium, a "non-temporary tangible medium" such as a ROM (Read Only Memory), a tape, a disk, a card, a semiconductor memory, a programmable logic circuit, or the like can be used. In addition, a RAM (Random Access Memory) for developing the above program may be further provided. Also, the program may be supplied to the computer via any transmission medium (communication network, broadcast wave, etc.) capable of transmitting the program. Note that one aspect of the present invention can also be implemented in the form of a data signal embedded in a carrier wave in which the program is embodied by electronic transmission.

The present invention is not limited to the above-described embodiments, but can be modified in various ways within the scope of the claims, and can be obtained by appropriately combining technical means disclosed in different embodiments. is also included in the technical scope of the present invention.

Reference Signs List 1 estimation device 2 photographing device 3, 5 terminal device 4 storage device 10, 20, 50 control unit 11, 21, 51 communication unit 12, 25, 52 storage unit 22, 54 camera 24 light source 41 learning data set 53 input/output unit REFERENCE SIGNS LIST 100 characteristic estimation system 101 captured data acquisition unit 102 analysis result acquisition unit 103 generation unit 104 input data acquisition unit 105 estimation unit 106 image processing unit 121 trained model 200 information processing device 501 acquisition unit 502 reception unit 503 output unit

Claims (8)

obtaining photographed data representing a photographed image of the soil material;
a step of obtaining analysis result data indicating analysis results of analyzing the characteristics of the soil material;
a step of generating a learned model by performing machine learning of the correlation between the photographed data and the analysis result data using the photographed data and the analysis result data corresponding to the photographed data;
A soil property estimation method executed by an information processing device ,
The analysis result data includes (i) parameter values of a predetermined function when a particle size accumulation curve indicating the analysis result of the particle size distribution of the soil material is approximated to the function, (ii) grains of the soil material Soil characteristics characterized by including at least one of the weight percentage of passing material for each range when the diameter is divided into a plurality of ranges, and (iii) data showing the measurement result of the water content ratio of the soil material. estimation method . a step of obtaining, as input data, photographed data representing a photographed image of the soil material;
estimating properties of the soil material by inputting the input data into the trained model;
The soil property estimation method according to claim 1, further comprising: further comprising a step of performing image processing on the photographed data;
3. The method of estimating soil properties according to claim 1 , wherein in the step of generating the learned model, the machine learning is performed using the photographed data subjected to the image processing. a step of obtaining, as input data, photographed data representing a photographed image of a soil material;
By inputting the input data into a trained model that has undergone machine learning of the correlation between the photographed data representing the photographed image of the soil material and the analysis result data indicating the analysis result of analyzing the characteristics of the soil material estimating properties of the soil material;
A soil property estimation method executed by an information processing device ,
The analysis result data includes (i) parameter values of a predetermined function when a particle size accumulation curve indicating the analysis result of the particle size distribution of the soil material is approximated to the function, (ii) grains of the soil material Soil characteristics characterized by including at least one of the weight percentage of passing material for each range when the diameter is divided into a plurality of ranges, and (iii) data showing the measurement result of the water content ratio of the soil material. estimation method . an input data acquisition unit that acquires input data representing an image of a soil material;
The input data is obtained by inputting the correlation between the photographed data representing the photographed image of the soil material and the analysis result data indicating the analysis result of analyzing the characteristics of the soil material into a trained model that has undergone machine learning. and an output unit for outputting output data indicating characteristics of the soil material, comprising:
The analysis result data includes (i) parameter values of a predetermined function when a particle size accumulation curve indicating the analysis result of the particle size distribution of the soil material is approximated to the function, (ii) grains of the soil material A device characterized by including at least one of the weight percentage of passing mass for each range when the diameter is divided into a plurality of ranges, and (iii) data indicating the measurement result of the water content ratio of the soil material. The analysis result data includes parameter values of a predetermined function when a particle size accumulation curve indicating the analysis result of the particle size distribution of the soil material is approximated to the function,
6. The apparatus according to claim 5 , wherein the output unit outputs a particle size additive curve appearing by applying the parameter values obtained by the trained model to the function. 6. A program for causing a computer to function as the device according to claim 5 , the program for causing the computer to function as the input data acquisition section and the output section. An estimating device and a terminal device,
The estimation device is
an input data receiving unit for receiving input data representing a photographed image of a soil material;
By inputting the input data into a trained model that has undergone machine learning of the correlation between the photographed data representing the photographed image of the soil material and the analysis result data indicating the analysis result of analyzing the characteristics of the soil material an estimating unit that estimates the properties of the soil material and transmits output data indicating the estimation result;
with
The terminal device
an acquisition unit that acquires photographed data representing an image of the soil material photographed as input data;
a receiving unit that transmits the input data to the estimating device and receives the output data transmitted from the estimating device as a response to the input data;
an output unit that outputs data related to the characteristics represented by the output data ,
The analysis result data includes (i) parameter values of a predetermined function when a particle size accumulation curve indicating the analysis result of the particle size distribution of the soil material is approximated to the function, (ii) grains of the soil material A system characterized by including at least one of the weight percentage of passing mass for each range when the diameter is divided into a plurality of ranges, and (iii) data showing the measurement result of the water content ratio of the soil material. .

Images