JP7054641B2 - Structure characterization system - Google Patents

Description

The present invention relates to a structure property evaluation system.

Conventionally, when designing seismic structures of structures such as bridges, it has been proposed to construct a model of the structure related to the target structure and perform analysis by the finite element method (FEM) (FEM). For example, see Patent Document 1). By performing FEM analysis, various information (structural characteristic parameters) related to the structural characteristics of the structure can be obtained.

Japanese Unexamined Patent Publication No. 2006-195713

However, depending on the structure, it may be difficult to apply the FEM analysis because there is not enough detailed information about the structure for performing the FEM analysis.

The present invention has been made in view of the above, and an object of the present invention is to provide a structure characteristic evaluation system capable of appropriately and easily acquiring structural characteristic parameters of a structure.

In order to achieve the above object, the structure characteristic evaluation system according to one embodiment of the present invention comprises a moving image acquisition unit that acquires a moving image of an image of a structure and a moving image acquired by the moving image acquisition unit. A parameter that calculates the structural characteristic parameter related to the structure by the feature amount calculation unit that acquires the feature amount related to the displacement of the structure and the feature amount by substituting the feature amount into the mechanical model related to the structure and performing reinforcement learning. It has a calculation unit and.

INDUSTRIAL APPLICABILITY According to the present invention, a structure characteristic evaluation system capable of appropriately and easily acquiring structural characteristic parameters of a structure is provided.

It is a figure which shows the structure of the structure property evaluation system which concerns on embodiment of this invention. It is a figure which shows the structure of the feature amount calculation part. It is a figure which shows typically the example of the moving image which copied the object structure, and the division into an element. This is an example of a graph of vibration data of a structure. It is a figure explaining the outline of the algorithm used for reinforcement learning. It is a flowchart which shows the process executed by the structure characteristic evaluation system which concerns on embodiment of this invention. It is a figure which shows the example of the displacement when a vehicle passes. It is a figure which shows the structure of the feature amount calculation part which concerns on the modification. It is a figure which shows the hardware configuration of the server included in the structure characteristic evaluation system which concerns on embodiment of this invention.

Hereinafter, embodiments for carrying out the present invention will be described in detail with reference to the accompanying drawings. In the description of the drawings, the same elements are designated by the same reference numerals, and duplicate description will be omitted.

FIG. 1 is a diagram showing a structure characteristic evaluation system 1 according to the present embodiment. Further, FIG. 2 is a diagram showing a feature amount calculation unit included in the structure characteristic evaluation system 1. As shown in FIG. 1, the structure characteristic evaluation system 1 includes a camera 10 and a server 20. The structure characteristic evaluation system 1 is a system for calculating structural characteristic parameters of a structure. The structure to be judged is a bridge, a building, or the like. Further, the structural characteristic parameters of the structure include, for example, the shape (thickness of each part, etc.) and material characteristics (density, Young's modulus, etc.) of the members constituting the structure, but are not limited thereto. .. The structure characteristic evaluation system 1 has a function of estimating the structural characteristic parameters of each part when it is assumed that the structure operates based on a certain dynamic model. The structural characteristic parameters calculated by the structural characteristic evaluation system 1 are used, for example, to determine the construction contents such as reinforcement or repair of the structure. In the structure characteristic evaluation system 1, the server 20 calculates the structural characteristic parameters using the moving image captured by the camera 10.

Subsequently, the configurations of the camera 10 and the server 20 included in the structure characteristic evaluation system 1 will be described. The camera 10 is an imaging device that captures a structure for which structural characteristic parameters are calculated and acquires a moving image of the structure. The camera 10 is fixedly installed in advance at a position where the structure can be imaged. As the camera 10, a well-known camera capable of taking an image with a resolution capable of detecting vibration or the like of each part of the structure can be used. The camera 10 and the server 20 can send and receive information to and from each other. The camera 10 images the structure at a preset frame rate (fps) and transmits the captured moving image to the server 20.

Information related to the displacement of the structure when an external force is applied is used for the calculation of the structural characteristic parameter by the structure characteristic evaluation system 1. Therefore, the moving image captured by the camera 10 is a moving image for detecting the displacement of the structure. When the structure is a bridge, for example, the camera 10 captures vibrations caused by wind, vibrations caused by the passage of a vehicle, and the like. Then, the obtained moving image is transmitted from the camera 10 to the server 20.

As shown in FIG. 1, functionally, the server 20 has a moving image acquisition unit 21, a moving image data holding unit 22, a feature amount calculation unit 23, an initial value data holding unit 24, a dynamic model holding unit 25, and a dynamic model. It includes an operation unit 26 and a characteristic parameter holding unit 27. The initial value data holding unit 24, the dynamic model holding unit 25, and the mechanical model operation unit 26 serve as parameter calculation units for calculating structural characteristic parameters related to the structure by reinforcement learning in the structure characteristic evaluation system 1 according to the present embodiment. Function.

The moving image acquisition unit 21 is a functional unit that acquires a moving image of an image of a structure. The moving image acquisition unit 21 acquires a moving image of the structure. The moving image acquisition unit 21 outputs the acquired moving image to the moving image data holding unit 22.

The moving image data holding unit 22 is a functional unit that stores data (moving image data) related to the moving image acquired by the moving image acquisition unit 21. The saved moving image data is sent to the feature amount calculation unit 23 and used for the feature amount calculation by the feature amount calculation unit 23.

The feature amount calculation unit 23 is a functional unit that detects the feature amount related to the displacement of the structure from the moving image data. The "feature amount related to the displacement of the structure" includes, for example, the feature amount related to the vibration at a plurality of parts of the structure. However, information different from the feature amount related to the vibration at a plurality of parts of the structure may be used as the feature amount related to the structure. Specifically, as the feature amount related to the displacement of the structure, the feature amount related to the deflection of the structure or the feature amount related to the strain of the structure can also be used. In the present embodiment, the feature amount calculation unit 23 detects the displacement in subpixel units from the moving image by a conventional technique such as POC (Phase Only Correlation), and vibrates at a plurality of parts of the structure. A case of calculating the feature amount related to the displacement by detecting it will be described.

Specifically, the feature amount calculation unit 23 includes a moving image dividing unit 31, a moving image selection unit 32, an element dividing unit 33, and a feature amount calculating unit 34.

First, as shown in FIG. 3, the moving image dividing unit 31 divides the acquired moving image into individual frames (images) constituting the moving image. Next, the moving image selection unit 32 selects moving image data (frame data included in the moving image data) to be used for calculating the feature amount from the moving image data divided into individual frames. A plurality of moving image data may be selected, or the moving image data in a part of the time zone (for example, the time zone in which the target structure is displaced) is separated from one moving image data. May be selected. The element division unit 33 extracts (divides into partial images) partial images (phase images) of a plurality of regions at preset positions in each frame included in the moving image data selected by the moving image selection unit 32. ). In this embodiment, each region is referred to as an element E. The size of the element E is preset. The position of the element E is the position where the structure is shown in the moving image. The position of the element E is, for example, set in advance by the user of the structure characteristic evaluation system 1 and stored in the feature amount calculation unit 23. Alternatively, the feature amount calculation unit 23 may detect the position where the structure is shown in the moving image and set the detected position as the position of the element E.

Next, the feature amount calculation unit 34 calculates the positional deviation of the same (positional) element E between the preceding and following frames with sub-pixel accuracy. The time difference Δt between the preceding and following frames is the reciprocal of the frame rate (fps) (Δt = 1 / fps). The feature amount calculation unit 34 calculates the positional deviation in the y direction of the moving image, for example, when determining the state of the bridge. This is because the deterioration of the bridge proceeds in response to vibration in the vertical direction (vertical direction). The feature amount calculation unit 34 calculates the positional deviation (displacement y) between the frames before and after each element from the moving image as the feature amount related to the vibration of the structure. As shown in FIG. 4, the feature amount related to the vibration obtained from the moving image by the feature amount calculation unit 34 is a function related to the element i and the value of the displacement y for each time (time) t, that is, y (i, t). ). The feature amount calculation unit 34 outputs the obtained function related to the vibration to the dynamic model operation unit 26 as feature amount information.

The initial value data holding unit 24 is a functional unit that holds information related to the initial values of the parameters used in the mechanical model corresponding to the structure. The information related to the initial value held in the initial value data holding unit 24 is the information related to the structural characteristic parameter as the initial value applied to the mechanical model when the structural characteristic parameter is calculated using the mechanical model described later. Is. Therefore, if there is information about the structural characteristic parameters set at the time of designing the structure, the information about the structural characteristic parameters at the time of designing is held in the initial value data holding unit 24 as initial value data. If there is no information on the structural characteristic parameters at the time of design, for example, refer to the design document, for example, information on general characteristics of the material, or in the case of a member standardized by JIS, etc. Dimensional information specified in the standard can be used as the initial value data. Further, the structural characteristic parameters previously calculated by using the structural characteristic evaluation system 1 can be retained as initial value data. In this way, the initial value data applied to the dynamic model can be changed as appropriate.

The mechanical model holding unit 25 is a functional unit that holds information related to the mechanical model used for calculating the structural characteristic parameters of the structure. The mechanical model used for calculating the structural characteristic parameters of the structure is a known physical model related to the structure. For example, when the structure is a bridge, the governing equation (governing differential equation) of the beam is used as a dynamic model. As the governing equation of the beam, the so-called Euler-Bernoulli beam, the governing equation related to the Timoshenko beam and the like are known. For example, in the case of Euler-Bernoulli beam, the following equation (1) is a dynamic model based on the governing equation.

In addition, F (i, t) in the above formula (1) is a function related to an external force, and ρ, A, E, I and the like are structural characteristic parameters in this embodiment (for example, E is a structure). Young's modulus). The structural characteristic parameters that can be calculated using the dynamic model depend on what structural characteristic parameters the dynamic model contains.

The mechanical model holding unit 25 holds a mechanical model used for calculating structural characteristic parameters. Since the mechanical model to be used can be selected according to the type of structure (bridge, steel tower, etc.), it can be configured to be held for each type of structure. Further, even in the same type of structure, for example, the mechanical model to be used may be changed according to the structure (shape) or the like. Therefore, the mechanical model holding unit 25 may hold mathematical formulas obtained from a plurality of modeling of the same type of structure (for example, a bridge), that is, a plurality of types of mechanical models.

The mechanical model operation unit 26 obtains the dynamic model held by the dynamic model holding unit 25, the information related to the initial value held by the initial value data holding unit 24, and the feature amount obtained by the feature amount calculation unit 23. In combination, it has a function to calculate structural characteristic parameters.

Regarding the calculation of structural characteristic parameters, a type of machine learning is performed after applying the feature quantities related to the vibration of the structure obtained from the moving image of the camera 10 to the dynamic model to which the information related to the initial values is applied. The structural feature parameters are calculated using Reinforcement Learning.

Examples of the reinforcement learning algorithm that can be used for calculating the structural characteristic parameters according to the present embodiment include Actor-Critic, Q-learning, and Deep-Q-Network. In this embodiment, the case where the Actor-Critic algorithm is applied among the reinforcement learning algorithms mentioned above will be outlined. However, the method is not limited to these reinforcement learning methods, and can be appropriately changed according to the mechanical model, the type of structural characteristic parameter, and the like.

Generally, in reinforcement learning, when an agent, environment, behavior, and reward are given, the work of learning to maximize the reward (measure) is performed. Further, TD error learning (Temporal Difference Learning) is known as one of the general methods of reinforcement learning. The learning method of TD error learning is to evaluate oneself and update the evaluation value. In order to update the evaluation value, the work of bringing the error (TD error) between the estimated evaluation value and the evaluation value obtained when actually acting is brought closer to 0 is performed.

The Actor-Critic algorithm is a method of TD error learning, and has a feature that an agent is composed of an Actor and a Critic. An example of the Actor-Critic arithmetic process is shown in FIG.

As shown in FIG. 5, the Actor first recognizes the state _St from the environment. Then, the agent determines a _policy and takes an action at based on the probability density function given to the Actor. As a result, the state _shifts from St to _{St + 1} . As a result, Critic receives a reward rt _{+ 1} in response to changes in state.

そして、Criticは、評価値のＴＤ誤差が０に近付くように、状態価値関数を更新する。更新は以下の数式（３）に基づいて行われる。

なお、数式（２）、（３）において、γは割引率（未知の状態評価関数であるために発生する割引を示す）であり、αは学習率（次の状態へのＴＤ誤差の反映率を示す）である。 At this time, Critic compares the estimated value with the reward and calculates the TD error. Assuming that the evaluation value at time t is the state value function V (St), the TD error δ _t at time _t + 1 can be calculated by the following mathematical formula (2).

Then, Critic updates the state value function so that the TD error of the evaluation value approaches zero. The update is performed based on the following formula (3).

In the formulas (2) and (3), γ is the discount rate (indicating the discount generated because it is an unknown state evaluation function), and α is the learning rate (the reflection rate of the TD error in the next state). Shows).

Also, in the Actor, referring to the TD error, change the policy so that the TD error becomes 0, and take action (at _{+ 1} ). By repeating this, the situation where the TD error becomes smaller is reached. can do.

The above Actor-Critic algorithm is applied to the mechanical model and structural characteristic parameters according to this embodiment. First, the left side of the mechanical model (governing equation) used for calculating the structural characteristic parameters is applied to the state St in Actor-Critic. Further, the action at corresponds to the update ( _numerical change) of the structural characteristic parameter included in the dynamic model. The probability density function corresponding to the policy is represented by a function _{π Φ} (at | _st ), and is a _probability density function that takes an action at when the state _st . Φ is a policy parameter and means a set according to the initial value of the structural characteristic parameter and the parameter that specifies the shape of the probability density function.

なお、上記は一例である。構造特性パラメータが互いに影響し合うと改定して、多変量正規分布を利用して確率密度関数を記述してもよい。 The probability density function corresponding to the above measure is based on the following equation (4), assuming that the structural characteristic parameters contained in the mechanical model are independent of each other, and that each parameter is normally distributed. ) Can be expressed. In the formula (4), Z is a normalization constant. Further, the mean value μ and the variance σ included in the formula (4) can be updated as in the following formulas (5) and (6), respectively.

The above is an example. The probability density function may be described by using the multivariate normal distribution, which is revised when the structural characteristic parameters influence each other.

As the reward _rt , a reward corresponding to the error in the state _st as a _result of performing the action at is set. The error is the state s when the free _damping vibration after the load contributing to the vibration disappears in the governing equation (mathematical formula (1)) that calculated the mechanical model st as a _result of performing the action at. If the above error, which is the calculation of the error for the free damped vibration of _t , is equal to or less than the predetermined threshold value, a positive reward is given according to the value, and if it is larger than the predetermined threshold value ε, the error is given. , Negative reward will be given according to the value. With such a configuration, the reward can be set so that it is preferable that the error is small.

In the dynamic model operation unit 26, as described above, the Actor's policy is changed so that the TD error becomes small, and the state value function of Critic is updated by the number of features obtained from the moving image data. .. After the iteration is finished, the structural characteristic parameters are determined based on the Actor's strategy. As a result, structural characteristic parameters after reinforcement learning are obtained.

The characteristic parameter holding unit 27 is a functional unit that holds the structural characteristic parameters calculated by the mechanical model operation unit 26. The structural characteristic parameters held by the characteristic parameter holding unit 27 are output from the server 20 to an external device or the like as needed.

Since the structural characteristic parameters calculated from the moving image selected by the moving image selection unit 32 are the characteristics of the structure under certain conditions, there is a possibility that they are over-learned. Therefore, by using the initial value as the structure characteristic parameter calculated last time and repeating the learning again using different moving images, it is possible to calculate the structure parameter considered to be optimal for all situations. .. Therefore, the structural characteristic parameters calculated by the dynamic model operation unit 26 are not only held by the characteristic parameter holding unit 27 but also held by the initial value data holding unit 24, and are appropriately used by the mechanical model operation unit 26. The default.

Next, a process (operation method performed by the structure characteristic evaluation system 1) executed by the structure characteristic evaluation system 1 according to the present embodiment will be described with reference to the flowchart shown in FIG.

First, a moving image of a structure (structure) is photographed by the camera 10. The captured moving image data is sent from the camera 10 to the server 20 and acquired by the moving image acquisition unit 21 of the server 20 (S01). The moving image data acquired by the moving image acquisition unit 21 is held by the moving image data holding unit 22.

Next, the moving image selection unit 32 selects the moving image data to be used for calculating the feature amount related to the structure. Next, in the element dividing unit 33, it is determined whether or not there are a plurality of frames related to the same structure by referring to the moving image data (S02). When there are not a plurality of frames in which the same structure is imaged (S02-NO), information related to vibration cannot be obtained, so the process is repeated from the acquisition of the moving image (S01), or the moving image selection unit 32 is used. Repeat from the selection of moving image data. On the other hand, when there are a plurality of frames in which the same structure is imaged (S02-YES), the feature amount is calculated using the frames. First, in the element division unit 33, the structure included in each moving image (frame) is divided into a plurality of elements (S03). Next, in the feature amount calculation unit 34, for each divided element, the position shift (displacement y), which is a minute movement amount between the frames before and after each element, is compared with other frames in chronological order. It is calculated as a feature amount related to the vibration of an object (S04). The function y (i, t) related to the displacement y is a space-time function. The feature amount related to the vibration thus obtained is sent from the feature amount calculation unit 23 to the mechanical model operation unit 26.

Next, the dynamic model operation unit 26 selects a mechanical model to be used for calculating the structural characteristic parameters from the mechanical models held by the dynamic model holding unit 25. Then, the differential coefficient required for the dynamic model is calculated based on the function y (i, t) (spatiotemporal function) which is the feature amount related to the vibration sent from the feature amount calculation unit 23, and the calculation result is the dynamics. Substitute in the model (S05). As illustrated in the above formula, the dynamic model may include the derivative of the function y related to vibration. If the dynamic model operation unit 26 has a differential coefficient used in the dynamic model, the differential coefficient is calculated from a function that is a feature quantity, and this is applied to the dynamic model.

Next, the dynamic model operation unit 26 updates the structural characteristic parameters included in the dynamic model by reinforcement learning (S06). As the initial value of the structural characteristic parameter, the value held in the initial value data holding unit 24 is used. Further, when using the same dynamic model as the previous one, as described above, the structural characteristic parameters updated by the previous reinforcement learning can be used. Further, for reinforcement learning, for example, an Actor-Critic algorithm or the like can be used. By performing reinforcement learning, the structural characteristic parameters are updated. After that, the structural characteristic parameters obtained by reinforcement learning are output (S07).

The process related to the calculation of the structural characteristic parameters by the mechanical model operation unit 26 can be appropriately changed. Hereinafter, some modification examples will be described.

As described above, the same structure may be divided into a plurality of elements, and the feature amount may be calculated for each element. In this case, structural characteristic parameters are calculated by reinforcement learning for each feature amount. That is, by obtaining a plurality of feature quantities from the moving image data obtained by imaging the same structure, a plurality of calculation results (for each element obtained in the moving image) can be obtained for the structural characteristic parameters of the same item. In some cases. In that case, the configuration may be such that "one structural characteristic parameter" to be output as "structural characteristic parameter obtained from moving image data" is determined based on a plurality of structural characteristic parameters related to the same item. When a plurality of structural characteristic parameters related to the same item are obtained, how to determine "one structural characteristic parameter" can be appropriately changed. For example, the average value of a plurality of parameters obtained for the same item may be used as "one structural characteristic parameter". Further, depending on the characteristics of the item of the structural characteristic parameter, for example, the minimum value among the plurality of parameters obtained for the same item may be set as "one structural characteristic parameter".

Further, the dynamic model operation unit 26 evaluates the calculation result at the time of updating (S06) by reinforcement learning of the structural characteristic parameters included in the dynamic model, and recalculates if necessary based on the evaluation result. May be. For example, if the variance of the structural characteristic parameter calculated as a result of reinforcement learning is larger than a predetermined threshold value, a part of the dynamic model may be changed based on a predetermined condition to perform recalculation. good.

Further, the dynamic model holding unit 25 holds in advance a plurality of mechanical models in which the setting of the boundary conditions (limit values of each parameter, etc.) when calculating the structural characteristic parameters is changed, and the mechanical model operation unit 26 holds the structure. It may be configured to select an appropriate one based on the calculation conditions of the characteristic parameters and perform the calculation. Further, the variance of the probability density function used in the policy of the dynamic model used for calculating the structural characteristic parameters may be changed according to the fluctuation of the surrounding environment of the structure (for example, seasonal fluctuation).

Further, in the case of reinforcement learning using the dynamic model by the mechanical model operation unit 26, the values that can be taken by each structural characteristic parameter are restricted in advance to prevent an erroneous value from being calculated by the reinforcement learning. May be. For example, for a parameter that takes only a substantially positive value, recalculation can be performed when a value of 0 or less is selected during reinforcement learning.

By the above processing, the structural characteristic parameters related to the structure are calculated and output from the moving image data captured by the camera 10.

It is preferable that the calculation of the structural characteristic parameters using the above mechanical model is performed using a moving image including various vibrations (displacements) generated in the structure. With such a configuration, it is possible to prevent calculation of structural characteristic parameters that reflect only specific vibrations when a specific external force is applied.

On the other hand, when the structure is a bridge, the structural characteristic parameters can be calculated using only the moving image related to the vibration when the vehicle passes over the bridge.

As shown in FIG. 7, it is assumed that the vibration generated by the passage of a vehicle on the bridge changes from the forced vibration due to the passage of the vehicle to the free damping fluctuation as the time t progresses. Can be done. Therefore, it is possible to calculate the feature amount on the assumption that the bridge vibrates with the waveform shown in FIG. 7.

When calculating the feature amount from the moving image when the vehicle is passing, the feature amount calculating unit 23 is instructed by the position of the load (vehicle) that vibrates the structure in the moving image, as shown in FIG. A load situation recognition unit 35 for evaluating movement and the like is provided. The load situational awareness unit 35 acquires information related to the position and movement of the vehicle to be loaded from the moving image data, and sends this to the moving image selection unit 32, so that the moving image selection unit 32 is the vehicle to be loaded. You can select a moving image in consideration of the movement of. Further, the feature amount calculation unit 34 can calculate the feature amount in consideration of the vibration caused by the passage of the vehicle. Further, the load situational awareness unit 35 provides the dynamic model operation unit 26 with information related to the load, so that the vehicle corresponds to the dynamic model (for example, F (i, t) in the mathematical formula (1)). It may be configured to reflect the weight to be applied.

As described above, the specific procedure for calculating the feature amount can be appropriately changed in consideration of the features of the moving image data and the like.

As described above, the structure characteristic evaluation system 1 according to the present embodiment is a structure from a moving image acquisition unit 21 that acquires a moving image of an image of a structure and a moving image acquired by the moving image acquisition unit 21. The feature amount calculation unit 23 that acquires the feature amount related to the displacement of, and the parameter calculation unit (initial value) that substitutes the feature amount for the mechanical model related to the structure and calculates the structural characteristic parameter related to the structure by reinforcement learning. It has a data holding unit 24, a dynamic model holding unit 25, and a mechanical model operation unit 26). Having such a configuration makes it possible to appropriately and easily acquire the structural characteristic parameters of the structure.

Conventionally, when designing seismic structures of structures such as bridges, it is common to construct a model of the structure related to the target structure and obtain structural characteristic parameters etc. by using FEM analysis. Met. However, detailed design drawings and the like are required to create a model of the structure for performing FEM analysis. Therefore, for existing structures, it may be difficult to obtain detailed information for FEM analysis. Further, even when a design drawing or the like is obtained, some structures are constructed in a state different from the design drawing, and it may be difficult to make an accurate evaluation only with the design drawing.

Further, when calculating the structural characteristic parameters from the FEM analysis, the parameters of the model of the structure used for the FEM analysis are conventionally based on the sensor information obtained by installing a sensor such as an acceleration sensor on the structure. It was necessary to adjust. In order to obtain the sensor information, it is necessary to actually attach the sensor or the like to the structure, so that the work cost related to the acquisition of the sensor information is incurred. In addition, regarding the adjustment of the parameters of the model of the structure based on the sensor information, there is a problem that not only the work cost is incurred but also a specific engineer who has acquired the technique related to the adjustment of the parameters needs to deal with it. ..

On the other hand, in the structure characteristic evaluation system 1 according to the present embodiment, the information required for calculating the structural characteristic parameters is the feature amount related to the displacement of the structure obtained from the moving image. Further, in the structure characteristic evaluation system 1, the structural characteristic parameters can be calculated by applying this feature quantity to the mechanical model in the mechanical model operation unit 26 and performing reinforcement learning. Therefore, it is possible to significantly reduce the work cost for acquiring the information related to the feature amount, which is the information corresponding to the sensor information used conventionally. Further, since the structural characteristic parameters can be calculated by reinforcement learning instead of the conventional FEM analysis and parameter adjustment, it is not necessary to deal with a specific engineer, and the work cost can be reduced. In this way, it is possible to appropriately and easily acquire the structural characteristic parameters of the structure.

Further, in the above-mentioned structure characteristic evaluation system 1, the mechanical model is configured to be the governing equation of the beam.

When the structure is a bridge, FEM analysis has been performed using a complicated structural model for the bridge in the past, whereas in the structure characteristic evaluation system 1, a general beam control equation is used. The structural characteristic parameters are calculated. With such a configuration, a configuration that can be calculated more easily is realized, especially when calculating the structural characteristic parameters related to the bridge. Even if the structure is not a bridge, if the governing equation of the beam can be used as a dynamic model, the structural characteristic parameters can be calculated using a structural model that is simpler than the conventional one.

Further, the above-mentioned parameter calculation unit (initial value data holding unit 24, mechanical model holding unit 25, mechanical model operation unit 26) is described in the design document relating to the structure as the initial value of the structural characteristic parameter calculated by reinforcement learning. Apply the information provided. In this way, by configuring the reinforcement learning after reflecting the information described in the design document, the structural characteristic parameters by reinforcement learning are compared with the case where a value unrelated to the structure is applied as the initial value. Can be calculated more quickly and accurately.

In the above embodiment, the case where the structure characteristic evaluation system 1 is applied to a bridge has been described, but the structure (building) other than the bridge is also a structure in which displacement occurs due to wind, traffic vibration, or the like. If so, the structural characteristic parameters can be calculated in the same manner as for bridges.

For example, even when the target structure is a utility pole or a steel tower, the structure characteristic evaluation system 1 according to the present embodiment can be applied. When the structure is a utility pole, it can be considered as a cantilever and the Euler-Bernoulli governing equation can be applied in the same way as a bridge. In addition, when the structure is a steel tower, the features of each member constituting the steel tower are calculated from the moving image data, and the dynamic model is applied individually according to the configuration of each member, similar to a bridge. Structural characteristic parameters can be calculated. In this way, the structural characteristic parameters related to the structure can be calculated by appropriately selecting and setting the mechanical model according to the shape and the like of the target structure.

Since the structure characteristic evaluation system according to the present invention only needs to be able to acquire a moving image, it may be configured to be able to acquire a moving image from other than the structure characteristic evaluation system. For example, if the server 20 according to the present embodiment is configured to be able to acquire moving images from the camera 10 (other than the structure characteristic evaluation system), even if the structure characteristic evaluation system is configured only by the server 20. good.

(others)
The block diagram used in the description of the above embodiment shows a block of functional units. These functional blocks (components) are realized by any combination of hardware and / or software. Further, the means for realizing each functional block is not particularly limited. That is, each functional block may be realized by one physically and / or logically coupled device, or directly and / or indirectly by two or more physically and / or logically separated devices. (For example, wired and / or wireless) may be connected and realized by these plurality of devices.

For example, the server 20 in one embodiment of the present invention may function as a computer that processes the server 20 in the present embodiment. FIG. 9 is a diagram showing an example of the hardware configuration of the server 20 according to the present embodiment. The server 20 described above may be physically configured as a computer device including a processor 1001, a memory 1002, a storage 1003, a communication device 1004, an input device 1005, an output device 1006, a bus 1007, and the like.

In the following description, the word "device" can be read as a circuit, a device, a unit, or the like. The hardware configuration of the server 20 may be configured to include one or more of the devices shown in the figure, or may be configured not to include some of the devices.

For each function in the server 20, by loading predetermined software (program) on hardware such as the processor 1001 and the memory 1002, the processor 1001 performs an operation, and the communication device 1004 communicates with the memory 1002 and the storage 1003. It is realized by controlling the reading and / or writing of data.

Processor 1001 operates, for example, an operating system to control the entire computer. The processor 1001 may be composed of a central processing unit (CPU) including an interface with a peripheral device, a control device, an arithmetic unit, a register, and the like. For example, each function of the server 20 may be realized by the processor 1001.

Further, the processor 1001 reads a program (program code), a software module and data from the storage 1003 and / or the communication device 1004 into the memory 1002, and executes various processes according to these. As the program, a program that causes a computer to execute at least a part of the operations described in the above-described embodiment is used. For example, each function of the server 20 may be realized by a control program stored in the memory 1002 and operated by the processor 1001. Although it has been described that the various processes described above are executed by one processor 1001, they may be executed simultaneously or sequentially by two or more processors 1001. Processor 1001 may be mounted on one or more chips. The program may be transmitted from the network via a telecommunication line.

The memory 1002 is a computer-readable recording medium, and is composed of at least one such as a ROM (Read Only Memory), an EPROM (Erasable Programmable ROM), an EEPROM (ElectricallyErasable Programmable ROM), and a RAM (Random Access Memory). You may. The memory 1002 may be referred to as a register, a cache, a main memory (main storage device), or the like. The memory 1002 can store a program (program code), a software module, or the like that can be executed to carry out the method according to the embodiment of the present invention.

The storage 1003 is a computer-readable recording medium, for example, an optical disk such as a CD-ROM (Compact Disc ROM), a hard disk drive, a flexible disk, an optical magnetic disk (for example, a compact disk, a digital versatile disk, a Blu-ray). It may consist of at least one (registered trademark) disk), smart card, flash memory (eg, card, stick, key drive), floppy (registered trademark) disk, magnetic strip, and the like. The storage 1003 may be referred to as an auxiliary storage device. The storage medium described above may be, for example, a database, server or other suitable medium containing memory 1002 and / or storage 1003.

The communication device 1004 is hardware (transmission / reception device) for communicating between computers via a wired and / or wireless network, and is also referred to as, for example, a network device, a network controller, a network card, a communication module, or the like. For example, each function of the server 20 may be realized by the communication device 1004.

The input device 1005 is an input device (for example, a keyboard, a mouse, a microphone, a switch, a button, a sensor, etc.) that receives an input from the outside. The output device 1006 is an output device (for example, a display, a speaker, an LED lamp, etc.) that outputs to the outside. The input device 1005 and the output device 1006 may have an integrated configuration (for example, a touch panel).

Further, each device such as the processor 1001 and the memory 1002 is connected by a bus 1007 for communicating information. The bus 1007 may be composed of a single bus or may be composed of different buses between the devices.

Further, the server 20 includes hardware such as a microprocessor, a digital signal processor (DSP: Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), a PLD (Programmable Logic Device), and an FPGA (Field Programmable Gate Array). The hardware may implement some or all of each functional block. For example, the processor 1001 may be implemented on at least one of these hardware.

Although the present embodiment has been described in detail above, it is clear to those skilled in the art that the present embodiment is not limited to the embodiment described in the present specification. This embodiment can be implemented as an amendment or modification without departing from the spirit and scope of the present invention as determined by the description of the scope of claims. Therefore, the description of the present specification is for the purpose of illustration and does not have any limiting meaning to the present embodiment.

The notification of information is not limited to the embodiments / embodiments described herein, and may be performed by other methods. For example, information notification includes physical layer signaling (eg, DCI (Downlink Control Information), UCI (Uplink Control Information)), higher layer signaling (eg, RRC (Radio Resource Control) signaling, MAC (Medium Access Control) signaling, etc. It may be carried out by broadcast information (MIB (Master Information Block), SIB (System Information Block)), other signals, or a combination thereof. Further, the RRC signaling may be referred to as an RRC message, and may be, for example, an RRC Connection Setup message, an RRC Connection Reconfiguration message, or the like.

The processing procedures, sequences, flowcharts, and the like of each aspect / embodiment described in the present specification may be rearranged in order as long as there is no contradiction. For example, the methods described herein present elements of various steps in an exemplary order and are not limited to the particular order presented.

The input / output information and the like may be stored in a specific place (for example, a memory) or may be managed by a management table. Information to be input / output may be overwritten, updated, or added. The output information and the like may be deleted. The input information or the like may be transmitted to another device.

The determination may be made by a value represented by 1 bit (0 or 1), by a boolean value (Boolean: true or false), or by comparing numerical values (for example, a predetermined value). It may be done by comparison with the value).

Each aspect / embodiment described in the present specification may be used alone, in combination, or may be switched and used according to the execution. Further, the notification of predetermined information (for example, the notification of "being X") is not limited to the explicit one, but is performed implicitly (for example, the notification of the predetermined information is not performed). May be good.

Software, whether called software, firmware, middleware, microcode, hardware description language, or other names, is an instruction, instruction set, code, code segment, program code, program, subprogram, software module. , Applications, software applications, software packages, routines, subroutines, objects, executable files, execution threads, procedures, features, etc. should be broadly interpreted.

Further, software, instructions, and the like may be transmitted and received via a transmission medium. For example, the software may use wired technology such as coaxial cable, fiber optic cable, twisted pair and digital subscriber line (DSL) and / or wireless technology such as infrared, wireless and microwave to website, server, or other. When transmitted from a remote source, these wired and / or wireless technologies are included within the definition of transmission medium.

The information, signals, etc. described herein may be represented using any of a variety of different techniques. For example, data, instructions, commands, information, signals, bits, symbols, chips, etc. that may be referred to throughout the above description are voltages, currents, electromagnetic waves, magnetic fields or magnetic particles, light fields or photons, or any of these. It may be represented by a combination of.

The terms described herein and / or the terms necessary for understanding the present specification may be replaced with terms having the same or similar meanings.

The terms "system" and "network" used herein are used interchangeably.

Further, the information, parameters, etc. described in the present specification may be represented by an absolute value, a relative value from a predetermined value, or another corresponding information. ..

The names used for the parameters mentioned above are not limited in any way. Further, mathematical formulas and the like using these parameters may differ from those expressly disclosed herein.

As used herein, the terms "determining" and "determining" may include a wide variety of actions. "Judgment" and "decision" are, for example, judgment, calculation, computing, processing, deriving, investigating, looking up (eg, table). , Searching in a database or another data structure), ascertaining can be considered as a "judgment" or "decision". Also, "judgment" and "decision" are receiving (for example, receiving information), transmitting (for example, transmitting information), input (input), output (output), and access. It may include (for example, accessing data in memory) to be regarded as "judgment" or "decision". In addition, "judgment" and "decision" are considered to be "judgment" and "decision" when the things such as solving, selecting, choosing, establishing, and comparing are regarded as "judgment" and "decision". Can include. That is, "judgment" and "decision" may include considering some action as "judgment" and "decision".

The terms "connected", "coupled", or any variation thereof, mean any direct or indirect connection or connection between two or more elements and each other. It can include the presence of one or more intermediate elements between two "connected" or "combined" elements. The connection or connection between the elements may be physical, logical, or a combination thereof. As used herein, the two elements are by using one or more wires, cables and / or printed electrical connections, and, as some non-limiting and non-comprehensive examples, radio frequencies. By using electromagnetic energies such as electromagnetic energies with wavelengths in the region, microwave region and light (both visible and invisible) regions, they can be considered to be "connected" or "coupled" to each other.

The phrase "based on" as used herein does not mean "based on" unless otherwise stated. In other words, the statement "based on" means both "based only" and "at least based on".

As used herein by designations such as "first", "second", etc., any reference to that element does not generally limit the quantity or order of those elements. These designations can be used herein as a convenient way to distinguish between two or more elements. Therefore, references to the first and second elements do not mean that only two elements can be adopted there, or that the first element must somehow precede the second element.

As long as "include", "including", and variations thereof are used herein or within the scope of the claims, these terms are similar to the term "comprising". In addition, it is intended to be inclusive. Moreover, the term "or" as used herein or in the claims is intended to be non-exclusive.

In the present specification, a plurality of devices shall be included unless the device has only one device apparently in context or technically. The entire disclosure is intended to include more than one, unless the context clearly indicates the singular.

1 ... structure characteristic evaluation system, 10 ... camera, 20 ... server, 21 ... moving image acquisition unit, 22 ... moving image data holding unit, 23 ... feature amount calculation unit, 24 ... initial value data holding unit, 25 ... dynamic model Holding unit, 26 ... Mechanical model operation unit, 27 ... Characteristic parameter holding unit.

Claims (3)

A moving image acquisition unit that acquires moving images of structures,
A feature amount calculation unit that acquires a feature amount related to displacement of the structure from a moving image acquired by the moving image acquisition unit, and a feature amount calculation unit.
A parameter calculation unit that substitutes the feature amount for the mechanical model related to the structure and calculates the structural characteristic parameter related to the structure by reinforcement learning.
Have,
The feature amount related to the displacement of the structure is the feature amount related to the vibration of the structure.
The feature amount calculation unit divides the moving image into individual frames constituting the moving image, and selects frame data to be used for calculating the feature amount from the moving image data divided into the individual frames. The division of each frame data included in the selected frame data into a plurality of partial images and the vertical positional deviation of the partial images between the preceding and following frames are used as feature quantities related to the vibration of the structure. A structure property evaluation system that calculates and executes . The structure characteristic evaluation system according to claim 1, wherein the mechanical model is a governing equation of a beam. The structure characteristic evaluation system according to claim 1 or 2, wherein the parameter calculation unit applies the information described in the design document relating to the structure as the initial value of the structural characteristic parameter calculated by reinforcement learning.

Images