Radiation source estimation device and radiation source estimation method

Abstract

To accurately estimate the distribution of radiation sources from a small number of measurement results.SOLUTION: A radiation source estimation apparatus 1 comprises: a radiation source vector creation part 13 which creates a radiation source vector ω indicating a radiation source that is supposed to exist in each of a plurality of regions obtained by dividing a target zone; a dosage vector creation part 15 which creates a dosage vector y indicating a measurement value at a measurement point; a matrix calculation unit 16 which calculates a matrix X corresponding to the number of radiation sources indicated by the radiation source vector ω and the number of measurement points; and a radiation source estimation part 17 which estimates the distribution of radiation sources in the target zone. The radiation source estimation part 17 considers the product of the radiation source vector ω and the matrix X as a pseudo vector of the dosage vector y, and estimates the distribution of the radiation sources by performing inverse analysis that restores the radiation source vector ω from the pseudo vector with the radiation source vector ω as a sparse vector.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a radiation source estimation device and a radiation source estimation method for estimating the distribution of a radiation source in a target area. [Background technology]

[0002] In order to grasp the contamination status in an area contaminated by radioactive materials, it is important to measure the radiation dose using a measuring device and estimate the distribution of the radiation source, which is the contaminated area. One of the techniques for estimating the radiation source is known, for example, as disclosed in Patent Document 1.

[0003] Patent Document 1 discloses a technique related to an energy source estimation method that searches for a pattern vector d of an energy source that generates a pattern vector that best approximates a pattern vector u whose elements are the distribution of a measured field. The technique disclosed in Patent Document 1 is characterized in that the pattern matching between the pattern vectors u and d is evaluated on wavelet spectrum vectors U and D obtained by subjecting the pattern vectors u and d to discrete wavelet transform, respectively. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 08-334538 Summary of the Invention [Problem to be solved by the invention]

[0005] In order to accurately estimate the distribution of radiation sources in a contaminated area of ​​radioactive materials, it is necessary to measure radiation doses at many measurement points. This type of measurement involves exposure to radiation and requires time and effort. In particular, when there are many radiation sources with high radiation doses, this type of measurement leads to an increase in the amount of exposure. Moreover, when the contaminated area is inside a building with a complex structure, this type of measurement is even more difficult. Therefore, since it is practically difficult to obtain measurement results at many measurement points in a contaminated area of ​​radioactive materials, there is a demand for a technology that can accurately estimate the distribution of radiation sources from a small number of measurement results. The technology disclosed in Patent Document 1 does not take such circumstances into consideration at all, and there is room for improvement.

[0006] The present invention has been made in consideration of the above circumstances, and has an object to provide a radiation source estimation device and a radiation source estimation method that are capable of accurately estimating the distribution of radiation sources from a small number of measurement results. [Means for solving the problem]

[0007] In order to achieve the above object, a radiation source estimation device of the present invention is a radiation source estimation device that estimates a distribution of radiation sources in a target area from measured values ​​of radiation doses at measurement points within the target area, and includes a radiation source vector creation unit that creates radiation source vectors indicative of radiation sources hypothesized to exist in each of a plurality of areas obtained by dividing the target area, a dose vector creation unit that creates dose vectors indicative of the measured values ​​at the measurement points, a matrix calculation unit that calculates a matrix corresponding to the number of radiation sources and the number of measurement points indicated by the radiation source vector, and a radiation source estimation unit that estimates the distribution of the radiation sources in the target area, wherein the radiation source estimation unit estimates the distribution of the radiation sources by regarding a product of the radiation source vector and the matrix as a pseudo-vector of the dose vector and treating the radiation source vector as a sparse vector and performing an inverse analysis to restore the radiation source vector from the pseudo-vector.

[0008] With this configuration, the radiation source estimation device can accurately estimate the distribution of radiation sources from a small number of radiation dose measurement results in a target area such as a contaminated area of ​​radioactive materials. As a result, the radiation source estimation device can accurately estimate the distribution of radiation doses in the target area, and can accurately grasp the contamination status of radioactive materials. In particular, the radiation source estimation device can accurately estimate the distribution of radiation sources from a small number of radiation dose measurement results even inside a building having a complex structure. As a result, the radiation source estimation device can be effectively applied to high-dose areas where it is difficult to obtain measurement results due to exposure of workers and measuring instruments, and can reduce the effort and labor of measurement and the amount of radiation exposure. Note that the present invention is the world's first to realize the ability to estimate the distribution of radiation sources from a small number of measurement results by inverse analysis in the field of radiation measurement.

[0009] In a further preferred embodiment, the radiation source estimation unit performs the inverse analysis using LASSO, which searches for the radiation source vector that minimizes an objective function obtained by adding a regularization term including an L1 norm of the radiation source vector to an error term including a square of the L2 norm of the difference between the dose vector and the pseudo vector.

[0010] The distribution of radiation sources in the target area is often such that high-intensity radiation sources are sparsely present. In other words, the radiation dose is often zero or uniform in areas other than those where high-intensity radiation sources are present. This means that the distribution of radiation sources in the target area is characterized by being sparse. By focusing on this characteristic, the radiation source estimation unit performs inverse analysis using LASSO, which can significantly reduce the amount of calculation compared to a method of searching for a radiation source vector that minimizes an objective function including the L0 norm, and can efficiently restore the radiation source vector. Therefore, the radiation source estimation device can accurately and efficiently estimate the distribution of radiation sources from the results of measurement of a small number of radiation doses.

[0011] In a further preferred embodiment, the radiation source estimation unit groups, among the radiation sources indicated by the radiation source vectors, the radiation sources corresponding to predetermined structures existing in the target area, and performs the inverse analysis using Grouped LASSO.

[0012] With this aspect, the radiation source estimation unit can accurately estimate the distribution of radiation sources in accordance with reality, taking into account the influence of complex structures present in the target area on the distribution of radiation sources. Therefore, even if the target area has a complex structure, the radiation source estimation device can accurately estimate the distribution of radiation sources from a small number of measurement results.

[0013] In a further preferred embodiment, the radiation source estimation device further includes a visualization unit that creates a map that visualizes the distribution of the radiation sources in the target area, and the visualization unit creates a VR image or an AR image by superimposing the distribution of the radiation sources estimated by the radiation source estimation unit on a CG model or a captured image of the target area, and creates the map using the created VR image or AR image.

[0014] With this aspect, the radiation source estimation device can notify the worker of, for example, where a high-intensity radiation source is present in the target area in a format that the worker can intuitively understand, and thus can effectively support decontamination work and the like performed by the worker at the work site.

[0015] In a further preferred aspect, the radiation source estimation device further comprises a dust behavior calculation unit which calculates the behavior of dust floating or deposited in the target area due to wind conditions, wherein the radiation source vector creation unit creates the radiation source vector by assuming that the radiation source includes the dust, the matrix calculation unit calculates the matrix including a matrix indicating the degree to which radiation emitted from the dust to the measurement point contributes to the measured value, the radiation source estimation unit estimates a distribution of the radiation source including the dust at a predetermined time, the dust behavior calculation unit calculates the position of the dust after a predetermined time has elapsed from the predetermined time using the estimated distribution of the dust at the predetermined time as an initial condition, and the radiation source estimation unit estimates the distribution of the radiation source after the predetermined time has elapsed from the predetermined time based on a calculation result of the dust behavior calculation unit.

[0016] This aspect allows the radiation source estimation unit to estimate the distribution of radiation sources in the target area by considering not only static radiation sources but also dynamic radiation sources. Therefore, the radiation source estimation unit can estimate how the distribution of radiation sources in the target area changes over time, and can estimate the distribution of radiation sources more accurately. Therefore, the radiation source estimation device can estimate the distribution of radiation sources with higher accuracy from a small number of measurement results.

[0017] In a further preferred aspect, the radiation source estimation device further includes an estimation accuracy determination unit that determines whether or not an estimation accuracy of the distribution of the radiation sources by the radiation source estimation unit is equal to or higher than a predetermined standard, and a re-estimation instructing unit that, when it is determined that the estimation accuracy is below the standard, causes the radiation source estimation unit to re-estimate the distribution of the radiation sources by adding new measurement values ​​at new measurement points.

[0018] With this aspect, even if the estimation accuracy of the radiation source estimation unit is insufficient due to a small number of measurement points, the radiation source estimation device can increase the number of measurement points and gradually improve the estimation accuracy. Therefore, the radiation source estimation device can accurately estimate the distribution of radiation sources from a minimum number of measurement results.

[0019] Further, a radiation source estimation method of the present invention is a radiation source estimation method for estimating a distribution of radiation sources in a target area from measured values ​​of radiation doses at measurement points within the target area, the method comprising: creating a radiation source vector indicating radiation sources hypothesized to exist in each of a plurality of areas obtained by dividing the target area; creating a dose vector indicating the measured values ​​at the measurement points; calculating a matrix corresponding to the number of the radiation sources indicated by the radiation source vector and the number of the measurement points; and estimating the distribution of the radiation sources in the target area, wherein estimating the distribution of the radiation sources includes regarding a product of the radiation source vector and the matrix as a pseudo-vector of the dose vector, treating the radiation source vector as a sparse vector, and performing an inverse analysis to restore the radiation source vector from the pseudo-vector, thereby estimating the distribution of the radiation sources.

[0020] With this configuration, the radiation source estimation method can accurately estimate the distribution of radiation sources in a target area such as an area contaminated by radioactive materials from a small number of radiation dose measurement results. Effect of the Invention

[0021] According to the present invention, it is possible to provide a radiation source estimation device and a radiation source estimation method capable of estimating the distribution of radiation sources with high accuracy from a small number of measurement results. [Brief description of the drawings]

[0022] [Figure 1] 1 is a diagram showing the configuration of a radiation source estimation apparatus according to an embodiment of the present invention. [Diagram 2] FIG. 2 is a diagram for explaining an outline of a method for estimating a radiation source by the radiation source estimation device. [Diagram 3] FIG. 1 is a diagram showing the relationship between a source vector, a dose vector, and a matrix. [Figure 4] FIG. 1 shows an example of a 3D CG model of a target area. [Diagram 5] FIG. 13 is a diagram showing an example of a map visualizing the distribution of radiation sources. [Figure 6] 4 is a flowchart showing an outline of a process performed by the radiation source estimation device. [Figure 7] FIG. 2 is a diagram for explaining prerequisites for verifying the validity of the radiation source estimation device. [Figure 8] 8(a) is a diagram showing the distribution of radiation sources in the test data shown in FIG. 7, and FIG. 8(b) is a diagram showing the estimation result of the distribution of radiation sources by the radiation source estimation device. [Figure 9] FIG. 9(a) is a graph showing the calculation results of the radiation dose distribution based on the radiation source of the test data shown in FIG. 8(a), and FIG. 9(b) is a graph showing the calculation results of the radiation dose distribution based on the estimated radiation source results shown in FIG. 8(b). [Figure 10] 11 is a graph comparing the radiation source of the test data with the estimation result of the radiation source by the radiation source estimation device. [Figure 11] 11 is a graph comparing the calculation results of radiation doses based on the radiation source of the test data and the calculation results of radiation doses based on the estimation results of the radiation source by the radiation source estimation device. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0023] Hereinafter, embodiments of the present invention will be described with reference to the drawings. Components with the same reference numerals in each embodiment have the same functions in each embodiment unless otherwise specified, and description thereof will be omitted.

[0024] [Configuration of radiation source estimation device] Fig. 1 is a diagram showing the configuration of a radiation source estimation device 1 of this embodiment. Fig. 2 is a diagram for explaining an outline of a radiation source estimation method by the radiation source estimation device 1. Fig. 3 is a diagram showing the relationship between a radiation source vector ω, a dose vector y, and a matrix X. Fig. 4 is a diagram showing an example of a three-dimensional CG model of a target area. Fig. 5 is a diagram showing an example of a map visualizing the distribution of radiation sources.

[0025] The radiation source estimation device 1 is a device equipped with a system that estimates the distribution of radiation sources in a target area by inverse analysis from measured values ​​of radiation doses at measurement points in the target area. The target area is an area for which the distribution of radiation sources is to be estimated. The target area is, for example, an area contaminated by radioactive materials, such as a nuclear power plant where a radioactive material leakage accident has occurred and the surrounding area. In this embodiment, the inside of a nuclear power plant building, which has a complex structure, is taken as an example of the target area to be described.

[0026] The radiation source estimation device 1 is configured by a so-called computer system. The radiation source estimation device 1 includes an input device 2, a communication device 3, a storage device 4, a display device 5, and a calculation processing device 10. The input device 2 is configured by a keyboard, a mouse, or the like. The communication device 3 is configured by a communication device that communicates with an external device via a network. The storage device 4 is configured by an SSD, a HDD, or the like. The storage device 4 stores various data and the like used in the processing of the calculation processing device 10. The display device 5 is configured by various displays such as a liquid crystal display that displays the processing results of the calculation processing device 10 and the like. The calculation processing device 10 is configured by a CPU, a ROM, a RAM, and the like. The calculation processing device 10 realizes various functions of the radiation source estimation device 1 by the CPU executing a program stored in the ROM.

[0027] The hardware configuration of the radiation source estimation device 1 is not particularly limited to the configuration shown in Fig. 1. Some functions of the arithmetic processing device 10 may be realized in cooperation with an external computer (e.g., a supercomputer) connected to the communication device 3 via a computer network.

[0028] The calculation processing device 10 includes a target area information acquisition unit 11, a model creation unit 12, a radiation source vector creation unit 13, a measurement information acquisition unit 14, a dose vector creation unit 15, a matrix calculation unit 16, a radiation source estimation unit 17, a dose calculation unit 18, a visualization unit 19, a dust behavior calculation unit 20, an estimation accuracy determination unit 21, and a re-estimation instruction unit 22.

[0029] The target area information acquisition unit 11 acquires information regarding the structure of the target area (hereinafter also referred to as "target area information"). The target area information is, for example, point cloud data or polygon data of the target area. The target area information is generated in advance by 3D scanning or photo scanning the target area. The target area information acquisition unit 11 acquires the target area information from the input device 2 or the communication device 3.

[0030] The model creation unit 12 uses the target area information acquired by the target area information acquisition unit 11 to create a 3D CG model (hereinafter also referred to as a "structural model") that represents the structures in the target area, as shown in Fig. 4. The model creation unit 12 divides the created structural model into a plurality of regions (also referred to as meshes, grids, or elements) to create a mesh model. In this case, the model creation unit 12 divides each surface (walls, floors, ground, ceiling, surfaces of indoor structures, etc.) and space within the building represented by the structural model into a plurality of regions to create the mesh model.

[0031] The radiation source vector creation unit 13 creates a radiation source vector ω indicating a radiation source assumed to exist in each of the multiple regions obtained by dividing the target area. Specifically, the radiation source vector creation unit 13 assumes that a radiation source exists in each mesh of the mesh model created by the model creation unit 12, and sets a radiation source in each mesh. Then, as shown on the left side of Fig. 2, the radiation source vector creation unit 13 creates a column vector as the radiation source vector ω, in which the element values ​​are the intensity of radiation emitted from the radiation source set in each mesh, and the number of elements is the total number m of the set radiation sources. In other words, the dimension of the radiation source vector ω is m.

[0032] The measurement information acquisition unit 14 acquires measurement information obtained by measuring radiation doses at multiple measurement points within the target area. The measurement information includes measurement points indicating the measurement positions of the radiation doses and measurement values. The measurement values ​​of the radiation doses are, for example, air dose rates. The measuring instruments for measuring the radiation doses include survey meters, spectrometers, gamma cameras, or the like. The measurement information acquisition unit 14 acquires the measurement information from the input device 2 or the communication device 3.

[0033] The dose vector creation unit 15 creates a dose vector y indicating measurement values ​​obtained by measuring radiation doses at a plurality of measurement points. Specifically, the dose vector creation unit 15 refers to the measurement points and measurement values ​​included in the measurement information acquired by the measurement information acquisition unit 14. As shown in Fig. 2, the dose vector creation unit 15 creates a column vector as the dose vector y, in which the element values ​​are the measurement values ​​at each measurement point and the number of elements is the number p of measurement points. That is, the dimension of the dose vector y is p.

[0034] The matrix calculation unit 16 calculates the matrix X. The matrix X indicates the degree to which radiation emitted from a radiation source indicated by the radiation source vector ω to a measurement point contributes to a measurement value. For this reason, the matrix X is also called a contribution rate matrix. The matrix X is a matrix (p×m matrix) corresponding to the number m of radiation sources indicated by the radiation source vector ω and the number p of measurement points. The matrix calculation unit 16 calculates radiation transport between the radiation source vector ω and the dose vector y to calculate the matrix X. Specifically, the matrix calculation unit 16 calculates how radiation (e.g., gamma rays or X-rays) emitted from one radiation source assigned to one element of the radiation source vector ω is transported to each measurement point assigned to the element of the dose vector y, respectively, and creates one column of the matrix X. Next, the matrix calculation unit 16 calculates how radiation emitted from the next radiation source assigned to the next element of the radiation source vector ω is transported to each measurement point assigned to the element of the dose vector y, and creates the next column of the matrix X. By performing such a calculation m times, which is the number of dimensions of the source vector ω, the matrix calculation unit 16 can create the matrix X.

[0035] An example of a radiation transport calculation code used by the matrix calculation unit 16 is PHITS (Particle and Heavy Ion Transport code System) developed by the Japan Atomic Energy Agency. PHITS is a Monte Carlo calculation code that simulates the behavior of various radiations using nuclear reaction models and nuclear data libraries.

[0036] The radiation source estimation unit 17 estimates the distribution of radiation sources in the target area. Specifically, the radiation source estimation unit 17 regards the product of the radiation source vector ω and the matrix X as a pseudo vector of the dose vector y, and estimates the distribution of radiation sources in the target area by performing an inverse analysis to restore the radiation source vector ω from the pseudo vector, with the radiation source vector ω as a sparse vector. The radiation source estimation unit 17 includes a pseudo vector setting unit 171 and an inverse analysis unit 172.

[0037] The pseudo vector setting unit 171 calculates a product Xω of the radiation source vector ω created by the radiation source vector creation unit 13 and the matrix X calculated by the matrix calculation unit 16. The pseudo vector setting unit 171 sets the calculated product Xω in the inverse analysis unit 172 as a pseudo vector that pseudo-indicates the dose vector y created by the dose vector creation unit 15.

[0038] The inverse analysis unit 172 performs an inverse analysis to restore the radiation source vector ω from the pseudo vector set by the pseudo vector setting unit 171. Specifically, the inverse analysis unit 172 assumes that the relationship shown in the following formula (1) holds between the radiation source vector ω, the matrix X, and the dose vector y, as shown in FIG.

[0039]

number

[0040] Furthermore, the inverse analysis unit 172 sets the radiation source vector ω to a sparse vector. The radiation source vector ω is set to have a sufficiently large number of elements whose element values ​​are zero. In this case, the number s of elements whose element values ​​are not zero in the radiation source vector ω is sufficiently small compared to the dimension m of the radiation source vector ω (s< <m)。

[0041] Then, the inverse analysis unit 172 performs inverse analysis using a least absolute shrinkage and selection operator (LASSO) that searches for a radiation source vector ω that minimizes an objective function D shown in the following formula (2). That is, the inverse analysis unit 172 performs inverse analysis using a LASSO that searches for a radiation source vector ω that minimizes an objective function D obtained by adding a regularization term including the L1 norm of the radiation source vector ω to an error term including the square of the L2 norm of the difference between the dose vector y and the pseudo vector.

[0042]

number

[0043] In the above formula (2), the first term on the right hand side is an error term, and the second term on the right hand side is a regularization term. λ is a hyperparameter that is set in advance. The larger λ is, the sparser the radiation source vector ω is. The distribution of radiation sources in the target area is often such that high-intensity radiation sources are sparsely present. In other words, the radiation dose is often zero or uniform in areas other than those where high-intensity radiation sources are present. This means that the distribution of radiation sources in the target area is characterized by a sparse distribution. Therefore, it is reasonable for the inverse analysis unit 172 to treat the radiation source vector ω as a sparse vector.

[0044] The inverse analysis unit 172 can group the radiation sources corresponding to the predetermined structures existing in the target area among the radiation sources indicated by the radiation source vector ω, and perform the above inverse analysis using Grouped LASSO. That is, the inverse analysis unit 172 regards the grouped radiation sources as one element of the radiation source vector ω, and performs a process of searching for the radiation source vector ω that minimizes the above objective function D, thereby recovering the radiation source vector ω.

[0045] For example, consider a pipe and a wall near the pipe as a predetermined structure existing in the target area. Assume that the pipe is represented by a plurality of meshes, and a radiation source is set in each of the plurality of meshes. The radiation source set in each of the plurality of meshes representing the pipe is assigned to each of the plurality of elements corresponding to the pipe in the radiation source vector ω. Similarly, for the wall near the pipe, the radiation source set in each of the plurality of meshes representing the wall is assigned to each of the plurality of elements corresponding to the wall in the radiation source vector ω. In this case, the inverse analysis unit 172 sets the plurality of elements corresponding to the pipe in the radiation source vector ω in one group, and sets the plurality of elements corresponding to the wall in the radiation source vector ω in one group. Then, the inverse analysis unit 172 performs the above inverse analysis using Grouped LASSO. Then, the inverse analysis unit 172 regards the plurality of elements corresponding to the pipe in the radiation source vector ω as one element, and regards the plurality of elements corresponding to the wall in the radiation source vector ω as one element, and performs a process of searching for a radiation source vector ω that minimizes the above objective function D. As a result, the inverse analysis unit 172 does not normally infer that a radiation source exists in both the element group corresponding to the piping and the element group corresponding to the wall, but can infer that a radiation source exists in either one of the element groups.

[0046] In a building having a complex structure, it is not easy to grasp the distribution of radiation sources due to the structure. For example, when measuring the radiation dose with a gamma camera, if there is a pipe on the front side of the measurement direction, it may be difficult to determine whether the radiation source is on the pipe or on the wall behind the pipe. The radiation source estimation unit 17 groups radiation sources corresponding to predetermined structures present in the target area and can perform the above-mentioned inverse analysis using Grouped LASSO. This allows the radiation source estimation unit 17 to accurately estimate the distribution of radiation sources in accordance with reality, taking into account the influence of complex structures present in the target area on the distribution of radiation sources. Therefore, the radiation source estimation device 1 can accurately estimate the distribution of radiation sources from a small number of measurement results even if the target area has a complex structure.

[0047] The dose calculation unit 18 calculates the distribution of radiation doses in the target area based on the distribution of radiation sources estimated by the radiation source estimation unit 17. The dose calculation unit 18 calculates how radiation emitted from the radiation sources estimated by the radiation source estimation unit 17 is transported in the target area. The dose calculation unit 18 calculates the distribution of radiation doses in the target area using a radiation transport calculation code such as PHITS, similar to the matrix calculation unit 16.

[0048] The visualization unit 19 creates a map that visualizes the distribution of radiation sources in the target area, as shown in Fig. 5. Specifically, the visualization unit 19 creates a VR (virtual reality) image or an AR (augmented reality) image by superimposing the distribution of radiation sources estimated by the radiation source estimation unit 17 on a CG model or a captured image of the target area. The visualization unit 19 creates the map using the created VR image or AR image. This VR image or AR image may be a still image or a moving image. This VR image or AR image may be a three-dimensional image or a two-dimensional image.

[0049] For example, the visualization unit 19 superimposes the distribution of radiation sources estimated by the radiation source estimation unit 17 on the structural model, which is a three-dimensional CG model created by the model creation unit 12, to create a VR image to which the estimated radiation sources are added. Alternatively, the visualization unit 19 superimposes the distribution of radiation sources estimated by the radiation source estimation unit 17 on a captured image of the target area acquired in real time by a mobile terminal or the like of a worker working in the target area, to create an AR image to which the estimated radiation sources are added. Then, the visualization unit 19 displays the created VR image or AR image on the display device 5 such as a liquid crystal display, or transmits it to an external device via the communication device 3 to display it on the external device. Note that the visualization unit 19 can superimpose not only the distribution of radiation sources estimated by the radiation source estimation unit 17, but also the radiation dose calculated by the dose calculation unit 18 on the CG model or captured image of the target area to create a VR image or an AR image.

[0050] The visualization unit 19 can create a VR image to make it appear as if a radiation source exists on the surface of a structure in a virtual space that reproduces the target area. The visualization unit 19 can create an AR image to make it appear as if a radiation source exists on the surface of a structure at a work site in the target area. This allows the radiation source estimation device 1 to notify a worker of, for example, where a high-intensity radiation source exists in the target area in a format that allows the worker to intuitively understand. Therefore, the radiation source estimation device 1 can effectively support decontamination work, etc., performed by workers at the work site.

[0051] The dust behavior calculation unit 20 calculates the behavior of dust floating or deposited in the target area depending on the wind conditions. There may be radiation sources that adhere to the surfaces of structures in the target area and remain there over time (hereinafter also referred to as "static radiation sources"), as well as radiation sources that adhere to dust floating or deposited in the target area and move over time (hereinafter also referred to as "dynamic radiation sources"). The radiation source estimation device 1 can estimate the distribution of dynamic radiation sources as well as static radiation sources in the target area by calculating the behavior of dust floating or deposited in the target area depending on the wind conditions. The distribution of dynamic radiation sources may be such that the position of the radiation source changes over time.

[0052] When estimating the distribution of radiation sources in the target area taking into account dynamic radiation sources, the radiation source vector creation unit 13 creates a radiation source vector ω on the assumption that the radiation source includes dust. That is, the radiation source vector creation unit 13 creates a radiation source vector ω by setting not only a radiation source in a mesh representing a structure in the target area, but also a radiation source in a mesh representing dust floating or deposited in the target area. The dimension of the radiation source vector ω including dust (radiation source vector ω considering dynamic radiation sources) is larger than that of the radiation source vector ω not including dust (radiation source vector ω of only static radiation sources). The matrix calculation unit 16 calculates a matrix X including a matrix indicating the degree to which radiation emitted from dust to a measurement point contributes to a measurement value. The number of elements in each row of the matrix X including dust is larger than that of the matrix X not including dust.

[0053] The radiation source estimation unit 17 estimates the distribution of radiation sources including dust at a predetermined time. The dust behavior calculation unit 20 calculates the position of dust after a predetermined time has elapsed from the predetermined time, using the dust distribution estimated by the radiation source estimation unit 17 at the predetermined time as an initial condition. The dust behavior calculation unit 20 calculates the position of dust after the predetermined time has elapsed by calculating the advection-diffusion phenomenon of dust using, for example, a CFD wind analysis calculation code. Then, the radiation source estimation unit 17 estimates the distribution of radiation sources after the predetermined time has elapsed from the calculation result of the dust behavior calculation unit 20. Furthermore, the dose calculation unit 18 can calculate the distribution of radiation doses from the estimated distribution of radiation sources after the predetermined time has elapsed.

[0054] This allows the radiation source estimation unit 17 to estimate the distribution of radiation sources in the target area by considering not only static radiation sources but also dynamic radiation sources. Therefore, the radiation source estimation unit 17 can estimate how the distribution of radiation sources in the target area changes over time, and can estimate the distribution of radiation sources more accurately. Therefore, the radiation source estimation device 1 can estimate the distribution of radiation sources with higher accuracy from a small number of measurement results.

[0055] The estimation accuracy determination unit 21 determines whether the estimation accuracy of the radiation source distribution by the radiation source estimation unit 17 (hereinafter also referred to as the "estimation accuracy of the radiation source estimation unit 17") is equal to or higher than a predetermined standard. For example, the estimation accuracy determination unit 21 determines whether the difference between the radiation dose near each radiation source estimated by the radiation source estimation unit 17 and the measurement value at each measurement point closest to each radiation source is within a predetermined allowable range. In this way, the estimation accuracy determination unit 21 determines whether the estimation accuracy of the radiation source estimation unit 17 is equal to or higher than a predetermined standard. The standard and allowable range of the estimation accuracy can be determined in advance by experiments, simulations, etc.

[0056] If the difference is within the allowable range, the estimation accuracy determination unit 21 determines that the estimation accuracy of the radiation source estimation unit 17 is equal to or higher than the standard. In this case, the estimation accuracy determination unit 21 determines that the estimation accuracy of the radiation source estimation unit 17 is sufficiently high, and determines that the estimation result of the radiation source estimation unit 17 is valid. If the difference is outside the allowable range, the estimation accuracy determination unit 21 determines that the estimation accuracy of the radiation source estimation unit 17 is below the standard. In this case, the estimation accuracy determination unit 21 determines that the estimation accuracy of the radiation source estimation unit 17 is insufficient, and notifies the re-estimation instruction unit 22 to that effect.

[0057] When the re-estimation instruction unit 22 determines that the estimation accuracy of the radiation source estimation unit 17 falls below the above-mentioned standard, it adds new radiation dose measurement values ​​at new measurement points and causes the radiation source estimation unit 17 to re-estimate the distribution of radiation sources in the target area.

[0058] Specifically, the re-estimation instruction unit 22 instructs the measurement information acquisition unit 14 to re-acquire measurement information so as to add new measurement values ​​of radiation doses at new measurement points. At this time, the re-estimation instruction unit 22 calculates a prior probability distribution P(ω I ) is set. Then, the re-estimation instruction unit 22 uses Bayesian estimation shown in the following formula (3) to identify new measured values ​​of radiation doses at new measurement points and instructs re-acquisition of measurement information. Note that the use of Bayesian estimation makes it possible to utilize prior measurement information and to re-acquire reliable measurement information, so that the radiation source estimation device 1 can estimate the distribution of radiation sources with higher estimation accuracy.

[0059]

number

[0060] On the left side of the above equation (3), P(ω I |ω) indicates the posterior distribution probability. On the right-hand side of the above equation (3), P(ω I ) denotes the prior probability distribution, and P(ω|ω I ) is a certain situation (ω I) indicates the probability distribution that the data (ω) is obtained, and P(ω) indicates the probability distribution that the data (ω) is obtained on average.

[0061] The re-estimation instruction unit 22 displays an instruction to re-acquire the measurement information on the display device 5, or transmits it to an external device via the communication device 3 to display it on the external device. When the measurement information to which a new radiation dose measurement value at a new measurement point is added becomes available from the input device 2 or the communication device 3, the measurement information acquisition unit 14 re-acquires the measurement information. The dose vector creation unit 15 recreates the dose vector y based on the re-acquired measurement information. One column of elements of the recreated dose vector y is added compared to before the re-creation. The matrix calculation unit 16 recreates the matrix X using the recreated dose vector y. One or more rows of elements of the recreated matrix X are added compared to before the re-creation. The addition of multiple rows means that multiple measurement points are added. The radiation source estimation unit 17 re-estimates the distribution of the radiation source in the target area based on the recreated matrix X and dose vector y. The radiation source estimation device 1 repeats the addition of the measured values ​​and the re-estimation by the radiation source estimation unit 17 as described above until the estimation accuracy of the radiation source estimation unit 17 reaches or exceeds a reference level.

[0062] As a result, even if the estimation accuracy of the radiation source estimation unit 17 is insufficient due to a small number of measurement points, the radiation source estimation device 1 can increase the number of measurement points and gradually improve the estimation accuracy. Therefore, the radiation source estimation device 1 can accurately estimate the distribution of radiation sources from a minimum number of measurement results.

[0063] FIG. 6 is a flowchart showing an outline of the processing performed by the radiation source estimation device 1.

[0064] The radiation source estimation device 1 acquires target area information, such as point cloud data of the target area (step S1), and creates a structural model of the target area using the acquired target area information (step S2).Then, the radiation source estimation device 1 creates a radiation source vector ω based on the created structural model (step S3).

[0065] The radiation source estimation device 1 also acquires measurement information including the measured values ​​and measurement points of radiation doses in the target area (step S4), and creates a dose vector y based on the acquired measurement information (step S5).The radiation source estimation device 1 then performs radiation transport calculation between the radiation source vector ω and the dose vector y (step S6), and creates a matrix X (step S7).

[0066] Then, the radiation source estimation device 1 regards the product of the radiation source vector ω and the matrix X as a pseudo vector of the dose vector y, and performs an inverse analysis to restore the radiation source vector ω from the pseudo vector, with the radiation source vector ω being a sparse vector. This allows the radiation source estimation device 1 to estimate the distribution of radiation sources in the target area (step S8).

[0067] Then, the radiation source estimation device 1 judges whether the estimation accuracy of the radiation source is equal to or higher than the standard (step S9). If the estimation accuracy of the radiation source is equal to or higher than the standard, the radiation source estimation device 1 judges that the estimated distribution of the radiation source is valid (step S9: Yes) and ends the process shown in FIG. 9. If the estimation accuracy of the radiation source is lower than the standard, the radiation source estimation device 1 judges that the estimated distribution of the radiation source is not valid (step S9: No). Then, the radiation source estimation device 1 instructs to reacquire measurement information to which new measured values ​​of radiation doses at new measurement points are added (step S10). Then, the radiation source estimation device 1 repeats the process from reacquiring the measurement information (step S4) to judging the estimation accuracy of the radiation source (step S9) until the estimation accuracy of the radiation source is equal to or higher than the standard.

[0068] [Verification of radiation source estimation device] FIG. 7 is a diagram for explaining prerequisites for verifying the validity of the radiation source estimation device 1. FIG. 8(a) is a diagram showing the distribution of radiation sources of the test data shown in FIG. 7. FIG. 8(b) is a diagram showing the estimation result of the distribution of radiation sources by the radiation source estimation device 1. FIG. 9(a) is a diagram showing the calculation result of the distribution of radiation doses based on the radiation sources of the test data shown in FIG. 8(a). FIG. 9(b) is a diagram showing the calculation result of the distribution of radiation doses based on the estimation result of the radiation source shown in FIG. 8(b). FIG. 10 is a graph comparing the radiation sources of the test data with the estimation result of the radiation source by the radiation source estimation device 1. FIG. 11 is a graph comparing the calculation result of the radiation doses based on the radiation sources of the test data with the calculation result of the radiation doses based on the estimation result of the radiation source by the radiation source estimation device.

[0069] In this embodiment, the validity of the radiation source estimation device 1 is verified by using a structural model of the target area shown in FIG. 4. FIG. 7 shows a top view of the inside of a building which is the target area shown in FIG. 4. As shown in FIG. 7, a total of 18 measurement points are provided in the building which is the target area, including a plurality of measurement points No. 1 to No. 9 at positions 100 cm and 200 cm above the building. The radiation sources of the test data for verification are provided on the surfaces of the walls A1 to A5 and the indoor structure B1. In this embodiment, the validity of the radiation source estimation device 1 is verified by verifying whether the estimation result of the radiation source by the radiation source estimation device 1 can reproduce the radiation source of the test data for verification.

[0070] The distribution of radiation sources in the test data shown in Fig. 8(a) represents the correct value of the estimation result of the radiation source shown in Fig. 8(b). Both Fig. 8(a) and Fig. 8(b) show that radiation sources are present in each of the walls A1 to A5 and the indoor structure B1. Comparing Fig. 8(b) with Fig. 8(a), it can be seen that the estimation result of the radiation source by the radiation source estimation device 1 can substantially reproduce the position of the radiation source and the radiation intensity of the test data.

[0071] Furthermore, FIG. 10 shows a graph comparing the radiation sources of the test data with the estimation result of the radiation source by the radiation source estimation device 1. The vertical axis of FIG. 10 indicates the intensity (relative value) of radiation emitted from each radiation source. The horizontal axis of FIG. 10 indicates the index (identification number) of the radiation source. The radiation sources with indexes=1, 3, 6, 11, 13, and 17 correspond to the radiation sources provided on the wall A3, the wall A2, the wall A5, the wall A4, the indoor structure B1, and the wall A1, which are the radiation sources of the test data shown in FIG. 7, respectively. In FIG. 10, square marks indicate the radiation sources of the test data, and triangular marks indicate the radiation sources estimated by the radiation source estimation device 1. Referring to FIG. 10, it can be seen that the estimation result of the radiation source by the radiation source estimation device 1 can substantially reproduce the position and radiation intensity of the radiation source of the test data.

[0072] Moreover, the radiation dose distribution based on the radiation source of the test data shown in Fig. 9(a) means the correct value of the calculation result of the radiation dose shown in Fig. 9(b). Comparing Fig. 9(b) with Fig. 9(a), it can be seen that the calculation result of the radiation dose distribution by the radiation source estimation device 1 can substantially reproduce the radiation dose distribution based on the radiation source of the test data.

[0073] Furthermore, Fig. 11 shows a graph comparing the radiation dose based on the radiation source of the test data with the calculation result of the radiation dose based on the estimation result of the radiation source by the radiation source estimation device 1. The vertical axis of Fig. 11 indicates the radiation dose at each evaluation point. The horizontal axis of Fig. 11 indicates the index (identification number) of the evaluation point of the radiation dose. In Fig. 11, square marks indicate the radiation dose based on the radiation source of the test data, and triangular marks indicate the radiation dose based on the estimation result by the radiation source estimation device 1. Referring to Fig. 11, it can be seen that the radiation dose calculated based on the estimation result of the radiation source by the radiation source estimation device 1 can approximately reproduce the radiation dose calculated based on the radiation source of the test data.

[0074] In this way, the radiation source estimation result by the radiation source estimation device 1 can approximately reproduce the radiation source of the test data. The radiation dose calculation result based on the radiation source estimation result by the radiation source estimation device 1 can approximately reproduce the radiation dose calculation result based on the radiation source of the test data. Therefore, it can be said that the radiation source estimation device 1 is valid.

[0075] As described above, the radiation source estimation device 1 of this embodiment is a radiation source estimation device that estimates the distribution of radiation sources in a target area from the measured values ​​of radiation doses at measurement points in the target area. The radiation source estimation device 1 includes a radiation source vector creation unit 13 that creates a radiation source vector ω indicating a radiation source assumed to exist in each of a plurality of regions obtained by dividing the target area, and a dose vector creation unit 15 that creates a dose vector y indicating a measured value at a measurement point. The radiation source estimation device 1 includes a matrix calculation unit 16 that calculates a matrix X corresponding to the number of radiation sources and the number of measurement points indicated by the radiation source vector ω, and a radiation source estimation unit 17 that estimates the distribution of radiation sources in the target area. The radiation source estimation unit 17 regards the product of the radiation source vector ω and the matrix X as a pseudo vector of the dose vector y, and performs an inverse analysis to restore the radiation source vector ω from the pseudo vector, with the radiation source vector ω as a sparse vector, thereby estimating the distribution of radiation sources.

[0076] With this configuration, the radiation source estimation device 1 of this embodiment can accurately estimate the distribution of radiation sources from a few radiation dose measurement results in a target area such as a contaminated area of ​​radioactive materials. As a result, the radiation source estimation device 1 can accurately estimate the distribution of radiation doses in the target area, and can accurately grasp the contamination status of radioactive materials. In particular, the radiation source estimation device 1 can accurately estimate the distribution of radiation sources from a few radiation dose measurement results even inside a building with a complex structure. As a result, the radiation source estimation device 1 can be effectively applied even in high-dose areas where it is difficult to obtain measurement results due to exposure of workers and measuring instruments, and can reduce the effort and labor of measurement and the amount of radiation exposure.

[0077] Furthermore, the radiation source estimation unit 17 performs inverse analysis using LASSO to search for a radiation source vector ω that minimizes an objective function D, which is obtained by adding a regularization term including the L1 norm of the radiation source vector ω to an error term including the square of the L2 norm of the difference between the dose vector y and the pseudo vector.

[0078] The distribution of radiation sources in the target area is often such that high-intensity radiation sources are sparsely present. In other words, the radiation dose is often zero or uniform in areas other than those where high-intensity radiation sources are present. This means that the distribution of radiation sources in the target area is characterized as a sparse distribution. By focusing on this characteristic, the radiation source estimation unit 17 performs inverse analysis using LASSO, which can significantly reduce the amount of calculation compared to a method of searching for a radiation source vector ω that minimizes an objective function including the L0 norm, and can efficiently restore the radiation source vector ω. Therefore, the radiation source estimation device 1 can accurately and efficiently estimate the distribution of radiation sources from a small number of radiation dose measurement results.

[0079] The radiation source estimation method of this embodiment can be realized by the radiation source estimation apparatus 1 performing the process shown in FIG.

[0080] That is, the radiation source estimation method of this embodiment is a method for estimating the distribution of radiation sources in a target area from the measured values ​​of radiation doses at measurement points in the target area. The radiation source estimation method includes creating a radiation source vector ω indicating a radiation source assumed to exist in each of a plurality of regions obtained by dividing the target area (step S3), and creating a dose vector y indicating the measured values ​​at the measurement points (step S5). The radiation source estimation method includes calculating a matrix X corresponding to the number of radiation sources and the number of measurement points indicated by the radiation source vector ω (step S7), and estimating the distribution of radiation sources in the target area (step S8). Estimating the distribution of radiation sources (step S8) involves regarding the product of the radiation source vector ω and the matrix X as a pseudo vector of the dose vector y, treating the radiation source vector ω as a sparse vector, and performing an inverse analysis to restore the radiation source vector ω from the pseudo vector, thereby estimating the distribution of radiation sources.

[0081] As a result, the radiation source estimation method of this embodiment can accurately estimate the distribution of radiation sources in a target area such as an area contaminated by radioactive materials from a small number of radiation dose measurement results.

[0082] Although the embodiments of the present invention have been described in detail above, the present invention is not limited to the above embodiments, and various modifications can be made without departing from the spirit of the present invention described in the claims. In the present invention, the configuration of one embodiment can be added to the configuration of another embodiment, the configuration of one embodiment can be replaced with the configuration of another embodiment, or a part of the configuration of one embodiment can be deleted. [Explanation of symbols]

[0083] 1...radiation source estimation device, 2...input device, 3...communication device, 4...storage device, 5...display device, 10...arithmetic processing device, 11...target area information acquisition unit, 12...model creation unit, 13...radiation source vector creation unit, 14...measurement information acquisition unit, 15...dose vector creation unit, 16...matrix calculation unit, 17...radiation source estimation unit, 171...pseudo vector setting unit, 172...inverse analysis unit, 18...dose calculation unit, 19...visualization unit, 20...dust behavior calculation unit, 21...estimation accuracy determination unit, 22...re-estimation instruction unit

Claims

1. A radiation source estimation device that estimates a distribution of radiation sources in a target area from measured values ​​of radiation doses at measurement points in the target area, comprising: a radiation source vector creation unit that creates a radiation source vector indicating a radiation source hypothesized to exist in each of a plurality of regions obtained by dividing the target area; a dose vector creation unit that creates a dose vector indicating the measurement values ​​at the measurement points; a matrix calculation unit that calculates a matrix corresponding to the number of the radiation sources indicated by the radiation source vector and the number of the measurement points; a radiation source estimation unit that estimates a distribution of the radiation sources in the target area, The radiation source estimation unit estimates the distribution of the radiation sources by regarding the product of the radiation source vector and the matrix as a pseudo vector of the dose vector, treating the radiation source vector as a sparse vector, and performing an inverse analysis to restore the radiation source vector from the pseudo vector. A radiation source estimation device comprising:

2. The radiation source estimation unit performs the inverse analysis using a LASSO that searches for the radiation source vector that minimizes an objective function obtained by adding a regularization term including the L1 norm of the radiation source vector to an error term including the square of the L2 norm of the difference between the dose vector and the pseudo vector.

2. The radiation source estimation apparatus according to claim 1 .

3. The radiation source estimation unit groups the radiation sources corresponding to predetermined structures present in the target area among the radiation sources indicated by the radiation source vectors, and performs the inverse analysis using Grouped LASSO.

3. The radiation source estimation apparatus according to claim 2 .

4. a visualization unit that creates a map that visualizes a distribution of the radiation sources in the target area; The visualization unit creates a VR image or an AR image by superimposing the distribution of the radiation sources estimated by the radiation source estimation unit on a CG model or a captured image of the target area, and creates the map using the created VR image or AR image.

2. The radiation source estimation apparatus according to claim 1 .

5. A dust behavior calculation unit that calculates the behavior of dust suspended or deposited in the target area due to wind conditions, the radiation source vector creation unit creates the radiation source vector on the assumption that the radiation source includes the dust; The matrix calculation unit calculates the matrices including a matrix indicating a degree to which radiation radiated from the dust to the measurement point contributes to the measurement value, The radiation source estimation unit estimates a distribution of the radiation source including the dust at a predetermined time, the dust behavior calculation unit calculates a position of the dust after a predetermined time has elapsed from the predetermined time using the estimated dust distribution at the predetermined time as an initial condition; The radiation source estimation unit estimates a distribution of the radiation source after the predetermined time has elapsed from a calculation result of the dust behavior calculation unit.

2. The radiation source estimation apparatus according to claim 1 .

6. an estimation accuracy determination unit that determines whether or not the estimation accuracy of the distribution of the radiation sources by the radiation source estimation unit is equal to or greater than a predetermined standard; and a re-estimation instruction unit that, when it is determined that the estimation accuracy falls below the standard, causes the radiation source estimation unit to re-estimate the distribution of the radiation source by adding new measurement values ​​at new measurement points.

2. The radiation source estimation apparatus according to claim 1 .

7. A radiation source estimation method for estimating a distribution of radiation sources in a target area from measured values ​​of radiation doses at measurement points in the target area, comprising: generating a source vector indicating a radiation source hypothesized to be present in each of a plurality of regions obtained by dividing the area of ​​interest; creating a dose vector indicative of the measurements at the measurement points; calculating a matrix corresponding to the number of the radiation sources indicated by the source vector and the number of the measurement points; and estimating a distribution of the radiation sources in the area of ​​interest, The estimation of the distribution of the radiation source includes estimating the distribution of the radiation source by regarding the product of the radiation source vector and the matrix as a pseudo vector of the dose vector, treating the radiation source vector as a sparse vector, and performing an inverse analysis to restore the radiation source vector from the pseudo vector. A radiation source estimation method comprising:

Images