Model search device, model search method, and program

Abstract

To perform highly accurate model retrieval.SOLUTION: A model search device that calculates a content rate of a database model by a query model, wherein the query model and the database model can be represented by a hierarchical structure including an assembly model, a subassembly, and a component, the model search device comprising: A model search apparatus comprising: a subassembly content ratio calculation unit configured to calculate a content ratio of a subassembly of a database model by a query model in each set direction; an assembly content ratio calculation unit configured to calculate a content ratio of the database model by the query model in the set direction based on the content ratio of the subassembly of the database model by the subassembly of the query model; and an inter-direction content ratio calculation unit configured to calculate the content ratio of the database model by the query model based on the content ratio of the database model by the query model in each set direction.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a model retrieval device, a model retrieval method, and a program. [Background technology]

[0002] CAD models, such as 3D CAD models, are used in various fields, including manufacturing. There is a method for searching for models similar to a target model from models recorded in a database. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-113025 [Non-patent literature]

[0004] [Non-Patent Document 1] Katayama, K. and Sato, W.: Subassembly Retrieval of 3D CAD Assembly Models with Different Layout of Components Based on Sinogram, IEICE Trans. Information and Systems, Vol.E102.D, No.4, pp.777-787 (online), DOI: 10.1587 / transinf.2018DAP0015 (2019). Summary of the Invention [Problem to be solved by the invention]

[0005] However, in the invention described in Patent Document 1, the position of the subassembly to be compared is taken into consideration by comparing the subassembly to be compared with subassemblies other than the subassembly to be compared at the same time. However, this method does not take the position into consideration sufficiently, and there is a possibility that the search accuracy will be low. An object of the present invention is to provide a model retrieval device, a model retrieval method, and a program with high retrieval accuracy. [Means for solving the problem]

[0006] One aspect of the present invention is a model retrieval device that calculates the inclusion rate of a database model in a query model or the inclusion rate of a query model in a database model, wherein the query model and the database model can be represented by a hierarchical structure consisting of assembly models, subassemblies, and parts, and the model retrieval device includes: a subassembly inclusion rate calculation unit that calculates the inclusion rate of a subassembly of the database model in a subassembly of the query model or the inclusion rate of a subassembly of the query model in a subassembly of the database model, in a direction that is set for each of the query model and the database model; an assembly inclusion rate calculation unit that calculates the inclusion rate of the database model in the query model or the inclusion rate of the query model in the database model, in the set direction, based on the inclusion rate of the subassembly of the database model in the subassembly of the query model or the inclusion rate of the subassembly of the database model in the subassembly of the database model; and an inter-directional inclusion rate calculation unit that calculates the inclusion rate of the database model in the query model or the inclusion rate of the query model in the database model, based on the inclusion rate of the database model in the query model or the inclusion rate of the query model in the database model for each of the set directions.

[0007] In the above aspect, the model retrieval device further includes: a component pair determination unit that determines pairs of components included in the database model and components included in the query model based on combinations of components included in the database model and components included in the query model for each of the set directions when the content rate of the database model in the query model for each of the set directions or the sum of the content rates of the query model in the database model is maximum; and an output unit that displays the database model and / or the query model on an external display device based on the determined pairs.

[0008] One aspect of the present invention is a model retrieval device that calculates the inclusion rate of a database model in a query model or the inclusion rate of a query model in a database model, wherein the query model and the database model can be represented by a hierarchical structure consisting of assembly models, subassemblies, and parts, and the model retrieval method includes: a subassembly inclusion rate calculation step that calculates the inclusion rate of a subassembly of the database model in a subassembly of the query model or the inclusion rate of a subassembly of the query model in a subassembly of the database model, in a direction that is set for each of the query model and the database model; an assembly inclusion rate calculation step that calculates the inclusion rate of the database model in the query model or the inclusion rate of the query model in the database model, in the set direction, based on the inclusion rate of the subassembly of the database model in the subassembly of the query model or the inclusion rate of the subassembly of the database model in the subassembly of the database model; and an inter-directional inclusion rate calculation step that calculates the inclusion rate of the database model in the query model or the inclusion rate of the query model in the database model, based on the inclusion rate of the database model in the query model or the inclusion rate of the query model in the database model for each of the set directions. [Effects of the Invention]

[0009] According to the present invention, a highly accurate model search can be performed. [Brief explanation of the drawings]

[0010] [Figure 1] FIG. 1 is a diagram illustrating a configuration of a model retrieval device according to a first embodiment. [Figure 2] 1 is an example of a query model and a database model. [Figure 3] FIG. 2 is a diagram illustrating a configuration of a model comparison unit according to the present embodiment. [Figure 4] 10 is a diagram showing an example of a method for determining an appropriate movement amount by a part approximation calculation unit. FIG. [Figure 5] 10 is a diagram showing an example of a method for calculating the content rate of subassemblies by a subassembly content rate calculation unit. FIG. [Figure 6] FIG. 10 is a diagram showing an example of a method for calculating an assembly content rate by an assembly content rate calculation unit. [Figure 7] 10 is a diagram showing an example of a method for calculating a content rate between a query model and a database model by an inter-direction content rate calculation unit. FIG. [Figure 8] 4 is a flowchart showing the operation of the model retrieval device according to the first embodiment. [Figure 9] 1 is a flowchart illustrating a method for calculating a content ratio between a query model and a database model. [Figure 10] FIG. 10 is a diagram illustrating an average value of estimation accuracy. [Figure 11] FIG. 10 is a diagram illustrating an average value of processing time. [Figure 12] FIG. 10 is a diagram showing details of processing time. [Figure 13] FIG. 10 is a diagram illustrating a configuration of a model retrieval device according to a second embodiment. [Figure 14] FIG. 10 is a diagram showing variables in the processing by the component pair determination unit 14. [Figure 15] FIG. 10 is a diagram showing a display example. [Figure 16] 10 is a flowchart showing the operation of the model retrieval device according to the second embodiment. [Figure 17]FIG. 10 is a diagram showing the accuracy rate in the experimental results of the second embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0011] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a diagram showing the configuration of a model retrieval device 1 according to the first embodiment. The model retrieval device 1 retrieves a query model m q The model retrieval device 1 retrieves the database model m d Among them, the query model m q Database model m that contains a high percentage of d Hereinafter, the query model m q is the database model m d The proportion of queries containing q Database model by m d It is defined as the content of

[0012] The model retrieval device 1 includes a query model acquisition unit 11, a model comparison unit 12, an output unit 13, and a storage unit 19. The query model acquisition unit 11 acquires a query model m q Get the query model m q is the CAD model to be searched.

[0013] The model comparison unit 12 compares the query model m q and the database model m d and calculate the comparison result. d is the query model m q This is the CAD model to be compared with the database model m d is stored in the storage unit 19. The comparison result is, for example, a query model m q Database model by m d The content rate of the query model m q Database model m with high content by d is.

[0014] The output unit 13 outputs the comparison result.

[0015] Query model m q and database model m d is an assembly model m. The assembly model m is decomposed into one or more subassemblies s. The subassemblies s are decomposed into one or more parts c. In other words, the query model m q and database model m d is composed of multiple components c. Query model m q and database model m d can be expressed as a hierarchical structure consisting of an assembly model m, a subassembly s, and a part c.

[0016] Also, the query model m q and database model m d A label is attached to part c in the table. The label may be, for example, the material, color, or manufacturer of part c. Information about the layout of labeled parts is called layout information. The subassembly to which a part belongs is determined, for example, by the part's shape and label. For example, parts with the same shape and label belong to one subassembly.

[0017] Figure 2 shows the query model m q and database model m d This is an example of a query model m q In this case, component c q belong to the green subassembly, yellow subassembly, etc. The database model m d Part C d The same is true for .

[0018] Below, the query model m q Database model by m d The procedure for calculating the content of will be explained. 3 is a diagram showing the configuration of the model comparison unit 12 according to this embodiment. The model comparison unit 12 includes a component similarity calculation unit 120, a subassembly content rate calculation unit 121, an assembly content rate calculation unit 122, an inter-direction content rate calculation unit 123, and a comparison result calculation unit 124.

[0019] The part approximation calculation unit 120 calculates the part approximation of the query model m q Part C q and the database model m d Part C d The degree of approximation between is calculated. The subassembly content rate calculation unit 121 calculates the subassembly content rate of the query model m q Subassemblies of q Database model by m d Subassemblies of d Calculate the content. The assembly content calculation unit 122 calculates the subassembly s q Subassemblies by d Based on the content of q Database model by m d Calculate the content. The inter-direction content rate calculation unit 123 calculates the query model m for each set direction v. q Database model by m d Based on the content of q Database model by m d Calculate the content. The comparison result calculation unit 124 calculates the database model m d The query model m calculated for each q Based on the content rate of the query model m q Database model m with high content by d Determine.

[0020] The inter-direction content rate calculation unit 123 calculates the inter-direction content rate of all the database models m d In the query model m qThe output unit 13 calculates the content ratio by the query model m calculated by the inter-direction content ratio calculation unit 123. q Each database model m d The content rate may be output.

[0021] The calculation method will be described in detail below. First, the method for calculating the degree of approximation between parts by the part approximation calculation unit 120 will be described. The part approximation calculation unit 120 first calculates a query model m q Direction of v q and the database model m d Direction of v d The part approximation calculation unit 120 sets the query model m q Parts included in C q and the database model m d Parts included in C d In calculating the degree of approximation between parts, the query model m q or the database model m d The degree of approximation between parts is calculated by translating the vector v q and direction v d The part approximation calculation unit 120 calculates the part c q and parts C d The degree of approximation is calculated for all combinations of

[0022] The method for calculating the degree of similarity between parts c is as follows: q and parts C d The feature values ​​are calculated in the set direction v. The feature values ​​are calculated using a method that uses the three-dimensional Radon transform. First, the model is rotated so that direction v is in the positive direction of the x-axis. Then, the volume of the area sliced ​​from part c by two planes perpendicular to the x-axis is calculated. Here, the positions of the two planes change at predetermined sampling intervals, scanning the section where part c exists. This allows part c q and parts C dThe volume calculated for each sampling interval is obtained using the formula: . The volume calculated for each sampling interval for part c is hereinafter referred to as the "partial volume." The cross-correlation of the partial volumes between specific parts is calculated, and the maximum value of the cross-correlation is calculated as the degree of approximation between the specific parts. For example, the partial volume calculated from a part is a one-dimensional array, and the one-dimensional array of partial volumes calculated for each part is shifted by one to calculate the cross-correlation. In other words, the correspondence between the partial volumes between specific parts is shifted and the correspondence that results in the highest degree of approximation is examined to calculate the degree of approximation between the specific parts.

[0023] Here, the calculation of the cross-correlation is performed using a discrete Fourier transform, which can simplify the calculation. The calculation of the discrete Fourier transform may be performed using a fast Fourier transform (FFT).

[0024] The part approximation calculation unit 120 calculates the part c d and parts C q By solving the assignment problem to maximize the sum of the proximity between d Part C q The maximum sum of the degrees of approximation calculated here is the maximum value of the sum of the degrees of approximation of the query model m in the set direction v and translation amount. q and the database model m d is the degree of approximation. The part approximation calculation unit 120 calculates the query model m q and the database model m d The part approximation calculation unit 120 calculates the approximation of the query model m q and the database model m d The translation amount that maximizes the degree of approximation is determined. Hereinafter, this translation amount is referred to as the appropriate translation amount.

[0025] 4 is a diagram showing an example of a method for determining an appropriate movement amount by the part similarity calculation unit 120. In the example shown in FIG. q The assembly model q1 and s q Consists of 2 assembly models q 1 is part C q 1 and c q Consists of 3 assembly models q 2, part C q 2 and c q 4 and c q 5. In the example shown in Figure 4, the database model m d The assembly model d 1 and s d Consists of 2 assembly models d 1 is part C d 1. Consists of an assembly model d 2, part C d 2 and c d It consists of 3. The part approximation calculation unit 120 calculates the part approximation using the query model m q Parts included in C q 1 to c q Five parts of 5 and database model m d Parts included in C d 1 to c d The part approximation calculation unit 120 calculates the degree of approximation between the part c and the three parts c. d and parts C q By solving the assignment problem that maximizes the degree of proximity between d Part C q In the example shown in FIG. 4, when the translation amount is L1, d 1 and c d 2 and c d 3 is c q 4 and c q 2 and c q 5, and the translation amount is L2, c d 1 and c d 2 and c d 3 is c q 3 and c q 1 and c q 4, and the translation amount is L3, c d 1 and c d 2 and c d 3 is cq 1 and c q 2 and c q 3, and the translation amount is L4. d 1 and c d 2 and c d 3 is c q 5 and c q 1 and c q It is assigned to 3.

[0026] The part approximation calculation unit 120 calculates the sum of the approximations in the allocation of part c for each translation amount. In the example shown in FIG. 4, when the translation amount is L1, the part approximation calculation unit 120 calculates the sum of the approximations in the allocation of part c for each translation amount. d 1 and c q Approximation of 4 and c d 2 and c q Approximation of 2 and c d 3 and c q The part approximation calculation unit 120 calculates the sum of the calculated approximations and the degree of approximation of L5. The same applies to the parallel movement amounts L2, L3, and L4. The part approximation calculation unit 120 determines the parallel movement amount when the sum of the calculated approximations is maximum as the appropriate movement amount.

[0027] Thereafter, the subassembly content rate calculation unit 121 calculates the content rate of the subassembly s of the query model when the parallel movement amount is the appropriate movement amount. q Subassemblies of database models by d The content of is calculated for all combinations of subassemblies s. The subassembly content rate calculation unit 121 calculates the content rate of the parts c for each combination of the subassemblies s. d and parts C q By solving the assignment problem that maximizes the degree of proximity between d Part C q Calculate the maximum sum of the degrees of approximation when the part c is assigned to d and parts C q The degree of approximation between the subassembly s and the query model is calculated by the part approximation calculation unit 120 in the process of determining the appropriate amount of movement. The maximum value of the sum of the degrees of approximation calculated here is the maximum value of the sum of the degrees of approximation between the subassembly s and the query model when the set direction and parallel movement amount are the appropriate amount of movement. qSubassemblies of database models by d The subassembly content rate calculation unit 121 may calculate the content rate of the subassemblies s depending on the combination of the subassemblies. d The number of parts in the subassembly s q The number of parts may be greater than the number of parts included in part c d The number of parts is c q In this case, the number of database models m that contain the subassembly may be larger than the number of d is excluded from the search.

[0028] This allows us to calculate the query model m q Subassemblies of q Database model by m d Subassemblies of d The content is calculated.

[0029] 5 is a diagram showing an example of a method for calculating the content rate of a subassembly by the subassembly content rate calculation unit 121. The subassembly content rate calculation unit 121 calculates the content rate of a subassembly s of a query model when the parallel movement amount is an appropriate movement amount. q Subassemblies of database models by d The calculation of the content rate of s is performed for all combinations of subassemblies s. In the example shown in FIG. 5, the subassembly content rate calculation unit 121 calculates the content rate of s q 1 by s d 1, s q 2 by s d 1, s q 1 by s d 2 and s q 2 by s d In the example shown in FIG. 5, the subassembly content rate calculation unit 121 calculates the content rate of part c d 1 of 1 q 1 and c q By solving the assignment problem for 3, s d 1 by s q Calculate the content of 1 and d 2 and c d 3c q 1 and cq By solving the assignment problem for 3, s d 2 by s q Calculate the content of 1 and d 1 of 1 q 2 and c q 4 and c q By solving the assignment problem for 5, s d 1 by s q Calculate the content of 2 and d 2 and c d 3c q 2 and c q 4 and c q By solving the assignment problem for 5, s d 2 by s q Calculate the content of 2. q by s d The content is calculated.

[0030] The assembly content calculation unit 122 calculates s q by s d Based on the content of q Database model by m d The assembly content rate calculation unit 122 calculates the content rate of the subassembly s q Subassemblies by d By solving the assignment problem that maximizes the content of m q by m d The maximum value of the content rate of the query model m in the set direction v is calculated. q Database model by m d is the content of

[0031] 6 is a diagram showing an example of a method for calculating the assembly content rate by the assembly content rate calculation unit 122. In the example shown in FIG. 6, the assembly content rate calculation unit 122 calculates the subassembly s d 1 and s d 2 by s q 1 and s q 2 and s q By solving the assignment problem that maximizes the content of 3, we can find the query model m in the given direction v.q Database model by m d In the example shown in FIG. 6, the assembly content rate calculation unit 122 calculates the content rate of s d 1 to s q Assign to 1, s d 2 to s q Assign to 2, s d 1 by s q The content of 1 and s d 2 by s q The sum of the content of 2 and the query model m in the set direction v q Database model by m d The content is calculated as the percentage of As a result, the assembly content rate calculation unit 122 calculates the query model m q Database model by m d Calculate the content.

[0032] Query model m q Direction of v q and database model m d Direction of v d The part approximation calculation unit 120, the subassembly content rate calculation unit 121, and the assembly content rate calculation unit 122 calculate the query model m q Database model by m d By calculating the content rate of each direction v, the query model m q Database model by m d The content is calculated.

[0033] The inter-direction content rate calculation unit 123 calculates the query model m for each direction v. q Database model by m d Based on the content of q Database model by m d The inter-direction content rate calculation unit 123 calculates the content rate of the direction v q The query model m q Direction by V d Database model m d By solving the assignment problem that maximizes the content of the query model m qDatabase model by m d Calculate the content.

[0034] FIG. 7 shows the query model m q Database model by m d 7 is a diagram illustrating an example of a method for calculating the inclusion rate of the direction v. In the example illustrated in FIG. q 1 in the direction v d Assign to 2, direction v q 2 in the direction v d Assign to 5, direction v q 3 in the direction v d Assign to 1, direction v q 4 in the direction v d Assign to 3, direction v q 1 and direction v d m in 2 q by m d The content of and direction v q 2 and direction v d m in 5 q by m d The content of and direction v q 3 and direction v d m in 1 q by m d The content of and direction v q 4 and direction v d m in 3 q by m d The sum of the content of m q by m d The final content of m q by m d The final content of the query model m q The database model m d is the content of

[0035] 8 is a flowchart showing the operation of the model retrieval device 1 according to the first embodiment. The query model acquisition unit 11 acquires a query model m q (Step S11). The model comparison unit 12 acquires the query model m q and the database model m d(Step S12). The output unit 13 compares the query model m q and the database model m d The comparison result is output (step S13).

[0036] Figure 9 shows the query model m q Database model by m d 1 is a flowchart showing a method for calculating the content rate of a query model m q Database model by m d The content rate of the query model m q Direction of v q and the query model m q Direction of v d The part approximation calculation unit 120 calculates the degree of approximation of the query model m for each translation amount (steps S101 to S108). q and the database model m d The part approximation calculation unit 120 calculates the degree of approximation between the query model m q and the database model m d The subassembly content rate calculation unit 121 determines the translation amount when the degree of approximation between the query model m and the query model m is maximum as the appropriate translation amount (step S105). q Subassemblies of q Database model by m d Subassemblies of d The assembly content rate calculation unit 122 calculates the content rate of the subassemblies s when the parallel movement amount is the appropriate movement amount (step S106). q by s d Based on the content of q Database model by m d The content rate of the initially set query model m is calculated (step S107). q Direction of v q and the query model m q Direction of v d Query model m in q Database model by md The model comparison unit 12 performs the operations from step S101 to step S108 on the query model m q Direction of v q and the query model m q Direction of v d By changing the query model m q Database model by m d By calculating the content of all directions v q and v d Query model m in q Database model by m d The content is calculated.

[0037] The inter-direction content rate calculation unit 123 calculates the inter-direction content rate in the direction v q and v d For each query model m q Database model by m d Based on the content of q Database model by m d The content rate of the query model m is calculated (step S108). q Database model by m d The content is calculated. The order of calculation is not limited to this. For example, the model comparison unit 12 first calculates the feature quantities of the part c in all directions v and translation amounts, and then compares the feature quantities of the query model m in all directions v and translation amounts. q Database model by m d Calculate the similarity of the subassembly s in all directions v, determine the appropriate movement amount, and q Subassemblies by d Calculate the content rate of the query model m in all directions v. q Database model by m d Calculate the content of and direction v q and v d For each query model m q Database model by m d Based on the content of q Database model by m dThe content may be calculated.

[0038] (Experimental results) The experimental results of the first embodiment will be described below. q Database model by m d The content rate of the query model m q Database model by m d We prepare five types of objects: Die, Clutch, Pump, Engine, and Bogie. For each object, we create a query model m q Database model by m d The content of was calculated. d As a result, 25 models were prepared for each type. The 25 models can be divided into groups of five, Type A to Type E, according to their type. The five models of Type A and Type B are models that are composed of one subassembly less than the query model. The five models of Type C and Type D are models that are composed of the same number of subassemblies as the query model, but one of the subassemblies is composed of fewer parts than the query model. The five models of Type E are models that are composed of one subassembly less than the query model, and one of the subassemblies is composed of fewer parts than the query model.

[0039] Among the five models of each type, one model is included in the query model. q 25 database models by m d The content rate of each was calculated, and the five models with the highest content rate were determined. The proportion of the five models included in the query model that were determined to have a high content rate was taken as the "estimated accuracy." If the content rate can be calculated with high accuracy, the five models included in the query model and the five models determined to have a high content rate will match. Query model m q and the database model m dThe estimation accuracy was calculated four times by randomly translating and rotating the direction v. The average value of the time required to calculate the estimation accuracy (processing time) was also calculated. q and v d The estimation accuracy was calculated by changing both to 10, 20, 30, and 40.

[0040] 10A and 10B are diagrams showing average values ​​of estimation accuracy. FIG. 10A shows the average value of estimation accuracy by the method of the first embodiment, and FIG. 10B shows the average value of estimation accuracy by the comparative method. In FIGS. 10A and 10B, the horizontal axis represents the direction v q and v d 10, the vertical axis represents the number of times ...

[0041] 11A and 11B are diagrams showing average values ​​of processing time. FIG. 11A shows the average value of processing time by the method of the first embodiment, and FIG. 11B shows the average value of processing time by the comparative method. In FIGS. 11A and 11B, the horizontal axis represents the direction v q and v d 11, the vertical axis represents the number of directions, and the vertical axis represents the processing time. As shown in Fig. 11, the processing time by the method of the first embodiment is shorter than the processing time by the comparative method, and the difference becomes more noticeable as the number of directions increases.

[0042] FIG. 12 shows the details of the processing time. “Projection” is a query model m q and database model m d The content rate is the time it takes to calculate the features for the components of the query model m q Database model by m d The comparison method is the method described in Non-Patent Document 1.

[0043] The method of the first embodiment can generate a query model m in a shorter time than the comparison method. q Database model by m dThis is because, while the feature amount calculated in the method of the first embodiment is a one-dimensional array of partial volumes, the feature amount calculated in the comparative method is two-dimensional data, and therefore, the method of the first embodiment can simplify the calculation of the degree of approximation of parts.

[0044] The model retrieval device 1 according to the second embodiment will be described below. Fig. 13 is a diagram showing the configuration of the model retrieval device 1 according to the second embodiment. In addition to the model retrieval device 1 according to the first embodiment, the model retrieval device 1 according to the second embodiment includes a component pair determination unit 14. In the model comparison retrieval device 1 according to the second embodiment, a query model acquisition unit 11 and a model comparison unit 12 that perform the same processing as the model comparison unit 1 according to the first embodiment will not be described.

[0045] The component pair determination unit 14 determines the query model m q Parts included in C q and the database model m d Parts included in C d The component pair determination unit 14 determines a component pair based on the calculation results of the subassembly content rate calculation unit 121, the assembly content rate calculation unit 122, and the inter-direction content rate calculation unit 123. In the calculation result by the inter-direction content rate calculation unit 123, the query model m q Direction of v q and the database model m d Direction of v d In addition, in the calculation result by the assembly content rate calculation unit 122, the combination of the query model m q Database model by m d Subassemblies when calculating the content of q and subassemblies d In other words, the combination of the determined query model m q Direction of v q and the database model m d Direction of v dEven in combination with subassemblies q and subassemblies d The combination has been decided.

[0046] In the calculation result by the subassembly content rate calculation unit 121, the subassembly s q Subassemblies by d When calculating the content of part c q and parts C d As described above, the combination of the determined query model m q Direction of v q and the database model m d Direction of v d Even in combination with subassemblies q and subassemblies d Since the combination of q Direction of v q and the database model m d Direction of v d In combination with part c q and parts C d The combination has been determined.

[0047] The component pair determination unit 14 determines the determined query model m q Direction of v q and the database model m d Direction of v d Part c in multiple combinations with q and parts C d Based on the combination of q Parts included in C q Among them, the database model m d Prescribed parts included in c d The part that is most frequently combined with the specified part c d The parts are determined to be pairs of parts.

[0048] Referring to Figure 7, when calculating the final content, q 1 and direction vd Combination with 2, direction v q 2 and direction v d Combination with 5, direction v q 3 and direction v d Combination with 1, direction v q 4 and direction v d 3 is determined. The component pair determination unit 14 determines the combination of the database model m d Prescribed parts included in c d For a query model m q Direction of v q 1 and database model m d Direction of v d In combination with 2, part c q Among them, the specified part c d and the query model m q Direction of v q 2 and the database model m d Direction of v d In combination with 5, part c q Among them, the specified part c d and the query model m q Direction of v q 3 and the database model m d Direction of v d In combination with 1, part c q Among them, the specified part c d and the query model m q Direction of v q 4 and the database model m d Direction of v d In combination with 3, part c q Among them, the specified part c d The part that is most frequently included among the four parts is the specified part c. d The component pair determination unit 14 determines the components that are paired with the database model m d All parts included in c d By performing the above calculations, the database model m d Each part included in C d and the query model m qParts included in C q Determine.

[0049] The processing by the component pair determination unit 14 is carried out by using a database model m d Parts included in C d the number of query models m q Parts included in C q The number of variables is the same as the product of the number of q Direction of v q and the database model m d Direction of v d In each combination with part c q and parts C d This is done by incrementing the value of the variable corresponding to the combination of parts c d The variable with the largest value and the corresponding part c q However, each part c d and the paired part c q is.

[0050] Here, the component pair determination unit 14 determines whether the component c q Among them, specified part c d The part that has the most combinations with the specified part c, but the number of combinations is less than a specified value, is called the specified part c. d In this case, it is not necessary to determine the part as a pair of the specified part c. d The pair of is not specifically determined.

[0051] The component pair determination unit 14 determines the query model m q Parts included in C q Among them, the database model m d Prescribed parts included in c d If there are two or more parts that are most frequently combined with the specified part c d It is not necessary to determine the parts as a pair.

[0052] The output unit 13 according to the second embodiment outputs a query model m based on the determination by the component pair determination unit 14. q and / or database model md The output unit 13 according to the second embodiment may or may not output the comparison result, similarly to the first embodiment.

[0053] Query model m q is output to an external display device, and d Parts that are paired with and parts that are not paired with are displayed differently. For example, part c d Parts paired with m are displayed with lower transparency than unpaired parts, making them more clearly visible. Figure 15 shows an example of a display. Figure 15(a) shows a query model m q , and Fig. 15(b) shows the database model m d FIG. 15(c) shows the query model m displayed by the output unit 13. q The query model m shown in Figure 15(c) q In the database model m d The parts paired with the part are displayed with a transparency of 0.8, and the database model m d Parts that are not paired with this part are displayed with a transparency of 0.2. For example, part c d The parts paired with may be displayed in color.

[0054] The query model m to be displayed q and the database model m d For example, the query model m q and the database model m d In the query model m, the paired parts are displayed in the same color. q and the database model m d In , paired parts are displayed with the same caption.

[0055] Query model m q Parts included in C q Among them, the database model md Prescribed parts included in c d If there are two or more parts with the most combinations with the query model m q The two or more components in the image may be displayed in a manner that indicates that the correspondence cannot be clearly determined, for example, by displaying the two or more components with a degree of transparency between the paired and unpaired components.

[0056] 16 is a flowchart showing the operation of the model retrieval device 1 according to the second embodiment. The query model acquisition unit 11 acquires a query model m q (Step S21). The model comparison unit 12 acquires the query model m q and the database model m d (Step S22). The component pair determination unit 14 compares the query model m q Parts included in C q and the database model m d Parts included in C d The output unit 13 determines a pair with the query model m based on the determination by the component pair determination unit 14 (step S23). q and / or database model m d (Step S24).

[0057] The experimental results of the second embodiment will be described below. q and the database model m d We verified whether the components contained in both the database model m can be accurately determined and displayed. d as the query model m q We prepared 100 models included in the database model m, and calculated the percentage of the 100 models that could correctly determine and display the parts. d Direction of v d The number of queries is set to 40, and the query model m q Direction of v q The number of questions was changed to 20, 40, 60, 80, and 100, and the percentage of correct answers was calculated.

[0058] 17 is a graph showing the accuracy rate in the experimental results of the second embodiment. In the graph shown in FIG. 17, the vertical axis represents the accuracy rate, and the horizontal axis represents the direction v q The correct answer rate is 0.9 or more, and the direction v q Increasing the number of q When the number of queries is 80, the accuracy rate is 1. As a result, in the second embodiment, the query model m q The database model m contained in d It can be seen that it is possible to determine accurately.

[0059] In the second embodiment, the query model m q In part C d The query model m is displayed so that the parts paired with it can be recognized. q In part C d The parts paired with are, in other words, the query model m q Therefore, in the second embodiment, the query model m q It is possible to visualize the part contained in the query model m q It is possible to show the part contained in

[0060] Also, the user can query the model m q The database model m contained in d Instead of being provided with a list of queries, the user is presented with a query model m q The database model m contained in d By displaying an image showing the database model m d Visually check the image displayed for each image and create the required database model m d In particular, the selection of the query model m q and database model m d When is large, the query model m q It is difficult to check the parts contained in a query model m q The database model m contained in dWhen an image showing the database model m is displayed, the contained parts can be easily identified. d This can reduce the burden of checking the

[0061] Other Embodiments One embodiment of the present invention has been described in detail above with reference to the drawings, but the specific configuration is not limited to that described above, and various design changes and the like are possible within the scope that does not deviate from the gist of the present invention.

[0062] In this embodiment, the model comparison unit 12 compares the query model m q Database model by m d However, the reverse may be true. In other words, the model comparison unit 12 calculates the content rate of the database model m d Query model m by q In this case, the subassembly content rate calculation unit 121 may calculate the content rate of the database model m d Subassemblies of d Query model m by q Subassemblies of q The assembly content rate calculation unit 122 calculates the content rate of the database model m d Query model m by q The content of is calculated and the database model m d 123 is the database model m d Query model m by q At this time, the component pair determination unit 14 calculates the content rate of the database model m d Parts included in C d Among them, the query model m q Prescribed parts included in c q The part that is most frequently combined with the specified part c q The parts are determined to be pairs of parts.

[0063] The components of the model search device 1 are realized by, for example, a hardware processor such as a CPU (Central Processing Unit) executing a program (software). Some or all of these components may be realized by hardware (including circuitry) such as an LSI (Large Scale Integration), an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or a GPU (Graphics Processing Unit), or may be realized by a combination of software and hardware. The program may be stored in advance in a storage device (a storage device having a non-transitory storage medium) such as an HDD (Hard Disk Drive) or flash memory, or may be stored in a removable storage medium (a non-transitory storage medium) such as a DVD or CD-ROM, and installed by inserting the storage medium into a drive device. [Explanation of symbols]

[0064] 1 Model search device, 11 Query model acquisition unit, 12 Model comparison unit, 13 Output unit, 19 Storage unit, 120 Part similarity calculation unit, 121 Subassembly content rate calculation unit, 122 Assembly content rate calculation unit, 123 Inter-direction content rate calculation unit, 124 Comparison result calculation unit

Claims

1. A model search device that calculates a content rate of a database model in a query model or a content rate of a query model in a database model, wherein the query model and the database model can be represented by a hierarchical structure consisting of an assembly model, a subassembly, and a part; a subassembly content rate calculation unit that calculates a content rate of a subassembly of the database model by a subassembly of the query model or a content rate of a subassembly of the query model by a subassembly of the database model in a direction set in each of the query model and the database model; an assembly content rate calculation unit that calculates a content rate of the database model by the query model or a content rate of the query model by the database model in the set direction based on a content rate of the subassembly of the database model by the subassembly of the query model or a content rate of the subassembly of the database model by the subassembly of the query model; an inter-directional inclusion rate calculation unit that calculates the inclusion rate of the database model in the query model or the inclusion rate of the query model in the database model based on the inclusion rate of the database model in the query model for each of the set directions or the inclusion rate of the query model in the database model; A model search device comprising:

2. a part approximation calculation unit that translates the query model and the database model by a predetermined amount in the set direction, and calculates the approximation between the query model and the database model; Furthermore, the subassembly content rate calculation unit calculates a content rate of the subassembly of the database model by the subassembly of the query model or a content rate of the subassembly of the database model by the subassembly of the query model, at the amount of translation when the degree of approximation is greatest. The model retrieval device according to claim 1 .

3. the component approximation calculation unit calculates the approximation between the query model and the database model by solving an assignment problem for maximizing the sum of the approximations between the components constituting the query model and the components constituting the database model; The model retrieval device according to claim 2 .

4. the component approximation calculation unit calculates feature amounts by calculating volumes of portions obtained by slicing the components constituting the query model and the components constituting the database model by two planes perpendicular to the set direction, and calculates a maximum value of cross-correlation of the feature amounts as approximation. The model retrieval device according to claim 3 .

5. the assembly content rate calculation unit calculates a maximum value of a sum calculated by solving an assignment problem for maximizing a sum of content rates of subassemblies of the database model by subassemblies of the query model or content rates of subassemblies of the query model by subassemblies of the database model as the content rate of the database model by the query model or the content rate of the query model by the database model in the set direction; The model retrieval device according to claim 1 .

6. the inter-direction inclusion rate calculation unit calculates, as the inclusion rate of the database model by the query model or the inclusion rate of the query model by the database model, a maximum value of a sum calculated by solving an assignment problem for maximizing the sum of the inclusion rates of the database model by the query model or the inclusion rates of the query model by the database model for each of the set directions; The model retrieval device according to claim 1 .

7. a component pair determination unit that determines a pair of a component included in the database model and a component included in the query model based on a combination of a component included in the database model and a component included in the query model for each of the set directions when the content rate of the database model in the query model for each of the set directions or the sum of the content rates of the query model in the database model is maximized; an output unit that displays the database model and / or the query model on an external display device based on the determined pairs; The model retrieval device according to claim 1 , further comprising:

8. the component pair determination unit determines, among the components included in the query model, a component that has been most frequently combined with a predetermined component included in the database model, or, among the components included in the database model, a component that has been most frequently combined with a predetermined component included in the query model, as a component that is paired with the predetermined component; The model retrieval device according to claim 7 .

9. the output unit displays, among the components included in the query model, components that are paired with components included in the database model with a lower transparency than components that are not paired. The model retrieval device according to claim 7 .

10. In the displayed database model and query model, the components are shown as paired components. The model retrieval device according to claim 7 .

11. A model search device that calculates a content rate of a database model in a query model or a content rate of a query model in a database model, wherein the query model and the database model can be represented by a hierarchical structure consisting of an assembly model, a subassembly, and a part; a subassembly content rate calculation step of calculating a content rate of a subassembly of the database model by a subassembly of the query model or a content rate of a subassembly of the query model by a subassembly of the database model in a direction respectively set in the query model and the database model; an assembly content rate calculation step of calculating a content rate of the database model by the query model or a content rate of the query model by the database model in the set direction based on a content rate of the subassembly of the database model by the subassembly of the query model or a content rate of the subassembly of the database model by the subassembly of the query model; an inter-directional inclusion rate calculation step of calculating an inclusion rate of the database model by the query model or an inclusion rate of the query model by the database model based on an inclusion rate of the database model by the query model for each of the set directions or an inclusion rate of the query model by the database model; A model search method having:

12. A program that causes a computer to execute the model retrieval method according to claim 11.

Images