Machine vision-based multi-span beam rapid modeling method and device

Abstract

The application provides a kind of machine vision-based multi-span beam rapid modeling method and equipment, comprising: model training module, based on the annotated sketch establishes multi-span beam structure information identification model;The multi-span beam structure sketch to be identified is input into the multi-span beam structure information identification model, and the multi-span beam structure modeling information file containing beam node information, beam unit information, beam node boundary condition information, beam unit load condition information and beam material attribute information is output.The application can solve the problem of tedious and time-consuming multi-span beam structure modeling process, realize the rapid extraction of multi-span beam structure modeling information by hand-drawing multi-span beam structure sketch, output standard modeling information file, which can be imported into structural mechanics solver and other professional computing software through interface, quickly establish multi-span beam structure calculation model, solve the problem of complex modeling process and long calculation time of traditional finite element software and other professional software, and improve the calculation and modeling efficiency.

Description

Technical Field

[0001] This invention relates to a method and device for rapid modeling of multi-span beams based on machine vision. Background Technology

[0002] The stress calculation of multi-span beam structures is applied in many engineering fields such as building structures, bridge engineering, and road engineering. It has a wide range of applications and a large calculation demand.

[0003] Existing calculations often require the use of specialized calculation software such as finite element software or structural mechanics solvers. Finite element software has a complex modeling process and requires high technical skills from operators; structural mechanics solvers require manual input of various parameters such as nodes, elements, loads, and boundary conditions, which is inefficient and time-consuming. Summary of the Invention

[0004] The purpose of this invention is to provide a method and device for rapid modeling of multi-span beams based on machine vision.

[0005] To address the above problems, this invention provides a rapid modeling method for multi-span beams using machine vision, comprising:

[0006] Step S1: The training information acquisition and processing module collects sample multi-span beam structure diagrams containing different working conditions, and collects beam node information, beam element information, beam node boundary condition information, beam element load condition information, and beam material property information. Each type of information is marked at the corresponding position on the corresponding multi-span beam structure diagram to obtain an annotated diagram.

[0007] Step S2: The model training module, based on the beam node information, beam element information, beam node boundary condition information, beam element load condition information, and beam material property information marked in the annotation diagram, uses artificial intelligence methods to establish a multi-span beam structure information recognition model.

[0008] Step S3: Input the simplified diagram of the multi-span beam structure to be identified into the multi-span beam structure information recognition model. The multi-span beam structure information recognition model identifies the beam node information, beam element information, beam node boundary condition information, beam element load condition information, and beam material property information in the simplified diagram of the multi-span beam structure to be identified, and outputs a multi-span beam structure modeling information file containing beam node information, beam element information, beam node boundary condition information, beam element load condition information, and beam material property information.

[0009] Furthermore, in the above method, in step S1, the node information includes: the number of nodes and the node number, wherein the number of nodes is n, and the nodes are numbered and labeled from left to right starting from the left side of the simplified diagram of the multi-span beam structure of the sample; wherein, a beam is called a beam unit, the beam units are connected end to end, and the two ends of each beam unit are beam nodes.

[0010] Furthermore, in the above method, in step S1, the beam element information includes: number of elements, element number, and element length; wherein, each beam element corresponds to an element length data L1 to L2. n-1 The unit length is mm; the number of units is the number of nodes - 1, and the units are numbered and labeled from left to right, starting from the left side of the simplified diagram of the multi-span beam structure in the sample. There is one beam unit between two beam nodes.

[0011] Furthermore, in the above method, in step S1, the beam node boundary condition information includes: the support type under beam node 1 and beam node n, wherein the support type includes three types: hinged support JZ, fixed support GZ, and directional support DX; the hinged support constrains the horizontal and vertical displacement of the beam node on it; the fixed support constrains the horizontal displacement, vertical displacement, and rotation angle of the beam node on it; the directional support constrains the horizontal displacement and rotation angle of the node; in the sample multi-span beam structure diagram, beam nodes other than the first and last are assumed to be hinged supports.

[0012] Furthermore, in the above method, in step S1, the load condition information of the beam unit includes: load quantity, load number, load type, load value, and load location; wherein, the load quantity is consistent with the number of beam units, which is n-1, and the loads are numbered and labeled from left to right starting from the left side of the sample multi-span beam structure diagram; the load type, load value, and load location are labeled.

[0013] Furthermore, in the above method, the load type includes: concentrated force Q. F Concentrated bending moment Q M He Junbuli Q JB There are 3 types in total;

[0014] Load value Q (unit: concentrated force) F The load value is in kN, and the concentrated bending moment Q M The load value is in kN·m, and the uniformly distributed force Q JB The load value is in kN / m; each beam element corresponds to one load value.

[0015] Concentration Q F and concentrated bending moment Q M The load location is: concentrated force Q F and concentrated bending moment Q M The ratio S of the distance from the load application point to the left end node of the element to the element length; the uniformly distributed force Q. JB The load locations are: S1, the ratio of the distance from the starting point of the uniformly distributed force to the beam node at the left end of the element to the element length, and S2, the ratio of the distance from the ending point of the uniformly distributed force to the beam node at the right end of the element to the element length; where, the concentrated force Q F and concentrated bending moment QM The load distance is the concentrated force Q F and concentrated bending moment Q M The distance from the load application point to the left end node of the element; the load distance of a uniformly distributed force is the distance from the starting point of the uniformly distributed force to the beam node at the left end of the element, or the distance from the ending point of the uniformly distributed force to the beam node at the right end of the element.

[0016] When there is no load on a beam element, the load type is selected as concentrated force Q. F The load value Q is 0, and the load location S is 0.

[0017] Furthermore, in the above method, in step S1, the material property information includes: beam elastic modulus E, in MPa, and beam section modulus I, in mm4.

[0018] Furthermore, in the above method, in step S3, the simplified diagram of the multi-span beam structure to be identified includes: material property information, support type, length of each beam element and load distance; and concentrated force Q among the load types. F Concentrated bending moment Q M and uniformly distributed force Q JB Various shapes of load arrows, with the load value marked on the load arrow.

[0019] According to another aspect of the present invention, a computer-readable storage medium is also provided, having stored thereon computer-executable instructions, wherein when executed by a processor, the computer-executable instructions cause the processor to perform the method described in any of the preceding claims.

[0020] According to another aspect of the present invention, a calculator device is also provided, comprising:

[0021] Processor; and

[0022] A memory configured to store computer-executable instructions, which, when executed, cause the processor to perform the method described in any of the preceding descriptions.

[0023] Compared with existing technologies, this invention, through a training information acquisition and processing module, collects simplified diagrams of multi-span beam structures under different working conditions, and collects beam node information, beam element information, beam node boundary condition information, beam element load condition information, and beam material property information. Each type of information is then labeled at its corresponding position on the simplified diagram to obtain a labeled diagram. The model training module, based on the labeled beam node information, beam element information, beam node boundary condition information, beam element load condition information, and beam material property information in the labeled diagram, uses artificial intelligence methods to establish a multi-span beam structure information recognition model. The simplified diagram of the multi-span beam structure to be recognized is input into the multi-span beam structure information recognition model, and the multi-span beam structure information recognition model identifies the beam structure to be recognized. This system identifies beam node information, beam element information, beam node boundary condition information, beam element load condition information, and beam material property information from simplified diagrams of multi-span beam structures. It then outputs a multi-span beam structure modeling information file containing these information. This solves the problems of tedious and time-consuming multi-span beam structure modeling processes, enabling rapid extraction of multi-span beam structure modeling information from hand-drawn simplified diagrams and outputting a standard modeling information file. This file can be imported into professional calculation software such as structural mechanics solvers via an interface to quickly establish a multi-span beam structure calculation model. This addresses the issues of complex and time-consuming modeling processes in traditional finite element software and other professional software, improving calculation and modeling efficiency. Attached Figure Description

[0024] Figure 1 This is a flowchart of a machine vision-based rapid modeling method for multi-span beams according to an embodiment of the present invention;

[0025] Figure 2 This is a schematic diagram of a hinge support according to an embodiment of the present invention;

[0026] Figure 3 This is a schematic diagram of a fixed support according to an embodiment of the present invention;

[0027] Figure 4 This is a schematic diagram of an embodiment of the directional support of the present invention;

[0028] Figure 5 This is a schematic diagram of a simplified multi-span beam structure to be identified according to an embodiment of the present invention;

[0029] Figure 6 It is identified and labeled. Figure 5 A schematic diagram of the node and unit numbers in the diagram;

[0030] Figure 7 It is identified and labeled. Figure 5 Schematic diagram of medium load. Detailed Implementation

[0031] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0032] like Figures 1 to 7 As shown, this invention provides a rapid modeling method for multi-span beams based on machine vision, comprising:

[0033] Step S1: The training information acquisition and processing module collects sample multi-span beam structure diagrams containing different working conditions, and collects beam node information, beam element information, beam node boundary condition information, beam element load condition information, and beam material property information. Each type of information is marked at the corresponding position on the corresponding multi-span beam structure diagram to obtain an annotated diagram.

[0034] Here, the sample multi-span beam structure diagram can be a hand-drawn multi-span beam structure diagram; the training information acquisition and processing module is mainly responsible for the acquisition and processing of training data, collecting multi-span beam structure diagrams containing multi-span beam structures under different working conditions, and collecting beam node information, beam element information, beam node boundary condition information, beam element load condition information, and beam material property information, and marking each type of information at the corresponding position on the corresponding multi-span beam structure diagram;

[0035] Step S2: The model training module, based on the beam node information, beam element information, beam node boundary condition information, beam element load condition information, and beam material property information marked in the annotation diagram, uses artificial intelligence methods to establish a multi-span beam structure information recognition model.

[0036] Here, the model form of the multi-span beam structure information recognition model is not limited. CNN, RNN and ResNet models can all be used to establish the multi-span beam structure information recognition model. To ensure the training effect, the training set needs to have more than 1,000 sets of data.

[0037] The training process of the model can use algorithms such as stochastic gradient descent and adaptive gradient algorithm to train the model until the loss function is less than a preset value, at which point the training ends.

[0038] Step S3: Input the simplified diagram of the multi-span beam structure to be identified into the multi-span beam structure information recognition model. The multi-span beam structure information recognition model identifies the beam node information, beam element information, beam node boundary condition information, beam element load condition information, and beam material property information in the simplified diagram of the multi-span beam structure to be identified, and outputs a multi-span beam structure modeling information file containing beam node information, beam element information, beam node boundary condition information, beam element load condition information, and beam material property information.

[0039] Here, as Figure 5 As shown, the input to the multi-span beam structure information recognition model is a hand-drawn structural sketch of various multi-span beams to be identified; the structural sketches of the multi-span beams to be identified are labeled with: material property information, support type, and the length of each beam element, such as... Figure 5 The numbers 4000, 4000, 5000, and 3000 are used to indicate the type of load: concentrated force Q. F Concentrated bending moment Q M Uniformly distributed force Q JB Arrows of various shapes, with the load value marked on the arrow, such as... Figure 5 The 10kN / m, 15kN / m and kN / m mentioned above, the load distance is as follows Figure 5 The numbers 500, 800, 300, and 600 are included.

[0040] like Figure 6 and 7 The model output is a multi-span beam structure modeling information file containing beam node information, beam element information, beam node boundary condition information, beam element load condition information, and beam material property information. Figure 6 Five nodes and four units were identified. Figure 7 The location of various loads was calculated and marked.

[0041] It is important to note that the units of the beam node information, beam element information, beam node boundary condition information, beam element load condition information, and beam material property information input in actual use should be consistent with the data units used during model training.

[0042] This invention solves the problem of cumbersome and time-consuming modeling process for multi-span beam structures. It enables the rapid extraction of modeling information for multi-span beam structures by simply drawing a simplified diagram of the structure by hand, and outputs a standard modeling information file. This file can be imported into professional calculation software such as structural mechanics solvers via an interface to quickly establish a calculation model of the multi-span beam structure. This solves the problems of complex modeling process and long calculation time in traditional professional software such as finite element software, and improves calculation and modeling efficiency.

[0043] In one embodiment of the machine vision-based rapid modeling method for multi-span beams of the present invention, step S1 includes the node information as follows: the number of nodes and the node number, wherein the number of nodes is n, and the nodes are numbered and labeled from left to right, starting from the left side of the simplified diagram of the sample multi-span beam structure. The node numbers are, for example, JD1 to JD1. n In this system, a single beam is called a beam element, and beam elements are connected end to end. The two ends of each beam element are beam nodes.

[0044] In one embodiment of the machine vision-based rapid modeling method for multi-span beams of the present invention, in step S1, the beam element information includes: number of elements, element number, and element length; wherein, each beam element corresponds to an element length data L1 to L2. n-1 The unit of element length is mm; the number of elements is the number of nodes - 1 (i.e., n-1), and the elements are numbered from left to right, starting from the left side of the simplified diagram of the multi-span beam structure in the sample. Two beam nodes constitute one beam element, and the element names can be DY1 to DY1. n-1 .

[0045] In one embodiment of the machine vision-based rapid modeling method for multi-span beams of the present invention, in step S1, the beam node boundary condition information includes: the support (support under the node) types under beam node 1 and beam node n, wherein the support types include: such as Figure 2 The hinge support JZ shown is as follows: Figure 3 The fixed support GZ shown and as follows Figure 5 There are three types of directional supports DX shown; the hinged supports constrain the horizontal and vertical displacements of the beam nodes on them; the fixed supports constrain the horizontal, vertical, and rotation angles of the beam nodes on them; the directional supports constrain the horizontal displacement and rotation angle of the nodes; in the sample multi-span beam structure diagram, beam nodes other than the first and last ones are assumed to be hinged supports.

[0046] In one embodiment of the machine vision-based rapid modeling method for multi-span beams of the present invention, in step S1, the beam element load condition information includes: load quantity, load number, load type, load value, and load location; wherein, the load quantity is consistent with the number of beam elements, which is n-1, and the loads are numbered and labeled from left to right, starting from the left side of the sample multi-span beam structural diagram, with the load number names being HZ1 to HZ respectively. n-1 Label the load type, load value, and load location;

[0047] Load types include: concentrated force Q F Concentrated bending moment Q M Uniformly distributed force Q JB There are 3 types in total;

[0048] Load value Q (unit: concentrated force) F The load value is in kN, and the concentrated bending moment Q M The load value is in kN·m, and the uniformly distributed force Q JB The load value is in kN / m; each element corresponds to one load value.

[0049] Concentration Q F and concentrated bending moment Q M The load location is: concentrated force Q F and concentrated bending moment Q MThe ratio S of the distance from the load application point to the left end node of the element to the element length; the uniformly distributed force Q. JB The load locations are: S1, the ratio of the distance from the starting point of the uniformly distributed force to the beam node at the left end of the element to the element length, and S2, the ratio of the distance from the ending point of the uniformly distributed force to the beam node at the right end of the element to the element length; where, the concentrated force Q F and concentrated bending moment Q M The load distance is the concentrated force Q F and concentrated bending moment Q M The distance from the load application point to the left end node of the element; the load distance of a uniformly distributed force is the distance from the starting point of the uniformly distributed force to the beam node at the left end of the element, or the distance from the ending point of the uniformly distributed force to the beam node at the right end of the element.

[0050] When there is no load on a certain element, the load type is selected as concentrated force Q. F The load value Q is 0, and the load location S is 0.

[0051] In one embodiment of the machine vision-based rapid modeling method for multi-span beams of the present invention, in step S1, the material property information includes: beam elastic modulus E, in MPa, and beam section modulus I, in mm4.

[0052] According to another aspect of the present invention, a computer-readable storage medium is also provided, having stored thereon computer-executable instructions, wherein when executed by a processor, the computer-executable instructions cause the processor to perform the method described in any of the preceding claims.

[0053] According to another aspect of the present invention, a calculator device is also provided, comprising:

[0054] Processor; and

[0055] A memory configured to store computer-executable instructions, which, when executed, cause the processor to perform the method described in any of the preceding descriptions.

[0056] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0057] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0058] Obviously, those skilled in the art can make various modifications and variations to the invention without departing from the spirit and scope of the invention. Therefore, if these modifications and variations fall within the scope of the claims of the invention and their equivalents, the invention is also intended to include these modifications and variations.

Claims

1. A machine vision-based multi-span beam rapid modeling method, characterized in that, The method comprises the following steps: Step S1, training information collection and processing module, collecting sample multi-span beam structure diagrams containing different working conditions, and collecting beam node information, beam element information, beam node boundary condition information, beam element load condition information and beam material attribute information, and marking each kind of information on the corresponding position of the sample multi-span beam structure diagram to obtain a marked diagram; Step S2, model training module, based on the beam node information, beam element information, beam node boundary condition information, beam element load condition information and beam material attribute information marked in the marked diagram, a multi-span beam structure information recognition model is established by using an artificial intelligence method; Step S3, inputting the multi-span beam structure diagram to be identified into the multi-span beam structure information recognition model, the multi-span beam structure information recognition model identifies the beam node information, beam element information, beam node boundary condition information, beam element load condition information and beam material attribute information in the multi-span beam structure diagram to be identified, and outputs a multi-span beam structure modeling information file containing the beam node information, beam element information, beam node boundary condition information, beam element load condition information and beam material attribute information; The beam element load condition information comprises: load quantity, load number, load type, load value and load position; wherein the load quantity is consistent with the beam element quantity, which is n-1, and the load is numbered and marked from left to right starting from the left side of the sample multi-span beam structure diagram; the load type, load value and load position are marked; Load types include: concentrated force Q F , concentrated bending moment Q M and uniform force Q JB , three in total; Load value Q unit: concentrated force Q F Load value unit: kN, concentrated bending moment Q M Load value unit: kN·m, uniform force Q JB Load value unit: kN / m; each beam element corresponds to a load value; Concentration Q F and concentrated bending moment Q M The load location is: concentrated force Q F and concentrated bending moment Q M The ratio S of the distance from the load application point to the left end node of the element to the element length; the uniformly distributed force Q. JB The load locations are: S1, the ratio of the distance from the starting point of the uniformly distributed force to the beam node at the left end of the element to the element length, and S2, the ratio of the distance from the ending point of the uniformly distributed force to the beam node at the right end of the element to the element length; where, the concentrated force Q F and concentrated bending moment Q M The load distance is the concentrated force Q F and concentrated bending moment Q M The distance from the load application point to the left end node of the element; the load distance of a uniformly distributed force is the distance from the starting point of the uniformly distributed force to the beam node at the left end of the element, or the distance from the ending point of the uniformly distributed force to the beam node at the right end of the element. When there is no load on a beam element, the load type is selected as concentrated force Q F , the load value Q is 0, and the load position S is 0.

2. The machine vision based multi-span beam rapid modeling method of claim 1, wherein, In step S1, the node information comprises: node quantity and node number, wherein the node quantity is n, and the nodes are numbered and marked from left to right starting from the left side of the sample multi-span beam structure diagram; wherein one beam is called one beam element, and the beam elements are connected end to end, and the two ends of each beam element are beam nodes.

3. The machine vision based multi-span beam rapid modeling method of claim 2, wherein, In step S1, the beam unit information beam includes: unit quantity, unit number and unit length; wherein each beam unit corresponds to a unit length data L1~L n-1 , the unit length unit is mm; the unit quantity is node quantity-1, and the units are numbered and marked from left to right starting from the left side of the sample multi-span beam structure diagram, and one beam unit is between two beam nodes.

4. The machine vision based multi-span beam rapid modeling method of claim 2, wherein, In step S1, the beam node boundary condition information comprises: support types under beam node 1 and beam node n, wherein the support types comprise: hinge support JZ, fixed support GZ and directional support DX, a total of 3 kinds; the hinge support restricts the horizontal displacement and vertical displacement of the beam node thereon; the fixed support restricts the horizontal displacement, vertical displacement and rotation angle of the beam node thereon; the directional support restricts the horizontal displacement and rotation angle of the node; the beam nodes other than the first and last nodes in the sample multi-span beam structure diagram are defaulted as hinge supports.

5. The machine vision based multi-span beam rapid modeling method of claim 1, wherein, In step S1, the material attribute information comprises: beam elastic modulus E, unit: MPa, beam sectional resistance moment I.

6. The machine vision based multi-span beam rapid modeling method of claim 1, wherein, In step S3, the multi-span beam structure sketch to be identified is marked with: material attribute information, support type, length of each beam unit, and load distance; various shapes of load arrows of the types of concentrated force Q F , concentrated bending moment Q M , and uniform force Q JB , with load values marked on the load arrows.

7. A computer-readable storage medium having stored thereon computer- executable instructions, wherein, The computer executable instructions are executed by the processor to make the processor execute the method of any one of claims 1 to 6.

8. A computing device, wherein, The method comprises the following steps: A processor; And A memory arranged to store computer executable instructions which, when executed, cause the processor to execute the method of any one of claims 1 to 6.

Images