Three-dimensional model generation system

Abstract

The goal is to easily generate a three-dimensional model of an object from a paper drawing representing that object. [Solution] A system for generating a three-dimensional model of an object to be modeled, The imaging means 31 of the three-dimensional model generation system 30 captures a paper drawing 20 in which multiple figures 20a representing the object 19, consisting of the object to be modeled 10 or its constituent elements 11 to 14, are viewed from different directions, thereby obtaining graphic image data 39 of the object 19. The graphic image data 39 is input to the generation unit 40. The generation unit 40 is composed of, for example, an AI system that stores a trained model 41, and the trained model 41 generates a three-dimensional model 49 of the object 19 from the graphic image data 39.

Description

Technical Field

[0001] The present invention relates to a system for generating a three-dimensional model of industrial products such as products, parts, and materials in various industrial fields, and particularly relates to a system for generating a three-dimensional model from paper drawings.

Background Art

[0002] Conventionally, drawings of various industrial products are often stored and managed as paper drawings by directly drawing on drawing paper or by drawing on a PC (computer) and printing out.

[0003] Patent Document 1 discloses creating a two-dimensional plan view from a three-dimensional design drawing created by three-dimensional CAD.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] A three-dimensional model created by three-dimensional CAD can easily and clearly grasp the three-dimensional shape of an object such as an industrial product. On the other hand, if the original drawing is a paper drawing, it is not easy to perform three-dimensional modeling. In view of such circumstances, an object of the present invention is to provide a system that can easily generate a three-dimensional model of a modeling object from a paper drawing representing the modeling object such as an industrial product.

Means for Solving the Problems

[0006] To solve the above problems, the present invention is a system for generating a three-dimensional model of a modeling object, An imaging means for obtaining graphic image data of the object by capturing a paper drawing in which a plurality of figures of the object to be modeled or its components are viewed from different directions, A generation unit that generates a three-dimensional model of the object from the aforementioned graphic image data, It is characterized by having the following features.

[0007] Preferably, the generation unit includes a trained model generated by machine learning, which generates a three-dimensional model of the training object from training geometric image data of the training object.

[0008] Preferably, the generating unit is A template storage unit that stores a three-dimensional template model in which the variable elements of the three-dimensional shape of the object to be modeled are represented by indeterminate parameters, An extraction unit that extracts definitive information of the variable element from the graphic image data, A fitting unit that generates the three-dimensional model from the three-dimensional template model by applying the confirmed values ​​to the indeterminate parameters, Includes.

[0009] Preferably, the object to be modeled is a building structure or building materials, or a civil engineering structure or civil engineering materials. [Effects of the Invention]

[0010] According to the present invention, a three-dimensional model of the object to be modeled can be easily generated from a paper drawing representing the object. [Brief explanation of the drawing]

[0011] [Figure 1] Figure 1 is a schematic explanatory diagram illustrating a three-dimensional model generation method using a three-dimensional model generation system according to the first embodiment of the present invention. [Figure 2] Figure 2 is a schematic explanatory diagram illustrating a three-dimensional model generation method using a three-dimensional model generation system according to a second embodiment of the present invention. [Modes for carrying out the invention]

[0012] Embodiments of the present invention will be described below with reference to the drawings. <First Embodiment (Figure 1)> As shown in Figure 1, the three-dimensional model generation system 30 generates a digital three-dimensional model 49 from one or more paper drawings 20. The paper drawing 20 shows multiple figures 20a, such as six-view drawings 21-23 and cross-sectional drawings 24, 25, which show the object 19 consisting of the object to be modeled 10 and its constituent elements 11 from different directions. Preferably, each figure 20a is accompanied by additional information 20b (only partially shown in Figure 1), such as dimension lines including dimensional values ​​and the material of the constituent elements 11.

[0013] The object to be modeled 10 in this embodiment is, for example, a building chimney 2 attached to a building 1. The building chimney 2 includes a plurality of chimney tubes 2a stacked vertically, support arms 2b that support each chimney tube 2a to the building frame 1b, an inspection opening 2c at the bottom, a cap section 2d at the top, and other components 11. Figures 20a include a front view 21, a plan view 22, a side view 23 of the building chimney 2, a plan cross-sectional view 24, a front cross-sectional view 25, etc. of the chimney tubes 2a which are components 11.

[0014] The three-dimensional model generation system 30 comprises an imaging means 31 and a generation unit 40. The imaging means 31 can be a digital camera, a dedicated scanner, a multifunction device with a scanner function, etc. A smartphone or laptop computer with a camera function may also be used as the imaging means 31.

[0015] The imaging means 31 captures the paper drawing 20, thereby obtaining graphic image data 39 of the object 19. Preferably, the graphic image data 39 includes not only image data of each graphic 20a itself, but also additional image data of additional information 20b such as dimension lines, dimension values, and material indications.

[0016] The graphic image data 39 during imaging is originally acquired as raster data. Note that software for converting raster data into vector data, vectorization means such as artificial intelligence (AI), etc. may be added to the imaging data 31, or the acquired graphic image data 39 may be automatically converted into vector data by the vectorization means. Regarding the additional information 20b represented by characters and symbols such as dimensional values and material indications in the paper drawing 20, the image thereof may be converted into text data by OCR (optical character recognition) means.

[0017] As the generation unit 40, a computer such as a personal computer (PC), a server computer, or the cloud can be used. The generation unit 40 generates a three-dimensional model 49 of the object 19 from the graphic image data 39. The three-dimensional model 49 is a virtual three-dimensional model constructed on a virtual three-dimensional space, and is preferably displayable on the computer display 48. The three-dimensional model 49 may be three-dimensional CAD data that can be displayed and edited on the display by three-dimensional CAD software.

[0018] Preferably, the generation unit 40 is constituted by an artificial intelligence computer system (hereinafter referred to as "AI system") storing the learned model 41. The learned model 41 is constructed by performing machine learning (including deep learning) for inferring a three-dimensional model of a learning object from learning graphic image data in advance.

[0019] The learning graphic image data is graphic image data of a learning object. The learning object is not limited to a three-dimensional object identical or of the same type as the object to be modeled 10, but may be a three-dimensional object of a type similar to the object to be modeled 10, or a three-dimensional object of a type different from the object to be modeled 10, and a variety of three-dimensional objects can be used. Preferably, the learning graphic image data includes a plurality of graphic image data such as six views, perspective views, cross-sectional views, etc. of the learning object viewed from a plurality of directions.

[0020] The three-dimensional model generation system 30 is used as follows. The imaging means 31 images the paper drawing 20 representing the object 19. Thereby, the graphic image data 39 of the object 19 is obtained. The graphic image data 39 is input into the generation unit 40. The imaging means 31 and the generation unit 40 may be connected by wire or wirelessly, and the graphic image data 39 may be transferred from the imaging means 31 to the generation unit 40. Alternatively, the graphic image data 39 may be transferred from the imaging means 31 to a portable storage medium (not shown), and then transferred from the storage medium to the generation unit 40.

[0021] In the generation unit 40, the graphic image data 39 is input into the learned model 41. The learned model 41 automatically generates the three-dimensional model 49 of the object 19 from the graphic image data 39.

[0022] The generated three-dimensional model 49 is displayed on the computer display 42. Preferably, the three-dimensional model 49 can be viewed from any direction, cross-sectioned at any cross-section, enlarged or reduced arbitrarily, further disassembled by component, or only some components can be displayed, or some components can be erased and displayed on the computer display 42. Thereby, the shape and structure of the object 10 to be modeled can be easily grasped. Also, two-dimensional drawing data (for example, front view data, plane cross-sectional view data, etc.) can be easily created, displayed, or printed from the three-dimensional model 49.

[0023] The additional information 20b such as the dimensional values and materials imaged from the paper drawing 20 and converted into text data may be linked to the corresponding part or component of the three-dimensional model 49 and displayed. Thereby, the configuration of the object 10 to be modeled can be grasped in more detail.

[0024] Next, another embodiment of the present invention will be described. Regarding the configurations that overlap with the previously described embodiments in the following embodiments, the same reference numerals will be given in the drawings and the description will be omitted. <Second Embodiment (FIG. 2)> As shown in Figure 2, in the three-dimensional model generation system 30B according to the second embodiment of the present invention, the generation unit 50 includes a template storage unit 51, an extraction unit 52, and a fitting unit 53. The template storage unit 51 stores a three-dimensional template model 54.

[0025] The three-dimensional template model 54 is a three-dimensional model of the typical, standard, or general three-dimensional shape of the object to be modeled 10, and variable elements such as specific shape, dimensions, proportions, arrangement, and the number of identical parts are represented by indeterminate parameters (variables). For example, in the three-dimensional template model 54 of the building chimney 2, variable elements such as whether the chimney tube 2a is a square tube or a cylindrical shape, the dimensions and proportions of the chimney tube 2a such as height, width, and thickness, and the number of chimney tubes 2a are represented by indeterminate parameters.

[0026] The extraction unit 52 extracts definitive information of variable elements from the graphic image data 39 from the imaging means 31. For example, it extracts definitive information from the graphic image data 39 such as whether the chimney tube 2a is square or cylindrical, the dimensions and ratios of the chimney tube 2a such as height, width, and thickness, and the number of chimney tubes 2a. This extraction may be performed using artificial intelligence (AI) functionality. For dimensions and the like, the imaging data may be converted into text data using OCR means and extracted as definitive information.

[0027] The fitting unit 53 generates a three-dimensional model 49 from the three-dimensional template model 54 by fitting the extracted confirmed values ​​to the indeterminate parameters of the three-dimensional template model 54. This makes it easy to understand the shape and structure of the object being modeled 10.

[0028] The present invention is not limited to the embodiments described above, and various modifications can be made without departing from its spirit. For example, the object to be modeled 10 in this invention is not limited to a building chimney 2, but may also be other building structures or building materials (e.g., voids embedded in floor slabs), or civil engineering structures or civil engineering materials (e.g., tunnel supports, pre-supported steel pipes for tunnel auxiliary construction methods). Furthermore, this invention is not limited to building structures or building materials, or civil engineering structures or civil engineering materials, but can be applied to the generation of three-dimensional models of various industrial products. [Industrial applicability]

[0029] This invention can be applied to a system for generating three-dimensional models from paper drawings. [Explanation of Symbols]

[0030] 1. Building 2. Building chimneys (objects to be modeled) 10 Objects to be Modeled 11 Components 19 objects 20 Paper drawings 20a Figure 20b Additional Information 30,30B Three-Dimensional Model Generation System 31 Imaging means 39. Graphic image data 40 Generation part 41 Pre-trained models 42 Computer displays 49 Three-dimensional models 50 Generation part 51 Template Storage Unit 52 Extraction part 53 Fitting section 54 Three-dimensional template models

Claims

1. A system for generating a three-dimensional model of an object to be modeled, An imaging means for obtaining graphic image data of the object by capturing a paper drawing in which a plurality of figures of the object to be modeled or its components are viewed from different directions, A generation unit that generates a three-dimensional model of the object from the aforementioned graphic image data, The generating unit comprises, A template storage unit stores a three-dimensional template model that models the typical, standard, or general three-dimensional shape of the object to be modeled, and in which variable elements including the dimensions and proportions of the object to be modeled are represented by indefinite parameters. An extraction unit that extracts definitive information of the variable element from the graphic image data, A fitting unit that generates the three-dimensional model from the three-dimensional template model by applying the confirmed information to the indeterminate parameters, A three-dimensional model generation system characterized by including the following.

2. The three-dimensional model generation system according to claim 1, wherein the object to be modeled is a building structure or building materials or a civil engineering structure or civil engineering materials.

Images