A calculation system that generates 3D point clouds from 2D architectural drawing data

Abstract

We provide a system that generates 3D point clouds from 2D architectural drawing data. [Solution] A system that generates a 3D point cloud from a 2D CAD drawing of a three-dimensional shape composed of target parts to be generated, such as the walls of a building, constructs a polygon of the target part to be generated using the relative coordinates of lines and faces in a layer that identifies the target part to be generated contained in the 2D CAD drawing, and by understanding the relative coordinates of the polygon, the layer information of the text information that identifies the target part to be generated, and the relative positions and cross-reference relationships of the accompanying lines, structural information is automatically added to the polygon information, and a 3D point cloud of the target part to be generated can be generated.If the target parts to be generated for a building related to the drawing are given in a list in the architectural drawing, etc., a collection of symbols and various information of the target parts to be generated in the list is constructed, and a 3D point cloud of the target part to be generated can be generated by adding various information from the list based on the positional relationship between the symbols and target parts in the building drawing.

Description

[Technical Field]

[0001] The present invention relates to a system for generating 3D point cloud data from 2D CAD drawing data of a building, and in particular to a system for generating point cloud data based on 2D CAD drawing data, thereby enabling inspection of whether a building has been finished according to the design drawings. [Background technology]

[0002] During the construction process of buildings such as office buildings and apartments, various people involved in the construction work, including government officials, building owners, and supervisors, conduct numerous inspections based on laws and regulations and design documents, such as construction supervisor inspections, confirmation application inspections, and fire inspections. Among these inspections, inspections of the finished construction parts, conducted by contractors such as general construction companies and specialist contractors for each process, are inspections to confirm and prove that the building has been constructed in accordance with the design documents, and are one of the most important tasks for contractors in managing the quality of buildings.

[0003] Proper inspection of the finished building can lead to the early detection of non-conformities that do not conform to the design documents during construction and the prevention of construction mistakes. This is expected to improve the quality of buildings and reduce lost profits, so there is a demand for labor-saving and accurate inspections.

[0004] The above-mentioned finish inspections are generally conducted by the contractor at each stage of construction or before other inspections, such as construction supervisor inspections. For each type of construction, such as rebar work, concrete work, fittings work, and interior work, inspections are conducted to check the quality of materials, including standards, dimensions, and materials, as well as the construction status and construction methods, and the results of various tests, such as concrete strength tests. Inspection methods include visual inspection by skilled technicians with the appropriate skills for the work, inspection using measuring tools such as tape measures and transits, and inspection using documents such as specifications, test reports, and construction photographs.

[0005] An example of a finish inspection is the inspection of the finished product after concrete is poured in a concrete construction project. This inspection involves visual inspection, construction photographic recording, and measurements using measuring instruments to check the dimensions, flatness, and position of the concrete components, as well as the presence or absence of defects such as lumps or cavities, and to confirm that the construction area conforms to the design documents.

[0006] The challenges of this inspection include the fact that visual inspection requires the inspector to directly look at the construction site and make a judgment, which makes the inspection cumbersome, the inspection accuracy depends on the inspector's skill, and there is a possibility that non-conforming areas may be overlooked. Furthermore, recording construction photographs and measuring with measuring instruments requires recording and measuring specific areas, and then repeating this process at multiple locations to determine whether the entire construction site complies with the design documents, which is an issue that requires a lot of time and effort.

[0007] The above issues are common to the inspection of the finish of construction work other than concrete work.In particular, when inspecting the dimensions and position of components that make up the structure, such as walls and ceilings, and the dimensions and position of partition walls, flooring, and ceilings installed in interior construction, if any non-conformities are overlooked, it will require a lot of money and time, and will have a significant impact on the progress of other construction work, so it is desirable to have an inspection method that can accurately and quickly check the entire building.

[0008] Patent Document 1 does not consider the above problem, and there is room for improvement. [Prior art documents] [Patent documents]

[0009] [Patent Document 1] Patent Application No. 2024-056796 Summary of the Invention [Problem to be solved by the invention]

[0010] In the construction industry, particularly in the civil engineering field, methods using 3D point cloud data acquired by Lidar sensors are becoming increasingly popular. If the 3D point cloud data acquired by Lidar sensors can be compared with building design documents, it is possible to prevent non-conforming areas from being overlooked when inspecting the finished building and improve the efficiency of inspections.

[0011] However, to compare 3D point cloud data with design documents, it is necessary to create a 3D model based on the design documents.Although 3D models can currently be created using 3D CAD, construction sites generally create drawings using 2D CAD, and there are issues such as the need for specialized knowledge to handle 3D CAD and the time it takes to create a 3D model.

[0012] Therefore, one of the objectives of this invention is to provide a system that generates 3D point cloud data of a building when constructed correctly according to the drawings from drawing data created by 2D CAD commonly used at construction sites, compares the generated 3D point cloud data with the 3D point cloud data of the building acquired by a Lidar sensor, and supports the inspection of the finished building.The drawings used in the system provided by this invention can be any type of drawing, such as a skeleton drawing, structural drawing, floor plan detail drawing, or house sales drawing, as long as they are drawing data created by 2D CAD. [Means for solving the problem]

[0013] According to one aspect of the present invention, there is provided a system for generating a three-dimensional shape of a building, which is configured by a target portion of the building, from a two-dimensional CAD drawing using a three-dimensional point cloud, comprising: means for constructing a polygon of the part to be generated using relative coordinates of lines and faces in a layer that identifies the part to be generated included in the two-dimensional CAD drawing; means for automatically assigning structural information such as height and width to the polygon information of the part to be generated by grasping the relative coordinates of the polygon of the part to be generated, the layer information of the character information specifying the part to be generated, and the relative positions and cross-reference relationships of the lines attached thereto, and generating a 3D point cloud of the part to be generated at a certain resolution based on the height and width information; The present invention provides a system comprising:

[0014] According to another aspect of the present invention, there is provided a system for generating a three-dimensional point cloud from a two-dimensional CAD drawing of a three-dimensional shape of a building, which is composed of walls, beams, openings, floors, ceilings, stairs, corridors, etc. (hereinafter referred to as a generation target part), and the system comprises: A polygon of the part to be generated is constructed using the relative coordinates of the lines and faces in the layer that identifies the part to be generated contained in the 2D CAD drawing. By understanding the relative coordinates of the polygons of the part to be generated, the layer information of the character information that identifies the part to be generated (for example, character information inside beam symbols, etc.), and the relative position and cross-reference relationship of the accompanying lines, Structural information such as height and width is automatically added to the polygon information of the target area, and a 3D point cloud of the target area is generated at a certain resolution based on that height and width information. If the target area is a wall, the height information of each floor is used as the height information of the wall, and a polygon is generated by a closed curve analysis of the wall area or line information. A 3D point cloud is generated based on the height information of beams, etc. If the part to be generated is a beam, the character information constituting the beam symbol is identified, and based on the relative coordinates of the character information, the height information of each level in the beam polygon candidate is used as a reference, the starting point position of the beam polygon is identified from the character information, and a point cloud is generated for the beam width, beam height, and height from the floor. If the object to be generated is an opening near a window or door frame, a point cloud is generated based on the relative coordinates of the information inside the upper and lower parts of the building, and the start and end positions of the frame structure of the upper and lower parts of the window frame. If the object to be generated is the floor or ceiling of a room or hallway, the relative coordinates of the text information indicating the slab structure are identified, and the floor and ceiling polygons are generated as a point cloud from the reference position. If the object to be generated is a staircase, a polygonal line is drawn across the polygon representing the staircase, and the polygon that intersects with the polygonal line is given predetermined step information to generate a 3D point cloud. The individually generated objects are combined to create a point cloud of the entire room. Identify the character information indicating the sleeve, identify the sleeve diameter and horizontal and vertical position information, and remove the 3D point cloud of the corresponding generation target part, excluding the point cloud of the wall surface, and generate a 3D point cloud of all generation target parts, In addition, when the building's generation target objects related to the drawing are given in a list, a collection of symbols and various information for the generation target parts in the list is constructed, and a 3D point cloud is generated by adding various information from the list based on the positional relationship between the symbols and the target parts in the building's drawing, We will develop a system that detects construction defects inside buildings, etc. by comparing the distance between the generated point cloud and 3D point clouds measured at actual construction sites using Lidar sensors, SLAM, etc. [Effects of the Invention]

[0015] By comparing point clouds generated from 2D CAD drawing data and measured point clouds using this invention, areas of construction defects in buildings can be quantitatively identified, enabling a more sophisticated evaluation than conventional visual inspections. [Brief explanation of the drawings]

[0016] [Figure 1] FIG. 1 shows an outline of the system configuration of the present invention. [Figure 2] Figure 2 shows an overview of the internal structure of the system for generating a 3D point cloud of a building of the present invention, and shows examples of a system for generating a 3D point cloud from a 2D drawing, a system for measuring a building, and a system for comparison. [Figure 3] FIG. 3 shows an example of information processing according to the present invention, and in particular shows an example of information processing for generating a point cloud of a concrete wall surface. [Figure 4] FIG. 4 shows an example of information processing according to the present invention, and in particular shows an example of information processing for generating a point cloud of a beam. [Figure 5] FIG. 5 shows an example of information processing according to the present invention, and in particular shows an example of information processing for generating a point cloud of the floor and ceiling. [Figure 6] FIG. 6 shows an example of information processing according to the present invention, and in particular shows an example of information processing for generating a point cloud of the concrete wall surfaces above and below the opening. [Figure 7] FIG. 7 shows an example of information processing according to the present invention, and in particular shows an example of information processing for generating a point cloud of stairs. [Figure 8] FIG. 8 shows an example of information processing according to the present invention, and in particular shows an example of information processing for generating a point cloud of slopes of a corridor or a balcony. [Figure 9] FIG. 9 shows an example of information processing according to the present invention, and in particular shows an example of information processing for generating a point cloud by performing coordinate transformation on a two-dimensional elevation drawing. [Figure 10] FIG. 10 shows an example of information processing for combining generated point clouds and removing point clouds of the sleeve portion. [Figure 11] FIG. 11 shows an example of information processing of the present invention, particularly showing an example of information processing for generating a point cloud when various information about the object to be generated, such as a building, is given in a list drawing and a two-dimensional drawing of the structure is given with outline symbols, etc. [Figure 12] FIG. 12 shows an example of information processing for comparison with the measurement point cloud. [Figure 13] FIG. 13 shows an example of the results of the information processing of the present invention, and in particular shows an example of combining point clouds of the generated object. [Figure 14] FIG. 14 shows an example of an embodiment of a measurement point cloud. [Figure 15] FIG. 15 shows an example of an embodiment of the comparison result of the measurement point clouds. DETAILED DESCRIPTION OF THE INVENTION

[0017] Hereinafter, an embodiment of the present invention will be described.

[0018] In recent years, there has been an increase in the use of "point cloud data," a collection of multiple points with additional information such as 3D coordinates, color, normals, etc. Point cloud data allows for a three-dimensional understanding of the shape of an object, making it possible to grasp the shape of an object over a wide area more efficiently than measuring dimensions using measuring tools such as a tape measure.

[0019] For example, if a method could be developed to utilize point cloud data in the inspection of the finished concrete structure that is carried out after concrete pouring work, it would be possible to comprehensively and quickly check the entire concrete structure, which is expected to reduce the labor required for inspection and improve accuracy.

[0020] Furthermore, if a method could be developed to utilize point cloud data in inspections to check the overall dimensions and position of a construction site, such as inspecting the completion of interior construction work or inspecting the installation of steel frame work, just as in the inspection of the completion of concrete frame structures mentioned above, it would be possible to comprehensively and quickly check the construction site, which is expected to reduce the labor required for inspections and improve accuracy.

[0021] The inspection of the finish of the aforementioned concrete frame structure is generally carried out by visually checking it against a 2D drawing using measuring tools such as a tape measure. However, visual inspection is cumbersome, and the accuracy of the inspection depends on the skill of the inspector, which means there is a risk of overlooking non-conforming areas.

[0022] Therefore, in this invention, we have invented a system that generates 3D point cloud data of a building when constructed correctly according to the drawing data from 2D CAD drawing data of a skeleton drawing created using 2D CAD commonly used on construction sites, and compares the generated 3D point cloud data with 3D point cloud data of the building acquired by a Lidar sensor or the like to verify the constructed building against the contents of the 2D drawing, thereby inspecting the finished concrete skeleton structure.This invention makes it possible to efficiently verify the 3D point cloud data of the building after construction with the skeleton drawing and design documents, resulting in labor-saving and improved accuracy in inspecting the finished concrete skeleton structure.

[0023] Furthermore, according to the present invention, it is possible to generate 3D point cloud data corresponding to each drawing from 2D CAD drawing data other than structural drawings, such as floor plan detail drawings, structural drawings, and housing sales drawings.Therefore, by comparing the generated 3D point cloud data with the 3D point cloud data of the building acquired by a Lidar sensor, etc., it is possible to inspect whether the building has been finished according to the design documents in situations other than inspecting the finish of the concrete structural body. [Example]

[0024] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.

[0025] FIG. 1 shows an overview of a system according to an embodiment of the present invention, and shows an example of a system for generating a point cloud of the interior of a building, etc.

[0026] In S110, a two-dimensional CAD drawing of the building is input.

[0027] In S120, the user sets various information related to the two-dimensional CAD drawing.

[0028] In S130, 3D point cloud data is generated from the 2D CAD drawing based on the architectural drawing and various information.

[0029] At S140, point cloud data during construction is measured during on-site surveys using Lidar sensors and SLAM point cloud measurement devices.

[0030] In S150, check the uploaded 3D point cloud results.

[0031] In S160, the point cloud generated from the drawing is compared with the measured point cloud.

[0032] In S170, based on the comparison results, defective construction areas are determined.

[0033] In S180, the user views data on areas of poor construction.

[0034] In S190, data on defective construction areas is downloaded.

[0035] The backend is intended to run on a server, but it can also run on a user's PC. In this embodiment, an example has been described in which a concrete skeleton drawing is used as the 2D CAD drawing of a building. The 2D CAD drawing of a building is not limited to a concrete skeleton drawing, and other drawings may be used. For example, in addition to this embodiment, structural drawings, detailed floor plans, and pamphlet drawings (drawings for selling homes) of a building may also be applied.

[0036] Figure 2 shows an overview of the point cloud generation part on the server or user PC.

[0037] In S210, two-dimensional CAD drawing data is read.

[0038] In S220, the user sets various settings related to the 2D CAD drawing data.

[0039] In S230, the 2D CAD drawing data is analyzed and a point cloud is generated. The point cloud generation method is described in detail in S3 to S11 below. In S240, the measured point cloud data is compared with the measured point cloud data. The detailed comparison method is described in S12.

[0040] S250 detects and judges construction defects inside buildings, etc.

[0041] FIG. 3 shows an example of information processing according to the present invention, and in particular shows an example of information processing for generating a point cloud of a concrete wall surface.

[0042] In S310, a layer showing the line segments of the concrete wall surface is input. Specifically, a layer showing polygons colored as concrete parts or lines surrounding the concrete parts is selected.

[0043] S320 generates polygons by connecting the lines of the concrete wall surface, and also automatically generates a group of closed curves for the lines surrounding the concrete part. In S330, polygons of the concrete wall surface are identified.

[0044] In S340, a concrete wall polygon is generated on a two-dimensional plane XY coordinate system based on a group of closed curves surrounding the concrete part. A point cloud is arranged at a certain resolution inside this polygon and on the outer periphery.

[0045] In S350, the height between floors (the height from the floor of the floor where the relevant concrete wall is located to the floor of the floor above) is used as an input value, and a point cloud is generated at a certain resolution directly above the Z axis of the point cloud generated in S330.

[0046] In S320 and S330, a GUI program for manually correcting the line group allows correction and selection change of the closed curve group and concrete wall polygons.

[0047] FIG. 4 shows an example of information processing according to the present invention, and in particular shows an example of information processing for generating a point cloud of a beam.

[0048] In S410, layers indicating the beam boundary and beam symbol are input.

[0049] In S420, grouping of character information and line information relating to beam symbols is performed based on the relative distance, the structure of the lines of the beam symbols, the relative distance, and the amount of information, to generate a beam symbol set.

[0050] In S430, closed curve analysis is performed on the line segments that indicate the boundaries of the beams to generate candidates for beam polygons.

[0051] In S440, a beam polygon is searched for from among the beam polygon candidates based on the beam width and beam shape indicated by the beam symbol information, the rotation angle of the information, and the relative position with respect to the beam polygon candidate, and the beam polygon is determined.

[0052] In S450, the height between floors (the height from the floor of the floor where the relevant concrete wall is located to the floor of the floor above) is used as an input value, and based on the beam symbol information, a point cloud is generated for the outer periphery of the beam polygon and the lower surface of the beam inside.

[0053] In S460, a point cloud is generated at a certain resolution directly above the Z axis of the point cloud generated in S450 based on the beam width, beam configuration, and beam bottom height information from the floor in the beam symbol information.

[0054] In addition, S430 and S450 allow the user to manually modify and add lines or polygons using a GUI program that allows the user to modify or select and change the closed curve group and beam polygons.

[0055] FIG. 5 shows an example of information processing according to the present invention, and in particular shows an example of information processing for generating a point cloud of the floor and ceiling.

[0056] In S510, the line segments indicating the boundaries of floors and ceilings, steps, etc., and the layer groups indicating slab symbols are input.

[0057] In S520, character information and line information relating to slab symbols are grouped according to the relative distance, the structure of the lines of the beam symbols, the relative distance, and the amount of information, to generate a group of slab symbols.

[0058] In S530, closed curve analysis is performed on the line segments that indicate the boundaries of the floor and ceiling, steps, etc., to generate a group of polygons that make up the floor and ceiling.

[0059] In S540, a combination of each floor / ceiling composing polygon group and slab symbol group is searched for based on the relative position or designated position of the slab symbol and the floor / ceiling composing polygon group, and each floor / ceiling composing polygon group is determined.

[0060] In S550, information on the difference in floor height between floor levels is input, the height of the floor and ceiling polygon group is determined based on the slab symbol, and the perimeter line of the floor and ceiling polygon group and a point cloud within the polygon group are generated.

[0061] In S560, floor height information between floors and a point cloud in the Z direction based on the floor height information are generated.

[0062] In S570, a point cloud in the Z direction is generated based on floor height information and ceiling height information between floors.

[0063] S580 simultaneously assigns floor and ceiling heights to polygons that do not have any information attached and generates point clouds.

[0064] In S530 and S540, a GUI program that allows you to manually modify and add lines or add polygons makes it possible to modify or select and change the polygons that make up the closed curves and floor / ceiling.

[0065] FIG. 6 shows an example of information processing according to the present invention, and in particular shows an example of information processing for generating a point cloud of the interior of a building above and below an opening.

[0066] In S610, the line segments indicating the boundaries of the openings and the layer group indicating the opening symbols are input.

[0067] In S620, a set of aperture symbols is generated by grouping character information and line information related to the aperture symbols.

[0068] In S630, a closed curve analysis is performed on the line segments that indicate the boundaries of the openings to generate opening polygon candidates.

[0069] In S640, combinations of aperture polygons and aperture symbols are generated based on the rotation angle of the aperture symbol, the classification, and the relative positions of the aperture polygon candidates.

[0070] In S650, the opening polygon is determined based on the combination of the opening polygon and the opening symbol, and a point cloud is generated inside the polygon above and below the opening.

[0071] In S660, the difference in floor elevation between floors is input, and based on the opening symbol, a point cloud of the concrete structure above and below the opening is generated in the vertical Z-axis direction.

[0072] In S670, a process of generating or removing a point cloud is added to reproduce the window frame support portion.

[0073] In addition, S630 and S650 allow the user to manually modify and add lines or polygons using a GUI program that allows the user to modify or select and change groups of closed curves and beam polygons.

[0074] FIG. 7 shows an example of information processing according to the present invention, and in particular shows an example of information processing for generating a point cloud in a staircase portion.

[0075] In S710, the layer of the line segments that make up the stairs is selected.

[0076] In S720, staircase polygon candidates are generated by combining the line segments that make up the staircase.

[0077] In S730, a polygon having a reference height is used as a starting position, and a polygonal line that crosses the staircase polygon candidate is created or selected, and the polygon that the polygonal line crosses is designated as a staircase polygon.

[0078] In S740, a point group is generated in the reference height polygon.

[0079] In S750, the height of each staircase polygon step is determined based on the input height information for each staircase step, and a point cloud is generated for each polygon.

[0080] FIG. 8 shows an example of information processing according to the present invention, and in particular shows an example of information processing for generating a point cloud of a corridor, a balcony, etc. In S810, the layer of the lines that make up the corridor, balcony, etc. is selected.

[0081] In S820, polygon candidates for generating the slope of a balcony, a corridor, etc. are generated by joining line segments, etc.

[0082] In S830, polygons that create slopes for balconies, corridors, etc. are finalized.

[0083] In S840, the relative Z coordinates used as the inclination references for the four or more corners for setting the inclination reference height are input.

[0084] In S850, an inclined plane is constructed based on the relative Z coordinate and the XY plane coordinate values ​​that serve as the inclination reference.

[0085] In S860, a point cloud is generated on an inclined plane at any resolution.

[0086] FIG. 9 shows an example of information processing according to the present invention, and in particular shows an example of information processing for generating a point cloud inside a building using structural polygons drawn on the XY plane.

[0087] In S910, a structural polygon drawn on a two-dimensional CAD drawing is selected.

[0088] In S920, coordinate conversion (XZ coordinate system or YZ coordinate system) of the structural polygon is performed to generate a point cloud around and inside the polygon.

[0089] In S930, the center line of the structural polygon is determined (dotted line).

[0090] In S940, the center line for placing the structural polygon is determined (solid line with arrow).

[0091] In S950, the structural polygons are replicated on the placement center line at a constant resolution.

[0092] An example of information processing for combining generated point clouds and removing point clouds of the sleeve portion will be described below.

[0093] In S1010, the point clouds of the construction object generated in S340, S460, S580, S670, S750, S860, and S950 are combined.

[0094] In S1020, the points corresponding to the sleeve are removed from the combined points. The position of the sleeve is determined by combining the character information and line information indicating the position information.

[0095] In S1030, absolute coordinates are assigned to the relative coordinates to align the point clouds of each floor.

[0096] In S1040, the point cloud of the building is completed.

[0097] FIG. 11 shows an example of information processing of the present invention, particularly an example of information processing for generating a point cloud when various information about a building is given in a list in an architectural drawing, etc., and only general symbols are given in the drawing of the structure.

[0098] In S1110, layers representing the target parts of the building to be generated are input. Here, the target parts of the building to be generated are everything that is actually shown on the drawings of the construction site, including, for example, walls, beams, floors, ceilings, openings, stairs, corridors, columns independent of walls, balconies and corridor waist-length walls, air conditioning equipment, toilets, kitchens, sinks, partition walls, windows, doors, built-in shelves, etc.

[0099] In S1120, the part to be generated indicated by the symbol is identified based on the relative position, width, etc. of the symbol and the part to be generated.

[0100] In S1130, grouping is performed on the symbols of the parts to be generated in the list in the architectural drawing and the width, height, depth, etc. of the objects.

[0101] In S1140, the width, height, depth, etc. based on the symbols of the grouped generation target parts are collated and assigned to the generation target parts of the two-dimensional drawing.

[0102] In S1150, a point cloud in the polygon of the generation target portion at the reference height is generated based on the reference height.

[0103] In S1160, a point cloud of the generation target portion is generated or a point cloud is removed.

[0104] In addition, in S1120, S1130, and S1140, the closed curve group and beam polygons can be corrected or selected and changed by a GUI program that allows lines to be manually corrected or added or polygons to be added.

[0105] FIG. 12 shows an example of information processing for comparison with the measurement point cloud.

[0106] In S1210, the point cloud generated from the 2D drawing and the measurement point cloud are roughly aligned manually or using an automatic algorithm.

[0107] In S1220, an automatic algorithm is used to precisely align the point cloud generated from the 2D drawing with the measured point cloud.

[0108] In S1230, the distances between the point groups are compared using the nearest point for each point group.

[0109] In S1240, construction defects in areas with large distances are checked. [Example]

[0110] <Generating 3D point cloud models from 2D drawings> Figure 13 shows an example of the results of generating a 3D point cloud model from a 2D drawing. The building is generated for each part according to the steps S3 to S9. It should also be noted that a similar 3D point cloud can be generated using the method of S11.

[0111] <Point cloud data merging> The point cloud data for each part of the building is combined using relative coordinates. By matching the coordinates on the relative coordinate system, the point cloud for the building is generated.

[0112] <Point cloud data processing> To reduce the computational load, the generated point cloud data can be thinned out or divided into rooms for analysis. In fact, Figure 13 shows a 3D point cloud model of a whole floor generated from a 2D drawing, and then specific rooms are extracted. This processing allows for high-speed analysis. [Example]

[0113] Figure 14 shows a point cloud measured by a SLAM point cloud measurement device. Any method for measuring the point cloud can be used, such as a fixed or SLAM Lidar sensor, as long as it can measure the point cloud inside the room.

[0114] <Measurement preparation and measurement execution> When measuring, a fixed or SLAM-type Lidar sensor is used to measure point clouds in every corner of the room. If the resolution of the measured point cloud is poor, the distance between the points will be large when compared with a point cloud generated from 2D CAD drawing data, and there is a risk that it will be recognized as an area of ​​poor construction, so point cloud measurement is carried out over a certain amount of time to obtain a large point cloud density.

[0115] Furthermore, since point cloud alignment often uses color information, it is desirable to measure using electric lights or other devices that allow the color information inside the building to be seen. On the other hand, Lidar sensors can reduce measurement accuracy due to diffuse reflection from sunlight, so care must be taken to avoid shadows caused by sunlight. Also, while photographing the interior of a building, there may be luggage or other items on site, but it is desirable to keep this to a minimum.

[0116] <Comparison of 3D generated point cloud and measurement results> Figure 15 shows a comparison, in millimeters, of the results of a 3D point cloud generated from a 2D CAD drawing and the point cloud measurements carried out during an on-site inspection. Areas where there is a large difference between the 3D point cloud generated from the 2D CAD drawing and the measured point cloud carried out during an on-site inspection may indicate areas of poor construction. However, areas with a large difference between the generated point cloud and the measured point cloud may contain items such as luggage, as described in Figure 14, or electrical wires hanging from the ceiling, so it is advisable to evaluate areas of poor construction, paying attention to the following points:

[0117] <Evaluation of construction defects> Areas where there is a possibility of large discrepancies in point cloud comparison include steps in floors and ceilings, areas around window and door frames on walls, beams, openings in walls, and sleeves on walls. For steps in floors and ceilings, the construction accuracy must be verified, taking into account the effects of shadows. For areas around window and door frames on walls, the construction accuracy must be evaluated, taking into account the thickness of the building's concrete structure. For beams, the construction accuracy must be verified to ensure that they are positioned appropriately. For openings in walls, careful verification must be made to ensure that the position, height, and width of the opening are set appropriately. For sleeves on walls, the position of the sleeve must be evaluated to ensure that it matches the 2D drawing.

[0118] Although the embodiments of the present invention have been described above, various alternatives, modifications, and variations are possible for those skilled in the art based on the above description, and the present invention includes the various alternatives, modifications, and variations described above within the scope of the present invention. Each information processing described in this specification can be realized by appropriately combining any hardware such as a processor or a computer and any software such as a program depending on the content of the information processing.

Claims

1. A system that generates a three-dimensional shape composed of a target part of a building from a two-dimensional CAD drawing as a three-dimensional point cloud, means for constructing a polygon of the part to be generated using relative coordinates of lines and faces in a layer that specifies the part to be generated included in the two-dimensional CAD drawing; means for automatically assigning structural information of height and width to the polygon information of the part to be generated by grasping the relative coordinates of the polygons of the part to be generated, the layer information of the character information specifying the part to be generated, and the relative positions and cross-reference relationships of the lines attached thereto, and generating a three-dimensional point cloud of the part to be generated at a certain resolution based on the height and width information; Equipped with The system is provided with a means for combining individually generated objects to generate a three-dimensional point cloud for the entire interior of a room, identifying character information indicating a sleeve, identifying the sleeve diameter and planar / vertical position information, and removing the three-dimensional point cloud for the corresponding generated object portion, thereby excluding the point cloud for the wall surface and generating a three-dimensional point cloud for all generated object portions.

2. The system of claim 1, further comprising a means for detecting construction defects inside a building by comparing the distance between the generated point cloud and a three-dimensional point cloud measured within an actual construction site.

3. The generation target portion of the building is at least one of a wall, a beam, an opening, a floor, a ceiling, a staircase, a corridor, a column independent of a wall, a balcony or a corridor's waist-length wall, an air conditioning system, a toilet, a kitchen, a sink, a partition wall, a window, a door, and a built-in shelf, The system according to claim 1 , wherein the character information identifying the part to be generated is character information within a component symbol.

4. If the target part is a wall, the height information of each floor is used as the height information of the wall, and a polygon is generated by a closed curve analysis of the wall area or line information, and a 3D point cloud is generated based on the height information of the beam. If the part to be generated is a beam, the character information constituting the beam symbol is identified, and based on the relative coordinates of the character information, the height information of each level in the beam polygon candidate is used as a reference, the starting point position of the beam polygon is identified from the character information, and a three-dimensional point cloud including the beam width, beam height, and height from the floor is generated. If the object to be generated is an opening near a window or door frame, a 3D point cloud is generated based on the relative coordinates of the information inside the upper and lower parts of the building, and the start and end positions of the frame structure of the upper and lower parts of the window frame. If the object to be generated is the floor or ceiling of a room or hallway, the relative coordinates of the text information indicating the slab structure are identified, and a 3D point cloud is generated from the reference position of the floor and ceiling polygons. The system of claim 1, wherein when the object to be generated is a staircase, a three-dimensional point cloud is generated by drawing a broken line that crosses a polygon representing the staircase and assigning predetermined step information to the polygon that intersects with the broken line.

5. The system described in claim 1, wherein when the building generation objects related to the drawing are given in a list, a collection of symbols of the generation objects in the list and various information is created, and a three-dimensional point cloud is generated by adding various information from the list based on the positional relationship between the symbols and the object objects in the building drawing.

6. The system described in claim 1, which compares the generated 3D point cloud with a 3D point cloud of a building measured by a Lidar sensor or SLAM within an actual construction site.

7. A system for generating a three-dimensional shape consisting of a target part of a building from a two-dimensional CAD drawing using a three-dimensional point cloud, means for constructing a polygon of the part to be generated using relative coordinates of lines and faces in a layer that specifies the part to be generated included in the two-dimensional CAD drawing; means for automatically assigning structural information of height and width to the polygon information of the part to be generated by grasping the relative coordinates of the polygons of the part to be generated, the layer information of the character information specifying the part to be generated, and the relative positions and cross-reference relationships of the lines attached thereto, and generating a three-dimensional point cloud of the part to be generated at a certain resolution based on the height and width information; Equipped with If the target part is a wall, the height information of each floor is used as the height information of the wall, and a polygon is generated by a closed curve analysis of the wall area or line information, and a 3D point cloud is generated based on the height information of the beam. If the part to be generated is a beam, the character information constituting the beam symbol is identified, and based on the relative coordinates of the character information, the height information of each level in the beam polygon candidate is used as a reference, the starting point position of the beam polygon is identified from the character information, and a three-dimensional point cloud including the beam width, beam height, and height from the floor is generated. If the object to be generated is an opening near a window or door frame, a 3D point cloud is generated based on the relative coordinates of the information inside the upper and lower parts of the building, and the start and end positions of the frame structure of the upper and lower parts of the window frame. If the object to be generated is the floor or ceiling of a room or hallway, the relative coordinates of the text information indicating the slab structure are identified, and a 3D point cloud is generated from the reference position of the floor and ceiling polygons. When the object to be generated is a staircase, the system generates a 3D point cloud by drawing a broken line that crosses the polygon representing the staircase and adding predetermined step information to the polygon that intersects with the broken line.

8. A program for executing the system according to any one of claims 1 to 7.

Images