Signature processing system

Abstract

To provide a signature processing system for reducing effort required for checking error in a construction report document and accepting the construction report document.SOLUTION: A system includes: two-dimensional codes 40A-40G in which a plurality of cells 401 with predetermined colors are arranged two-dimensionally; and a notch part 404 for detecting location of a signature area. A color combination represents information related to civil engineering work. The system further includes a recognition program which calculates reliability of whether a recognition of an image signature by a recognition part 162 is accurate according to positional relationship of each cell 401 constituting the two-dimensional codes 40A-40G (in-image signature) in a completion drawing or according to a shape profile of the in-image signature.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a tag used by being arranged at a civil engineering work site, and includes a tag associated with information related to civil engineering work, a photographing device that takes a plurality of photographs of the civil engineering work site, and a recognition unit for recognizing an in-image tag that is a tag in a composite image obtained by synthesizing the plurality of photographs, and relates to a tag processing system.

Background Art

[0002] In civil engineering work for laying underground structures such as gas conduits, the contractor who has carried out the civil engineering work creates and submits a construction report document describing the results of the civil engineering work to the contractor who ordered the civil engineering work. Specifically, the construction report document refers to drawings (construction drawings and as-built drawings) for showing the laying state of the underground structure, a work report for reporting the content of the civil engineering work and the results of the airtightness test performed after laying the conduit, a settlement statement for reporting the costs incurred according to the content of the civil engineering work, and the like. As a system for creating as-built drawings, for example, a construction drawing creation support system disclosed in Patent Document 1 is known.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Various information is required for the construction report document. For example, for creating as-built drawings, material information and dimensional information of the underground structure are required. Also, for creating a work report, information related to the work content is required. Further, for creating a settlement statement, in addition to the above-mentioned information, information on the site situation and the number of workers is required, and a lot of labor is spent to create the construction report document by describing a lot of information.

[0005] Under such circumstances, since OA equipment is required for creating construction report materials, at the civil engineering construction site, information necessary for creating construction report materials is memoed and remembered, and later, at the office of the contractor who carried out the civil engineering work, based on such memo and memory, the information necessary for creating construction report materials is written out clearly, transmitted to the person in charge of creation, and supplemented with explanations as necessary to create the construction report materials. Therefore, when creating construction report materials, the efficiency decreases because multiple personnel are involved, and mistakes occur in the memorization, recording, transmission, and input of information. For this reason, in addition to the labor for creating construction report materials, a lot of labor is spent on checking for mistakes and inspecting the construction report materials.

[0006] The present invention is for solving the above problems, and an object thereof is to provide a document processing system capable of reducing the labor involved in checking mistakes in construction report materials and inspecting the construction report materials.

Means for Solving the Problems

[0007] In order to solve the above problems, the document processing system of the present invention has the following configuration.

[0008] (1) In a document processing system including a document associated with information on the civil engineering work, which is arranged and used at the civil engineering work site, a photographing device that takes a plurality of photographs of the civil engineering work site, and a recognition unit for recognizing an in-image document that is the document appearing in the image taken by the photographing device, the document has a plurality of cells having a predetermined color arranged two-dimensionally, includes a notch for detecting the position of the area of the document, and represents the information by a combination of colors. A recognition program for calculating the reliability of whether the recognition of the in-image document by the recognition unit is accurately performed based on at least any one of the positional relationship of the cells constituting the in-image document, the number of the cells, the color of the cells, the shape of the in-image document, the color of the in-image document, and the information read from the in-image document.

[0009] (2) In the document processing system described in (1), it is preferable that the positional relationship is the distance between the centers of gravity of the respective cells constituting the document in the image, and the reliability is calculated based on the distance.

[0010] (3) In the document processing system described in (1) or (2), it is preferable that the positional relationship is the angle of the line segment when the centers of gravity of the respective cells are connected by a line segment, and the reliability is calculated based on the angle.

[0011] (4) In the document processing system described in (1), it is preferable that an outer circumscribed rectangle surrounding the document in the image is obtained based on the shape of the document in the image, and the reliability is calculated based on the area of the outer circumscribed rectangle.

[0012] (5) In the document processing system described in (1), it is preferable that an outer circumscribed rectangle surrounding the document in the image is obtained based on the shape of the document in the image, and the reliability is calculated based on the difference between the area of the document in the image and the area of the outer circumscribed rectangle.

[0013] (6) In the document processing system described in (1), it is preferable that an outer circumscribed rectangle surrounding the document in the image is obtained based on the shape of the document in the image, and the reliability is calculated based on the interior angles of the outer circumscribed rectangle.

[0014] (7) In the document processing system described in (6), it is preferable that the interior angles of the document in the image are calculated based on the shape of the document in the image, and the reliability is calculated based on whether the interior angles are within a predetermined threshold value.

[0015] (8) In the document processing system described in (1), it is preferable that the center of gravity of the document in the image is calculated based on the shape of the document in the image, and the reliability is calculated based on the center of gravity.

[0016] (9) In the document processing system described in (1), it is preferable that the area of the document in the image is calculated based on the shape of the document in the image, and the reliability is calculated based on the area.

[0017] (10) In the document processing system described in (1), it is preferable to calculate the area of each cell constituting the in-image document based on the shape of the in-image document, and calculate the reliability based on the area.

[0018] (11) In the document processing system described in (1), it is preferable to recognize the number of corners of the in-image document based on the shape of the in-image document, and calculate the reliability based on the number of corners.

[0019] (12) In the document processing system described in (1), it is preferable to include a positioning device used at the civil engineering site, and calculate the reliability based on the position information acquired by the positioning device and the information read from the in-image document.

[0020] (13) In the document processing system described in (1), it is preferable to compare the information read from the in-image document with the past civil engineering achievements and calculate the reliability.

[0021] (14) In the document processing system according to any one of (1) to (13), it is preferable to include a document generation unit that creates a construction report document describing the results of the civil engineering based on the information associated with the document, and the reliability calculated by the recognition program is described in the construction report document.

[0022] (15) In the document processing system described in (1), the recognition program determines whether the document in the image is a document or not according to the threshold values of each item among the number of cells constituting the document in the image, the shape of the cells, the color of the cells, the positional relationship of the cells, the shape of the document in the image, and the color of the document in the image. When resetting the threshold value in the item related to the determination of whether it is a document among each item, the reliability is only deducted (the first deduction) in the item where the resetting is performed. When resetting the threshold value in the item other than the item related to the determination of whether it is a document and the item related to the determination of whether it is a document among each item, in the item where the resetting is performed, the reliability is deducted in a small amount with respect to the first deduction for the item other than the item related to the determination of whether it is a document, which is preferable.

[0023] (16) In the document processing system described in (1), it is preferable that the recognition program calculates the reliability by extracting the cells from the image captured by the imaging device and determining whether the set of the cells is the document.

[0024] (17) In the document processing system described in (16), it is preferable that the recognition program determines whether the set of the cells is the document based on the shape of the set of the cells extracted from the image captured by the imaging device.

[0025] (18) In the document processing system described in (15) or (16), the recognition program determines whether the cell is a cell or not according to the threshold values of the number of cells, the shape of the cells, the color of the cells, and the positional relationship of the cells for extracting the cells constituting the document in the image. When resetting the threshold value in the item related to the determination of whether it is a cell among each item, the reliability is only deducted (the first deduction) in the item where the resetting is performed. When resetting the threshold value in the item other than the item related to the determination of whether it is a cell and the item related to the determination of whether it is a document among each item, in the item where the resetting is performed, the reliability is deducted in a small amount with respect to the first deduction for the item other than the item related to the determination of whether it is a cell, which is preferable.

Advantages of the Invention

[0026] According to the above-mentioned document processing system, since the reliability of whether the recognition of the in-image document is accurately performed is calculated based on the recognition program, by using this reliability in, for example, construction report materials, it is possible to check for mistakes in the construction report materials and reduce the labor required for inspection of the construction report materials. Specifically, for example, when information (such as the cost generated according to the work content of the civil engineering work) is described in the construction report materials based on the information associated with the document, the reliability of whether the corresponding document is accurately recognized is also noted. By doing so, when checking and inspecting the content of the construction report materials, it is only necessary to focus on checking those with low reliability, those with a reliability of 50 or less where the recognition program stopped recognizing, and those corrected by the operator, so the labor can be reduced.

Brief Description of the Drawings

[0027]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Embodiments for Carrying Out the Invention

[0028] An embodiment of the document processing system 1 of the present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram showing an example of the configuration of the document processing system 1 of this embodiment. FIG. 2 is a diagram showing an example of a state of photographing the civil engineering work site using a photographing device. FIG. 3 is a diagram showing an example of the laying state of the conduit 20 and the joint 21. FIG. 4(a) is a diagram showing an example of the two-dimensional codes 40A - 40G, and (b) is a diagram showing an example of the color of the cell and the number corresponding to the color. FIG. 9 is a diagram showing an example of the construction report material R.

[0029] The document processing system 1 is a system used to create a construction report material describing the results of civil engineering work for laying gas conduits 20 and joints 21 (an example of buried objects). The user is a contractor (construction contractor) who performs civil engineering work. The construction contractor uses the document processing system 1 to create the construction report material R and submits the construction report material R to the client who orders the civil engineering work. Here, the construction report material R refers to, for example, as shown in FIG. 9, a drawing D (construction drawing or as-built drawing) for showing the laying state of the buried object, a work report W for reporting the content of the civil engineering work and the results of the airtightness test performed after laying the conduit, a settlement statement S for reporting the expenses generated according to the content of the civil engineering work, and the like.

[0030] Although it does not matter how the embedded objects are laid, in this embodiment, for example, as shown in FIG. 3, the case where the work of laying the conduits 20A, 20B and the joints 21A, 21B, 21C is performed will be described. The work of laying the conduits 20A, 20B and the joints 21A, 21B, 21C is, more specifically, the work of laying the conduits 20A, 20B and the joints 21A, 21B, 21C in the installation groove 30a provided by excavating the ground 30 in order to connect the existing pipes 22A, 22B that have already been buried in the ground. The conduits 20A, 20B are not particularly limited, but for example, they are straight pipes with a length of 5 m and a nominal diameter of 100A. In the following, unless particularly necessary, when explaining the joints 21A, 21B, 21C and the joints 21A, 21B, 21C, the alphabets in the reference numerals are omitted and denoted as the conduit 20 and the joint 21.

[0031] As shown in FIG. 1, the document processing system 1 includes a construction report document creation device 10, a photographing device 11, a positioning device 12, a pre-input device 18, a dimension correction member 24, and a post-verification unit 25. The construction report document creation device 10 may be a device installed in the office of the construction company or a cloud server. Further, the photographing device 11, the positioning device 12, the pre-input device 18, and the dimension correction member 24, which are surrounded by a broken-line frame in FIG. 1, are devices used at the civil engineering construction site (collectively referred to as the on-site devices 2), and are connected to the construction report document creation device 10 via a communication line 19 such as the Internet. Further, the post-verification unit 25 may be a device installed in the office of the client of the civil engineering construction or a cloud server.

[0032] The photographing device 11 is a digital camera used by the worker at the civil engineering construction site to photograph the laying state of the conduit 20 and the joint 21. For example, as shown in FIG. 2, a handle 11a is connected to the photographing device 11, and the worker at the construction site can photograph the conduit 20 disposed in the installation groove 30a dug up by the civil engineering construction from above while standing on the ground 30. The photographing is performed after the conduits 20 and the joints 21 are laid in the installation groove 30a and before they are embedded, and about 200 digital images are photographed in one day's construction work.

[0033] The imaging device 11 transmits the captured digital image to the construction report document creation device 10 via the communication line 19. In this embodiment, it is assumed that the operator himself / herself uses the imaging device 11 for imaging. However, the imaging device 11 may be mounted on a remotely controllable unmanned aircraft such as a drone and used for imaging from above. Further, the operator may hold the imaging device 11 in his / her hand for imaging. In this case, the imaging position is lower than when using the handle 11a or a drone, and the imaging range per captured image becomes narrower. Therefore, the number of captured images is larger than when using the handle 11a or a drone.

[0034] Here, before the operator at the site performs imaging with the imaging device 11, the necessary two-dimensional codes 40A to 40F according to the work content of the civil engineering work are arranged on a predetermined location or on the settlement board 23 placed on the ground 30, and the laying states of the conduits 20 and joints 21 including the two-dimensional codes 40A to 40F are imaged. Note that the two-dimensional codes 40A to 40F are printed on paper or the like and can be carried. Also, the settlement board 23 is, for example, a portable blackboard or whiteboard. Note that it is desirable that the surface of the settlement board 23 be of a single color that is not used for the two-dimensional codes 40A to 40F. This is for improving the recognition rate of the two-dimensional codes 40A to 40F. Also, by arranging the two-dimensional codes 40A to 40F together on the settlement board 23, the recognition of the two-dimensional codes 40A to 40F can be performed intensively, leading to a reduction in recognition time and an improvement in recognition rate. Further, since there is a high probability that the same work will be done over several days in civil engineering work (that is, there is a high probability that the necessary two-dimensional codes 40A to 40F will be the same over several days), once the two-dimensional codes 40A to 40F are arranged on the settlement board 23, there is no need to prepare them later.

[0035] The two-dimensional codes 40A to 40F are, for example, the two-dimensional codes disclosed in JP-A-2020-160626. Specifically, as shown in FIG. 4(a), the two-dimensional codes 40A to 40F are formed in a concave character shape having a corner portion 402 and a bottom portion 403 by two-dimensionally arranging five square cells 401 and providing a notch portion 404. The cell 401 is composed of a color selected from a predetermined four or five colors, and the two-dimensional codes 40A to 40F are associated with information related to civil engineering work, that is, information necessary for the construction report material creating device 10 to create a construction report material, by a combination of colors.

[0036] The information associated with the two-dimensional codes 40A to 40F is specified as follows. When the recognition unit 162 described later recognizes the two-dimensional code 40, the colors of the respective cells 401 are read in the order shown by the arrow Y in FIG. 4(a). It should be noted that the respective cells 401 may be recognized in the order opposite to the direction shown by the arrow Y. As shown in FIG. 4(b), the number corresponding to red is 1, the number corresponding to blue is 2, and the number corresponding to green is 3 are predetermined. For example, since the two-dimensional code 40 shown in FIG. 4(a) is arranged in the order of blue, red, red, red, green, when the numbers corresponding to the respective colors are applied, it represents a sequence of numbers 21113. This sequence of numbers enables the information associated with the two-dimensional code to be specified. Further, a label 41 on which a sequence of numbers represented by an arrangement of colors is described is provided at the notch portion 404 of the two-dimensional codes 40A to 40F.

[0037] Here, the information (association information) associated with the two-dimensional codes 40A to 40F will be described. The association information is classified into six categories, for example, "materials", "work", "site conditions", "work scope", "workers and work vehicles", and "time and date".

[0038] The classification of "materials" includes, for example, information related to buried objects (buried object information). The buried object information refers to the "model number", "type", "name", "material", "size", "length", "price", etc. of the conduit 20 and the joint 21. The "model number" is an individual number for identifying the conduit 20 and the joint 21. The "type" indicates whether the buried object is the conduit 20 or the joint 21. The "name" represents the type of the conduit 20 and the joint 21, such as a PE straight pipe or a welded steel straight pipe. The "material" represents the material that constitutes the conduit 20 and the joint 21, such as polyethylene or steel. The "size" represents the diameter of the conduit 20 and the joint 21, such as 75A or 200A. The "length" represents the length of the conduit 20 and the joint 21. The "price" represents the price of the conduit 20 and the joint 21.

[0039] Therefore, the on-site workers place the two-dimensional code 40A associated with the buried object information of the laid conduit 20 and joint 21 at the site. In this embodiment, as shown in FIG. 3, on the surface of the conduit 20, the two-dimensional code 40A associated with the buried object information of the conduit 20 is pasted, and on the surface of the joint 21, the two-dimensional code 40A associated with the buried object information of the joint 21 is pasted and arranged in this way. Note that the reason for pasting the two-dimensional code 40A on the buried object is to enable the position of the buried object to be specified by acquiring the absolute coordinates of the two-dimensional code 40A by the positioning device 12, as will be described later.

[0040] The classification of "work" includes, for example, information related to the work content of civil engineering work, information related to the work performed on buried objects, and information related to the results of that work. The work content of civil engineering work refers to a new pipe connection work for connecting newly buried conduits to each other, an existing pipe connection work for connecting an existing conduit and a newly buried conduit, and the like. The work performed on buried objects refers to welding, processing, airtightness testing, etc. The airtightness testing is a test for confirming whether airtightness is maintained so that no gas leakage occurs at the connected parts of the conduits 20 (the parts connected by the joint 21).

[0041] Therefore, the on-site worker places the work performed on the buried object and the 2D code 40A associated with the work and its result near the location where the work was carried out, according to the work performed on the buried object and the result of that work. In the present embodiment, since the work of connecting the existing pipe 22A and the conduit 20A by the joint 21A is being carried out, the 2D code 40B associated with the information "existing pipe connection work" is placed on the surface or near the joint 21A. Furthermore, since the work of connecting the existing pipe 22B and the conduit 20B by the joint 21B is being carried out, the 2D code 40B associated with the information "existing pipe connection work" is placed on the surface or near the joint 21B. Furthermore, since the work of connecting the conduit 20A and the conduit 20B by the joint 21C is being carried out, the 2D code 40B associated with the information "new pipe connection work" is placed on the surface or near the joint 21C. Furthermore, assuming that an airtightness test was performed at the joint 21C portion, the 2D code 40B associated with the information "airtightness test" and the information regarding the pass or fail of the test is placed on the surface or near the joint 21C where the test was performed. Note that the reason for placing the 2D code 40B on the surface or near the joint 21 is to enable the identification of the location where the work was carried out by acquiring the absolute coordinates of the 2D code 40B with the positioning device 12, as will be described later.

[0042] The classification of "site situation" includes, for example, information related to location (such as downtown area, suburbs, arterial road, backlot, water area, etc.). Therefore, the on-site worker places the 2D code 40C associated with that location on the site according to the location where the civil engineering work was carried out. For example, in the present embodiment, assuming that civil engineering work was carried out along an arterial road, the 2D code 40C associated with the information "arterial road" is placed on the settlement board 23. Note that instead of placing the 2D code 40C on the site, it is also possible to acquire the information related to the location by the positioning device 12.

[0043] The classification of "working range" includes, for example, information related to backfilling of the installation trench 30a and restoration work such as paving. Therefore, the on-site workers place two-dimensional codes associated with the information of "backfilling" and "paving" on-site according to the range of backfilling and the range of paving. For example, in this embodiment, assuming backfilling and paving of the installation trench 30a, two-dimensional codes 40D associated with the information of "backfilling and paving" are placed at the four corners of the installation trench 30a. Placing the two-dimensional codes 40D at the four corners of the installation trench 30a is to enable identification of the range where backfilling, restoration work such as paving, etc. have been carried out by obtaining the absolute coordinates of the two-dimensional codes 40D with the positioning device 12, as will be described later. Note that the range where backfilling, restoration work such as paving, etc. have been carried out may also be determined using the elevation terrain data described later.

[0044] The classification of "workers and work vehicles" includes, for example, information related to the occupations of the workers (supervision, construction workers, security guards), information related to the types of work vehicles (dump trucks, construction vehicles, etc.), and information related to the tools used in the work.

[0045] Therefore, the on-site workers place two-dimensional codes 40E associated with the occupations of the workers and the types of work vehicles according to the occupations of the workers who have carried out the work on-site and the types of work vehicles and tools used on-site. In this embodiment, assuming that one supervisor and one construction worker have carried out work on-site and one dump truck has been used, a total of three two-dimensional codes 40E (that is, the two-dimensional code 40E associated with the information of "supervision", the two-dimensional code 40E associated with the information of "construction worker", and the two-dimensional code 40E associated with the information of "dump truck") are placed on-site. If multiple construction workers have carried out the work, the corresponding number of two-dimensional codes 40E associated with the information of "construction worker" are installed according to the number of workers. In addition, when expensive tools are used, in order to claim opportunity loss fees, the two-dimensional codes associated with those tools may be placed on-site together with the tools.

[0046] The classification of "time and season" includes, for example, information regarding the time zone when civil engineering work was carried out (daytime, late at night, holidays, etc.) and information regarding seasons (dry season, Obon, New Year period, winter, etc.).

[0047] Therefore, the on-site worker arranges the two-dimensional code 40F associated with the information of that time zone and season at the site according to the time zone and season when the work was carried out at the site. For example, in this embodiment, assuming that civil engineering work was carried out late at night, the two-dimensional code 40F associated with the information of "late at night" is arranged on the settlement board 23. Instead of arranging the two-dimensional code 40F at the site, it is also possible to acquire time information by the positioning device 12.

[0048] In addition, unlike the two-dimensional codes 40A - 40D, the two-dimensional code 40E associated with the information belonging to the classification of "worker and work vehicle" and the two-dimensional code 40F associated with the information belonging to the classification of "time and season" do not require position identification by the positioning device 12, and thus are collectively arranged on the settlement board 23.

[0049] Returning to the description of the voucher processing system 1, the positioning device 12 is a device capable of acquiring the absolute coordinates of the positioning target by using the positioning technology based on GPS.

[0050] On-site workers use the positioning device 12 to obtain at least three absolute coordinates from the two-dimensional codes 40A, 40B, 40C, 40D attached to the conduit 20 and the joint 21, the feature points at the construction site, and the ground control points (GCP) after the laying of the conduit 20 and the joint 21. Thereby, information about the absolute coordinates can be provided in drawings such as three-dimensional point cloud data generated by the drafting unit 161 described later. The positions of the two-dimensional codes 40A, 40B, 40C, 40D may be directly measured, but can also be obtained based on three-dimensional point cloud data or the like. In the case of the feature points at the construction site and the ground control points (GCP), if there are about six points at 20 m intervals, three-dimensional point cloud data or the like can be generated. Therefore, when the number of the two-dimensional codes 40A, 40B, 40C, 40D is more than that, the number of operations for measuring the feature points at the construction site and the GCP is less than that for measuring the two-dimensional codes 40A, 40B, 40C, 40D. Note that as the feature points at the construction site, cracks in the pavement, pavement joints, pavement lines, patterns of manholes, corners of L-shaped gutters, corners of apron blocks, corners of iron covers for gutters, public-private boundary piles, block reference points, etc. are used. In addition, depending on the absolute coordinates of the two-dimensional code 40A, the embedding positions of the conduit 20 and the joint 21 can be specified. Also, when welding, processing, or airtight testing is being performed on the embedded object, the position where the welding, processing, or airtight testing was performed can be specified by the absolute coordinates of the two-dimensional code 40B arranged in that case. Further, when there is a trouble such as water leakage L occurring in the installation groove 30a, the position where the trouble occurred can be specified by the absolute coordinates of the two-dimensional code 40C arranged in that case. Also, the position of the installation groove 30a can be specified by the absolute coordinates of the two-dimensional codes 40D arranged at the four corners of the installation groove 30a. Then, the positioning device 12 can transmit the information about the obtained absolute coordinates to the construction report document creation device 10 via the communication line 19.

[0051] The pre-input device 18 is a device for selecting the type of construction work, inputting permit information, adding missing materials, and inputting information such as the site situation map (pre-input information). Selecting the type of construction work enables mode selection according to the purpose of civil engineering work, such as water supply, sewerage, gas, optical cable, etc. The format of the construction report document R and the items described therein are changed according to the selected mode. Inputting permit information means inputting the unit price ordered and contracted for each permit. Since the number of orderers and construction companies is one-to-many in the civil engineering industry, it is smooth for the orderer to create and distribute the data (unit price information) or register it himself. Adding missing materials means adding materials, personnel, work, etc. that are not available in each mode. The timing of addition is at the time of inputting the unit price of the above-mentioned order contract information. Registering the site situation map means registering a downtown map, a suburban map, a trunk road map, a waterworks map, etc., and is used when not using the two-dimensional code 40C associated with the information belonging to the category of "site situation".

[0052] Also, according to the situation at the construction site, if there is an increase in unit price or an increase in cost, the amount is specified by the pre-input device 18. The case of an increase in unit price or an increase in cost is, for example, a case where the number of work interruptions increases due to congestion of buried objects or heavy traffic in the downtown area, resulting in a decrease in construction efficiency. In addition, when performing construction work in the suburbs, it takes time to move and the working hours are reduced. Trunk roads are constructed with strong materials for the road structure and take time to excavate, so it is conceivable that the construction efficiency will decrease. Furthermore, when performing construction work at a waterworks, the bottom of the excavation cannot be seen and it is difficult to dig up the soil, and soil and water need to be scooped up, so it is conceivable that the construction efficiency will decrease.

[0053] Also, according to the time or period when the civil engineering work is carried out, if there is an increase in unit price or an increase in cost, the amount is specified by the pre-input device 18. For example, it is considered that there is an increase in unit price or an increase in cost when it is difficult to arrange workers, such as late at night during the New Year holidays.

[0054] In addition, the pre-input device 18 can also associate new information with the two-dimensional code. The two-dimensional codes 40A - 40F described above are general-purpose products with pre-associated information. However, when an operator performs irregular work, the general-purpose two-dimensional codes 40A - 40F cannot be used. For example, there may be a problem such as water leakage L occurring in the installation groove 30a, and assuming that the corresponding work is carried out, a case where a two-dimensional code associated with the work content has not been prepared in advance can be considered. In such a case, the pre-input device 18 is used to associate information such as "water leakage countermeasure work" with a new two-dimensional code 40G. The content associated by the pre-input device 18 is transmitted to the construction report document creation device 10 via the communication line 19. Then, the operator arranges the two-dimensional code 40G associated with the information "water leakage countermeasure work" adjacent to the water leakage L. Note that arranging the two-dimensional code 40G adjacent to the water leakage L is to enable the positioning device 12 to obtain the absolute coordinates of the two-dimensional code 40C, thereby enabling the position of the water leakage L and the like to be specified, as will be described later.

[0055] The dimension correction member 24 is, for example, a measuring instrument such as a surveying gauge or a tape measure. When the positioning situation by the positioning device 12 is poor, the size represented in the three-dimensional drawing such as the three-dimensional point cloud data becomes inaccurate. In such a case, the dimension correction member 24 is used to correct the dimensions and improve the defect (details will be described later). In addition to the above-described measuring instruments, as the dimension correction member 24, an L-shaped side groove, a cast iron lid for the side groove, a manhole, etc. that can specify the size may also be used.

[0056] The construction report document creation device 10 includes a communication unit 13, a registration unit 14, a database 15, and a processing unit 16.

[0057] The communication unit 13 receives the digital image transmitted from the imaging device 11, the information regarding the absolute coordinates transmitted from the positioning device 12, and the information associated with the newly created two-dimensional code transmitted from the pre-input device 18. Then, the information received by the communication unit 13 is registered in the database 15 by the registration unit 14.

[0058] In the database 15, in addition to the information on the registered digital images and absolute coordinates, information belonging to the six categories of "materials", "work", "site conditions", "work scope", "workers and work vehicles", and "time and date" as already described is registered in advance. In addition, unit price information corresponding to the information belonging to the above six categories is also registered in advance in the database 15. The unit price information is, for example, the cost per meter of the conduit 20 specified by the embedded object information such as "model number" belonging to the "materials" category, the cost per time of the "airtightness test" belonging to the "work" category, the cost per person of the "construction worker" belonging to the "workers and work vehicles" category, and the like. Also, the information input by the pre-input device 18 is registered in the database 15.

[0059] The processing unit 16 includes a drawing unit 161, a recognition unit 162, an analysis unit 165, a reading unit 163, a document generation unit 164, and a collation unit 166.

[0060] The drawing unit 161 generates an as-built drawing (an example of a composite image) by synthesizing a plurality of digital images registered in the database 15. The as-built drawing is a drawing showing the laying state of the conduit 20 and the joint 21, and refers to, for example, three-dimensional drawings such as three-dimensional point cloud data, three-dimensional mesh data, three-dimensional CAD drawings, elevation terrain data, two-dimensional drawings such as ortho images, two-dimensional CAD drawings, and vector data serving as the basis for three-dimensional CAD drawings and two-dimensional CAD drawings.

[0061] The drawing unit 161 generates drawings as follows, for example. First, three-dimensional point cloud data is generated based on the digital image. The three-dimensional point cloud data is a drawing of the three-dimensional images of the conduit 20 and the joint 21 by a set of points, and can represent the laying state of the conduit 20 and the joint 21 with high precision.

[0062] Furthermore, the drafting unit 161 can generate 3D mesh data with a smaller data size based on the 3D point cloud data. The 3D mesh data is obtained by meshing the 3D point cloud data and converting it into polygon data, depicting the laying states of the conduits 20 and joints 21.

[0063] Furthermore, the drafting unit 161 can generate an orthoimage based on the 3D point cloud data. The orthoimage is an image depicting the laying states of the conduits 20 and joints 21 by orthographic projection.

[0064] Furthermore, the drafting unit 161 can generate vector data based on any one of the 3D point cloud data, 3D mesh data, and orthoimage. The vector data is a drawing that simply represents the laying states of the conduits 20 and joints 21 by representing the conduits 20 and joints 21 with point data, line data, surface data, and volume data. It is possible to include the buried object information as an attribute of the drawn lines and points in the vector data.

[0065] Furthermore, the drafting unit 161 can generate 3D CAD drawings and 2D CAD drawings based on the vector data.

[0066] The 3D CAD drawing is a drawing depicting 3D images of the conduits 20 and joints 21 using 3D functions. It is possible to include the buried object information of the conduits 20 and joints 21 included in the 2D code 40 as an attribute of the conduits 20 and joints 21 represented on the drawing.

[0067] In addition, the 2D CAD drawing refers to a plan view, cross-sectional view, side view, etc. that represent the laying state of the conduit 20 and the joint 21. In the 2D CAD drawing, the embedding information of the conduit 20 and the joint 21 included in the 2D code 40 can be held as the attributes of the conduit 20 and the joint 21 represented on the drawing. Also, since a plurality of layers can be provided in the 2D CAD drawing, for example, a layer that only illustrates the laying state of the conduit 20 and the joint 21, a layer that includes images of the 2D codes 40A - 40F arranged on site, a layer that includes information of the work report W and the settlement statement S, etc. can be provided.

[0068] In addition, the drafting unit 161 can hold information regarding absolute coordinates, acquired by the positioning device 12, in the 3D point cloud data, 3D mesh data, ortho image, 3D CAD drawing, and 2D CAD drawing described above.

[0069] The as-built drawing created by the drafting unit 161 as described above can be used as the drawing D of the construction report material R. For example, in the example shown in FIG. 9, the vector data created by the drafting unit 161 is used. In addition, 3D drawings or 2D drawings created by the drafting unit 161 may be used. Also, it is not always necessary to use the as-built drawing, and the construction drawing used for civil engineering work may be used as the drawing D of the construction report material R.

[0070] Next, the recognition unit 162 recognizes the 2D codes 40A - 40G (document vouchers in the image) in the as-built drawing (for example, a layer that includes images of the 2D codes 40A - 40G of the 2D CAD drawing or 3D point cloud data) drafted by the drafting unit 161 based on the digital image. For example, in the present embodiment, the recognition unit 162 recognizes the 2D code 40A (see FIG. 3) attached to the conduit 20 and the joint 21, the 2D codes 40B, 40B (see FIG. 3) near the joints 21A, 21B, the 2D codes 40C, 40E, 40F (see FIG. 3) on the settlement board 23, and the 2D code 40D (see FIG. 3) arranged at the four corners of the installation groove 30a.

[0071] The recognition of the two-dimensional codes 40A - 40F is performed as follows based on a recognition program. Fig. 5 is a flowchart of the recognition program. Fig. 6 is a diagram for explaining the recognition of the centroid distance (step S7) and the centroid angle (step S8) in the recognition program. Fig. 7 is a diagram for explaining a modification example of the recognition of the centroid angle (step S8) in the recognition program.

[0072] First, color recognition is performed (step S1). This is a step for extracting the colors used in the cells from the entire as-built drawing. Extracting the colors used in the cells from the entire as-built drawing means extracting what might be cells from the entire as-built drawing. Specifically, for each color to be extracted from the as-built drawing, ranges of hue, lightness, and saturation are set. For example, if it is red, the hue H (maximum 360) is 330 - 360, 0 - 10, the lightness S (maximum 100) is 50 - 100, and the saturation B (maximum 100) is 70 - 100. Then, the colors within the defined range are extracted from the as-built drawing (the extracted part is called the extraction part). Colors outside the above range are not extracted at this point, assuming they are not cells.

[0073] Next, rectangular shape recognition is performed (step S2). That is, it is determined whether the extraction part is a rectangular shape. This is because if the extraction part is not a rectangular shape, it is highly likely not to be a cell constituting the two-dimensional code. Specifically, a cell with a ratio of the area of the cell to the area when the cell is surrounded by the smallest rectangle of 0.8 or more, and a cell with its contour extracted and having 4 corners in a convex shape are determined to be rectangular shapes. Parts of the extraction part that are not recognized as rectangular shapes are excluded at this point, assuming they are not cells. This is because depending on the color recognition (step S1), things other than the cells shown in the as-built drawing may accidentally fall within the set ranges of hue, lightness, and saturation and be extracted. On the other hand, among the extraction parts, the parts determined to be rectangular shapes (referred to as rectangular parts) are parts that may be cells. Note that the as-built drawing is obtained by synthesizing digital images, and since the image is distorted during synthesis, the rectangular parts are likely to be distorted, not exact squares, but rectangles, parallelograms, etc.

[0074] Next, perform the same-color division of the square part (step S3). When cells of the same color are adjacent and arranged in the two-dimensional code, individual cells cannot be recognized. Therefore, this is a step to enable recognition of individual cells by performing division. For example, recognize the longitudinal direction of the portion determined to be square-shaped (square part). If the length in the longitudinal direction is 1.7 to 2.3 times the length of two cells, divide it into two parts. If the length in the longitudinal direction is 2.5 to 3.5 times the length of three cells, divide it into three parts. A portion whose length in the longitudinal direction is shorter than the length of two cells is not divided. Note that the square part after division is called the divided square part.

[0075] Next, correct the square part and the divided square part in the as-built drawing to a square (step S4). This is because during the process of synthesizing digital images, the image may be distorted, and the shape of cells that are originally square is likely to be distorted. For the correction, first, overlay a square with the normal side length on the centroid of the square part or the divided square part. When it overlaps with other adjacent square parts or divided square parts, move the square in the direction where the overlap disappears. Also, when the two-dimensional code is pasted in a cylindrical shape, a part of the square is missing, resulting in a rectangular square part or divided square part. For the correction, correct the missing part considering the concave shape. The correction is performed for the square part or the divided square part adjacent to multiple square parts or divided square parts when the length of the non-missing side is within 0.7 to 1.3, or when the area of the square part or the divided square part is 0.8 times or more of the area of the square. When the square part or the divided square part is alone or outside the range of the side length, it is excluded as not being a two-dimensional code at this point. Note that the square part or the divided square part after correction is called the corrected part.

[0076] Next, recognize the centroid of the corrected part (step S5). In this embodiment, since the corrected part is a square, the centroid is located at the center of the corrected part.

[0077] Next, perform concave shape recognition (step S6). Specifically, a set of 5 adjacent correction parts in the as-built drawing is regarded as one set, and the number of corners of the set of correction parts is calculated. In this embodiment, since the two-dimensional code is concave, those with a number of corners other than 8 are excluded as not being two-dimensional codes. Further, for the set of the upper 2 correction parts and the lower 3 correction parts, a set in which the upper 2 correction parts are separated is determined to be a concave shape. At this point, those that are not concave are excluded as not being two-dimensional codes. For those determined to have a concave arrangement of correction parts, the orientation of the concave shape is recognized based on the position of the notch part, and the numerical sequence represented by the color constituting the two-dimensional code is recognized.

[0078] Next, for those determined to have a concave set of correction parts in step S6, calculate the distance between the centroids (step S7). Specifically, for example, as shown in FIG. 6(a), with the correction part 503 located below the notch part 506 (corresponding to the notch part 404) of the set 50A of the correction parts 501-505, from the centroid G3 of the correction part 503, calculate the distances to the centroids G1, G2, G4, G5 of the respective correction parts 501, 502, 504, 505 (that is, the lengths of the respective line segments L1-L4 connecting from the centroid G3 to the centroids G1, G2, G4, G5). Then, a predetermined threshold value is set in advance (for example, L1 and L4 are 1.2 times the length of one side of the cell, and L2 and L3 are √2×1.2 times the length of one side of the cell), and it is determined whether the calculated distance is within the threshold value. As in the set 50A of the correction parts 501-505 shown in FIG. 6(a), if the distortion is small, it will be within the threshold value. On the other hand, as in the set 50B of the correction parts 501-505 shown in FIG. 6(b), if there is distortion, the distances between the centroids G1-G5 (the lengths of the line segments L11-L14) will be large and may exceed the threshold value. The set of correction parts that exceeds the threshold value is excluded as not being a two-dimensional code at this point.

[0079] Instead of calculating the lengths of the line segments L1 - L4 (L11 - L14) connecting the centers of gravity, as shown in Fig. 7(a), it is also possible to obtain the line segments L21 - L24 (L31 - L34) connecting adjacent centers of gravity G1 - G5 and calculate the lengths of the line segments. Then, a threshold value of 1.2 times the length of one side of the cell is set, and it is determined whether the calculated distance is within the threshold value. If the distortion is small, as in the set 50A of the correction parts 501 - 505 shown in Fig. 7(a), it will be within the threshold value. On the other hand, as in the set 50B of the correction parts 501 - 505 shown in Fig. 7(b), if there is distortion, the distance between the centers of gravity G1 - G5 (the lengths of the line segments L31 - L34) will increase, and there is a possibility of exceeding the threshold value. The set of correction parts that exceed the threshold value is excluded as not being a two-dimensional code at this point.

[0080] Next, for the set of correction parts for which it is determined that the distance between the centers of gravity is within the threshold value, the center-of-gravity angle is calculated (step S8). Specifically, for example, as shown in Fig. 6(a), the center line CL11 passing through the center of gravity G3 of the correction part 503 and the line segments L1 - L4 connecting the centers of gravity G1 - G5 are obtained, and the angles A1 - A4 of the line segments L1 - L4 with respect to the center line CL11 are calculated. Then, a predetermined threshold value is set in advance (for example, angles A1, A4 are 45° ± 5°, angles A2, A3 are 90° ± 5°), and it is determined whether the calculated angles A1 - A4 are within the threshold value. If the distortion is small, as in the set 50A of the correction parts 501 - 505 shown in Fig. 6(a), it will be within the threshold value. On the other hand, as in the set 50B of the correction parts 511 - 515 shown in Fig. 6(b), if there is distortion, the angles A11 - A14 of the line segments L11 - L15 with respect to the center line CL11 will be smaller or larger than when there is no distortion, and there is a possibility of exceeding the threshold value. The set of correction parts that exceed the threshold value is excluded as not being a two-dimensional code at this point.

[0081] Instead of calculating the angles A1 - A14 (A1 - A14) of the line segments L1 - L4 (L11 - L14) with respect to the center line CL11, as shown in FIG. 7(a), line segments L21 - L24 connecting adjacent centroids G1 - G5 may be obtained, and the angles A21 - A23 between the line segments L21 - L24 may be calculated. Then, it is determined whether the angle is within a threshold value. If the distortion is small, as in the set 50A of the correction units 501 - 505 shown in FIG. 7(a), it will be within the threshold value. On the other hand, as in the set 50C of the correction units 501 - 505 shown in FIG. 7(b), if there is distortion, the angles A31 - A33 between the line segments L21 - L24 will be smaller or larger than when there is no distortion, and there is a possibility of exceeding the threshold value. The set of correction units that exceeds the threshold value is excluded as not being a two-dimensional code at this point.

[0082] Next, label recognition is performed (step S9). Specifically, for example, it is determined whether there is a label 41 (see FIG. 4(a)) in the notch 506 of the set 50 of the correction units 501 - 505. Specifically, when the number of sets of black dots is 6 or more, it is determined as a label. The set 50 of the correction units 501 - 505 for which it is determined that there is a label 41 is determined to be a two-dimensional code, and the set of correction units for which it is determined that there is no label is excluded as not being a two-dimensional code at this point. Here, only the presence or absence of the label is determined, and the content written is not read. Instead of recognizing the set of black dots, character recognition or number recognition may also be used.

[0083] Through the above steps, for those determined to be two-dimensional codes, a reliability is calculated as an index of whether the recognition of the two-dimensional code is accurate (step S10). The calculation of the reliability is, for example, calculated by a point deduction method with a full score of 100 points. If the threshold values in steps S1 - S4, S6 - S9 are strict, the amount of point deduction for the reliability is small, while if the threshold values are loose, the amount of point deduction for the reliability is large.

[0084] After calculating the reliability, reset the thresholds in steps S1 - S4, S7 - S9 (step S11), and then repeat steps S1 - S10. In step S11, at least expand the thresholds of steps S1 - S4. This is because the thresholds in the first - week steps S1 - S4 are strict, and there is a possibility of overlooking what is actually a cell. In addition, if there is something that was excluded as not being a two - dimensional code in steps S7 - S9, expand the threshold of the step where the exclusion was made. This is because the thresholds in the first - week steps S6 - S9 are strict, and there is a possibility that what is actually a two - dimensional code has been excluded due to image distortion. Note that the point - deduction method when repeating is to deduct the total points accumulated from the first week to the last week.

[0085] When the cells overlooked in steps S1 - S4 are correctly recognized, even with the same threshold, they may be determined as two - dimensional codes in steps S6 - S9. Therefore, devise the method for expanding the threshold, the steps to repeat in the second week, the number of repetitions, and the point - deduction method for reliability. For example, only reset the thresholds of steps S1 - S4 and repeat steps S1 - S10. In this case, the steps that are not reset do not deduct reliability points. On the other hand, when resetting the thresholds of steps S1 - S4, S7 - S9 and repeating steps S1 - S10, since the thresholds of S7 - S9 are lowered too much compared to S1 - S4, reduce the amount of threshold expansion and reduce the number of points deducted for steps S7 - S9. Also, in step S1, there are thresholds for hue, brightness, and saturation for each color, resulting in multiple thresholds. Therefore, when only resetting one threshold for one color in step S1 and repeating steps S1 - S10, the number of repetitions is the largest. So, set the amount of reliability deduction for steps S2 - S4 relative to step S1 to be moderate, and further reduce the amount of reliability deduction for steps S7 - S9. With this device, it becomes possible to deduct reliability according to the change in the threshold and the number of repetitions.

[0086] For example, the ranges of hue, lightness, and saturation in step S1 (color recognition) are expanded by 20% every time the recognition program makes one full cycle. Each time it is expanded, the reliability calculated in step S10 is deducted by 10% of the full score (for example, if the full score is 100 points, it is deducted by 10 points). The threshold value in step S2 (quadrilateral recognition) is expanded by 10% every time the recognition program makes one full cycle. Each time it is expanded, the reliability calculated in step S10 is deducted by 5% of the full score (for example, if the full score is 100 points, it is deducted by 5 points). The threshold value in step S4 (square correction) is expanded by 10% every time the recognition program makes one full cycle. Each time it is expanded, the reliability calculated in step S10 is deducted by 5% of the full score (for example, if the full score is 100 points, it is deducted by 5 points). The threshold value in step S7 (center-of-gravity distance recognition) is expanded by 3% every time the recognition program makes one full cycle. Each time it is expanded, the reliability calculated in step S10 is deducted by 1% of the full score (for example, if the full score is 100 points, it is deducted by 1 point). The threshold value in step S8 (center-of-gravity angle recognition) is expanded by 3% every time the recognition program makes one full cycle. Each time it is expanded, the reliability calculated in step S10 is deducted by 1% of the full score (for example, if the full score is 100 points, it is deducted by 1 point). The threshold value in step S9 (label recognition) is expanded by 3% every time the recognition program makes one full cycle. Each time it is expanded, the reliability calculated in step S10 is deducted by 1% of the full score (for example, if the full score is 100 points, it is deducted by 1 point).

[0087] However, when the threshold value is reset, if the reliability calculated in step S10 becomes 50 or less (step S12: YES), the repetition of steps S1 - S10 is not performed, and the recognition program ends. In this case, a record with a reliability of 50 or less is recorded in the database 15. Regardless of the reliability value, if the operator modifies the sequence of the two-dimensional code, the modification history is recorded in the database 15. The drafting unit 161 combines and records the reliability recorded in the database 15 in the two-dimensional code in the ortho image shown in Figure ○○ (Figure 3 with the reliability added). Also, it can be held as an attribute of the conduit 20 and the joint 21 represented on the three-dimensional CAD drawing and the two-dimensional CAD drawing, and can also be combined and recorded.

[0088] Next, the analysis unit 165 analyzes the information acquired by the on-site device 2 and the two-dimensional codes 40A - 40F recognized by the recognition unit 162 (recognition program) based on the analysis program. FIG. 8 is a flowchart of the analysis program.

[0089] First, it is determined whether the positioning status by the positioning device 12 is good or not (step S21). The quality of the positioning status is determined by the speed of the positioning device 12 and the FIX solution. For example, when the speed is 0.01 m / s or less and a FIX solution is obtained, the positioning status is determined to be good (step S21: good). On the other hand, when it is determined that the positioning status is poor (step S21: no), this is recorded in the database 15 (step S22).

[0090] Next, it is determined whether correction of the three-dimensional drawing is necessary (step S23). If in step S21 it is determined that the positioning status is poor, it is determined that correction is necessary (step S23: necessary), and correction of the three-dimensional drawing is executed (step S24). Specifically, the dimensions of the dimensional correction member 24 in the three-dimensional drawing are measured, and if they are different from the actual dimensions, the size in the three-dimensional drawing is corrected by expanding or shrinking it to the actual dimensions. On the other hand, if in step S21 it is determined that the positioning status is good, it is determined that correction is not necessary (step S23: no).

[0091] Next, it is confirmed whether the absolute coordinates of the two-dimensional codes 40A, 40B, 40C, 40D have been acquired by the positioning device 12 (step S25). By checking the information registered in the database 15, if information about the absolute coordinates of the two-dimensional codes 40A, 40B, 40C, 40D acquired by the positioning device 12 at the civil engineering site is registered, it is determined that the absolute coordinates have been acquired (step S25: yes). On the other hand, if information about the absolute coordinates of the two-dimensional codes 40A, 40B, 40C, 40D is not registered in the database 15, it is determined that the absolute coordinates have not been acquired (step S25: no), and based on the three-dimensional drawing (for example, three-dimensional point cloud data), the coordinates of the two-dimensional codes 40A, 40B, 40C, 40D are calculated (step S26).

[0092] Next, it is determined whether azimuth calculation is necessary (step S27). In step S21, if it is determined that the positioning status is poor, it is determined that azimuth calculation is necessary (step S27: necessary), and azimuth calculation is executed (S28). Specifically, the azimuth is calculated based on the shadow shown in the 3D drawing, the time and position information obtained by the positioning device 12. On the other hand, in step S21 in FIG. 8, if it is determined that the positioning status is good, it is determined that azimuth calculation is unnecessary (step S27: no).

[0093] Next, it is determined whether acquisition of time and date information is necessary (step S29). When the recognition unit 162 recognizes the two-dimensional code 40F associated with the information belonging to the category of "time and date" from the drawing (for example, a layer including images of two-dimensional codes 40A - 40F of a 2D CAD drawing, 3D point cloud data, etc.) drawn by the drawing unit 161 based on the digital image, it is determined that acquisition of time and date information is unnecessary (step S29: no). In this embodiment, the recognition unit 162 recognizes the two-dimensional code 40F on the settlement board 23 (see FIG. 3) and determines that acquisition of time and date information is unnecessary. On the other hand, when the recognition unit 162 fails to recognize the two-dimensional code 40F for reasons such as the two-dimensional code 40F not being used at the site, it is determined that acquisition of time and date information is necessary (step S29: necessary), and the information on the year, month, day, and time is acquired by the positioning device 12 (GPS) (step S30).

[0094] Next, it is determined whether it is necessary to acquire information regarding the on-site situation (step S31). When the recognition unit 162 recognizes the two-dimensional code 40C associated with the information belonging to the category of "on-site situation" from the drawing (for example, a layer including images of the two-dimensional codes 40A - 40F of a 2D CAD drawing, 3D point cloud data, etc.) drawn by the drawing unit 161 based on the digital image, it is determined that it is not necessary to acquire information regarding the on-site situation (step S31: No). In the present embodiment, the recognition unit 162 recognizes the two-dimensional code 40C (see FIG. 3) on the settlement board 23 and determines that it is not necessary to acquire information regarding the on-site situation. On the other hand, when the recognition unit 162 fails to recognize the two-dimensional code 40C for reasons such as the two-dimensional code 40C not being used on-site, it is determined that it is necessary to acquire information regarding the on-site situation (step S31: Yes), and position information is acquired by the positioning device 12 (GPS) (step S32).

[0095] Next, it is determined whether it is necessary to acquire information regarding the work range (step S33). When the recognition unit 162 recognizes the two-dimensional code 40D associated with the information belonging to the category of "work range" from the drawing (for example, a layer including images of the two-dimensional codes 40A - 40F of a 2D CAD drawing, 3D point cloud data, etc.) drawn by the drawing unit 161 based on the digital image, it is determined that it is not necessary to acquire information regarding the work range (step S33: No). In the present embodiment, the recognition unit 162 recognizes the two-dimensional codes 40D (see FIG. 3) arranged at the four corners of the installation groove 30a and determines that it is not necessary to acquire information regarding the work range. On the other hand, when the recognition unit 162 fails to recognize the two-dimensional code 40D for reasons such as the two-dimensional code 40D not being used on-site, it is determined that it is necessary to acquire information regarding the work range (step S33: Yes), and the excavation volume and excavation area are calculated by analyzing the objects on the 3D drawing (for example, by analyzing the data in which the height differences of the elevation terrain data are color-coded, the excavation volume and excavation area can be calculated) (step S34). Then, the analysis program ends.

[0096] When the analysis by the analysis unit 165 is completed, the reading unit 163 reads from the database 15 the information associated with the two-dimensional codes 40A - 40F recognized by the recognition unit 162 or the result calculated by the analysis program, and the unit price information corresponding to the information. Then, based on the information and unit price information associated with the two-dimensional codes 40A - 40F read by the reading unit 163, the document generation unit 164 calculates the claim amount and automatically generates the settlement statement S and the work report W. Specifically, it is as follows.

[0097] First, the reading unit 163 reads from the database 15 the information associated with the information belonging to the category of "materials" associated with the two-dimensional code 40A recognized by the recognition unit 162, and the unit price information corresponding to the information. For example, the embedded object information such as "model number", and the unit price information corresponding to the embedded object information are read from the database 15.

[0098] Then, the document generation unit 164 calculates the claim amount as follows from the information read above. For example, if the unit price information of the read conduits 20A and 20B is 1,000 yen / m, and the lengths of the conduits 20A and 20B are respectively 5m (that is, 10m in total) as specified by the embedded object information, the claim amount is calculated as 10,000 yen. This corresponds to the item of "straight pipe 100A" in the settlement statement S. Also, if the unit price information of the read joint 21 is 1,000 yen / piece, and the number of joints 21 is 3 as specified by the number of two-dimensional codes 40A, the claim amount is calculated as 3,000 yen. This corresponds to the item of "socket 100A" in the settlement statement S. Note that the lengths of the conduits 20A and 20B may also be read by analyzing the objects on the three-dimensional drawing.

[0099] Next, the reading unit 163 reads from the database 15 the information associated with the "work" category corresponding to the two-dimensional codes 40B, 40B recognized by the recognition unit 162, and the unit price information corresponding to that information. For example, in the present embodiment, based on the two-dimensional codes 40B, 40B (see FIG. 3) near the joints 21A, 21B, the information of "existing pipe connection work" and the unit price information corresponding to "existing pipe connection work" are read from the database 15. In addition, based on the two-dimensional codes 40B, 40B (see FIG. 3) near the joint 21C, the information of "new pipe connection work" and "airtightness test" is obtained. Then, the unit price information corresponding to each of "new pipe connection work" and "airtightness test" is read from the database 15. Note that since the two-dimensional code 40B associated with the "airtightness test" also has information on whether the airtightness test is passed or failed, information such as "qualified" is also read from the database 15 together.

[0100] And the document generation unit 164 calculates the claim amount from the above-read information as follows. For example, if the unit price information of the "existing pipe connection work" read is 10,000 yen per location, since the number of locations where the "existing pipe connection work" was performed is 2, which is specified by the number of two-dimensional codes 40B, the claim amount is calculated as 20,000 yen. This corresponds to the item of "existing pipe connection work" in the settlement statement S. Also, if the unit price information of the "new pipe connection work" is 10,000 yen per location, since the number of locations where the "new pipe connection work" was performed is 1, which is specified by the number of two-dimensional codes 40B, the claim amount is calculated as 10,000 yen. This corresponds to the item of "new pipe connection work" in the settlement statement S. Furthermore, if the unit price information of the "airtightness test" is 10,000 yen per time, since the "airtightness test" was performed 1 time, which is specified by the number of two-dimensional codes 40B, the claim amount is calculated as 10,000 yen. This corresponds to the item of "airtightness test" in the settlement statement S.

[0101] Next, the reading unit 163 reads from the database 15 the information belonging to the category of "site conditions" associated with the two-dimensional code 40C recognized by the recognition unit 162, and the unit price information corresponding to that information. For example, in the present embodiment, based on the two-dimensional code 40C (see FIG. 3) on the settlement board 23, the information of "arterial road" and the unit price information corresponding to the "arterial road" are read from the database 15. In addition, when instead of arranging the two-dimensional code 40C at the site, information regarding the location is acquired by the positioning device 12, the reading unit 163 reads the corresponding unit price information from the database 15 in comparison with the information regarding the location acquired by the positioning device 12.

[0102] Then, the document generation unit 164 calculates the claim amount from the information read above as follows. For example, if the unit price information of the "arterial road" read is 10,000 yen per case and an invoice S for one case of construction is to be created, the claim amount is calculated as 10,000 yen. This corresponds to the item of "increased charge for arterial road" in the invoice S. In addition, when information regarding the site conditions is acquired by the analysis program (step S24 in FIG. 8), the claim amount is calculated based on the acquired information.

[0103] Next, the reading unit 163 reads from the database 15 the information belonging to the category of "working range" associated with the two-dimensional code 40D recognized by the recognition unit 162, and the unit price information corresponding to that information. For example, in the present embodiment, based on the two-dimensional code 40D (see FIG. 3) arranged at the four corners of the installation groove 30a, the information of "backfilling and paving" and the unit price information corresponding to "backfilling" and "paving" are read from the database 15.

[0104] Then, the document generation unit 164 calculates the claim amount from the information read above as follows. For example, if the unit price information of the "backfilling" read is 10,000 yen / m 3Then, the claim amount is calculated by multiplying the volume for backfilling by this. Note that the volume for backfilling is calculated by the material generation unit 164 based on the absolute coordinates of the two-dimensional code 40D acquired by the positioning device 12 (i.e., the absolute coordinates of the four corners of the installation groove 30a) and the depth of the installation groove 30a measurable on the three-dimensional drawing. For example, if the volume for backfilling is 10 m 3 Then, the claim amount is calculated to be 100,000 yen. This corresponds to the item of "backfilling cost" in the settlement statement S. Also, if the unit price information of the "pavement" read is 3,000 yen / m 2 Then, the claim amount is calculated by multiplying this by the area for pavement. Note that the area for pavement is calculated by the material generation unit 164 based on the absolute coordinates of the two-dimensional code 40D acquired by the positioning device 12 (i.e., the absolute coordinates of the four corners of the installation groove 30a). For example, if the area for pavement is 10 m 2 Then, the claim amount is calculated to be 30,000 yen. This corresponds to the item of "pavement cost" in the settlement statement S. In addition, when the excavation volume and excavation area are calculated based on the three-dimensional drawing by the analysis program (step S26 in FIG. 8), the claim amount is calculated based on the calculated excavation volume and excavation area.

[0105] Next, the reading unit 163 reads from the database 15 the information belonging to the category of "worker and work vehicle" associated with the two-dimensional code 40E recognized by the recognition unit 162, and the unit price information corresponding to that information. For example, in this embodiment, the information of "supervision", "construction worker", "dump" associated with the two-dimensional codes 40E, 40E, 40E on the settlement board 23, and the unit price information corresponding to each of "supervision", "construction worker", "dump" are read from the database 15.

[0106] Then, the document generation unit 164 calculates the billing amount from the above-read information as follows. For example, if the unit price information of "supervision" read is 20,000 yen per person, and the number of "supervision" is one person and is specified by the number of two-dimensional codes 40E, the billing amount is calculated as 20,000 yen. This corresponds to the item of "supervision" in the settlement statement S. Also, if the unit price information of "construction worker" read is 10,000 yen per person, and the number of "construction worker" is one person and is specified by the number of two-dimensional codes 40E, the billing amount is calculated as 10,000 yen. This corresponds to the item of "construction worker" in the settlement statement S. Furthermore, if the unit price information of "dump truck" read is 20,000 yen per unit, and the number of "dump truck" is one unit and is specified by the number of two-dimensional codes 40E, the billing amount is calculated as 20,000 yen. This corresponds to the item of "dump truck" in the settlement statement S.

[0107] Next, the reading unit 163 reads from the database 15 the information belonging to the category of "time and period" associated with the two-dimensional code 40F recognized by the recognition unit 162, and the unit price information corresponding to that information. For example, in this embodiment, based on the two-dimensional code 40F (see FIG. 3) on the settlement board 23, the information of "late at night" and the unit price information corresponding to "late at night" are read from the database 15. In addition, when instead of arranging the two-dimensional code 40F at the site, the positioning device 12 is used to acquire time information, the reading unit 163 reads from the database 15 the unit price information of the corresponding time zone in comparison with the time information acquired by the positioning device 12.

[0108] Then, the document generation unit 164 calculates the billing amount from the above-read information as follows. For example, if the unit price information of "late at night" read is 10,000 yen per case, and a settlement statement S for one case of construction work is created, the billing amount is calculated as 10,000 yen. This corresponds to the item of "late-night surcharge" in the settlement statement S. In addition, when the time and period information is acquired by the analysis program (step S22 in FIG. 8), the billing amount is calculated based on the acquired information.

[0109] The data generation unit 164 summarizes and outputs the results calculated as described above in a list like the settlement statement S shown in FIG. 9. In addition, the reliability calculated by the recognition program is also noted in the settlement statement S. When the quantity is two or more, the minimum reliability is also noted. For example, when there are three, like "Socket 100A" in the table, the average value of the reliability is described in the "Reliability" column, and the minimum reliability among the three is also noted in the "Minimum Reliability" column. The reliability is also noted in the drawing D. Note that the reliability described in the drawing D and the settlement statement S is just an example.

[0110] Furthermore, the data generation unit 164 also creates and outputs a work report W. The work report W describes the work content of the civil engineering work. In this embodiment, "Straight pipe 100A connection work" and "Airtightness test" are cited as the work content. Furthermore, for the "Airtightness test", the result is described. The "Straight pipe 100A connection work" in the work report W combines the "Existing pipe connection work" and the "New pipe connection work" using the 100A straight pipe and describes it as "Straight pipe 100A connection work". This notation of "Straight pipe 100A connection work" is because the reading unit 163 reads out the type of the conduit 20, which is "Straight pipe", and the nominal diameter of the conduit 20, which is "100A", from the buried object information associated with the two-dimensional code 40A, and then automatically adds the word "connection work" to these "Straight pipe" and "100A" and outputs it. Also, the "Airtightness test" in the work report W and the description of its result are due to the fact that the "Airtightness test" was performed by the two-dimensional code 40B and the result of the test was specified, so the specified content is output. Note that the content described in the work report W is just an example and is not limited to the content of the drawing. In addition, the reliability calculated by the recognition program is also noted in the work report W. When the quantity is two or more, the minimum reliability is also noted. Note that the reliability described in the work report W is just an example.

[0111] The construction report materials R created as described above are stored in the database 15 as data that can be viewed, for example, on a computer or the like. Then, the work report W is transmitted to the client who ordered the civil engineering work. This transmission may be carried out by the construction report materials creation device 10 via the communication line 19, or may be transmitted by the worker via email or the like.

[0112] The verification unit 166 and the post-verification unit 25 are devices for verifying the construction report materials R (especially the settlement statement S and the work report W) submitted to the client of the civil engineering work. Specifically, it is as follows. The settlement statement S and the work report W are submitted to the client for billing the construction cost. When the client registers the settlement statement S and the work report W with the post-verification unit 25, the post-verification unit 25 transmits the settlement statement S and the work report W to the construction report materials creation device 10 via the communication line 19. Then, the verification unit 166 verifies the transmitted settlement statement S and work report W with the settlement statement S and work report W registered in the database 15. The verification items are, for example, the numbers of the two-dimensional codes 40A - 40G, the number of the two-dimensional codes 40A - 40G, the positions of the two-dimensional codes 40A - 40G, the position of the site, the settlement items, the settlement quantity, the settlement amount, the report items, the report quantity, the report result, the positioning status, the positioning time, the analysis time, the period, the site status, the work scope, the presence or absence of utilization of the dimensional correction member 24, the presence or absence of correction of the direction, the presence or absence of utilization of the elevation terrain data, the reliability, etc. As a result, the client can focus on confirming the presence of overbilling, those with low reliability, those with a reliability of 50 or less for which the recognition program has aborted recognition, and those corrected by the worker, and the labor can be reduced.

[0113] (First Modified Example of the Recognition Program) The recognition program described above recognized the extraction of cells by recognizing the colors in the as-built drawing and recognized the set as a two-dimensional code. However, it may be possible to recognize the two-dimensional code by shape recognition. That is, the two-dimensional code may be recognized by searching for the shape (concave shape) of the two-dimensional code from the as-built drawing. Also, the recognition program described above calculates the reliability based on the positional relationship of each cell constituting the two-dimensional code (that is, the distance and angle between the centers of gravity). However, it may be possible to calculate the reliability based on the contour shape of the two-dimensional code in the as-built drawing. This will be specifically described below. FIG. 10 is a flowchart showing a modified example of the recognition program. FIG. 11 is a diagram for explaining the concave shape recognition (step S41) in the recognition program. FIG. 12 is a diagram for explaining the cell division (step S42) in the recognition program.

[0114] First, the concave shape is recognized (step S41). This is a step for extracting the concave shape shown in the as-built drawing as a two-dimensional code. The as-built drawing is changed from grayscale (maximum 256) to binary grayscale with a level of 226. In this embodiment, since the two-dimensional code is concave, an octagon with 8 corners is extracted. Those with other than 8 corners are excluded as not being two-dimensional codes. Further, octagons 60A (60B) in which two inner angles of 90° (angle A61, angle A62) and 270° (angle A63, angle A64) are adjacent to each other are extracted. Then, a threshold of ±5° is set for the inner angles of 90° (angle A61, angle A62) and 270° (angle A63, angle A64), and it is determined whether the inner angle is within the threshold. As shown in the octagon 60A in FIG. 11(a), if the distortion is small, it is within the threshold. On the other hand, as shown in the octagon 60B in FIG. 11(b), if there is distortion, there is a possibility of exceeding the threshold. The octagon that exceeds the threshold is excluded as not being a two-dimensional code at this point. After extraction, the orientation of the two-dimensional code 40 is confirmed based on the position of the notch.

[0115] Next, cell division is performed (step S42). Due to image processing, there may be a case where the gaps between cells 401 are filled as shown in FIG. 12(a), and each cell 401 is integrated. In such a case, based on the notch 404, it is divided into five cells 401. For example, as shown in FIG. 9(b), it is divided into five cells by a division line 405 provided based on the notch 404.

[0116] Next, the number of cells included in the two-dimensional code is recognized (step S43). In the present embodiment, since the two-dimensional code is composed of five cells, those other than five are excluded as not being two-dimensional codes. Further, it is recognized whether the two-dimensional code is composed of unused colors or the number of colors (step S44). In the present embodiment, if the two-dimensional code is composed of five-color cells, when there are cells other than five colors, they are excluded as not being two-dimensional codes.

[0117] Next, the cells divided in step 43 are corrected to squares (step 45). In the process of synthesizing digital images, the shape of the two-dimensional code may be distorted. If division is performed based on the notch 404 in a distorted two-dimensional code, the divided cells may not be square. Also, when the two-dimensional code is attached to a cylindrical shape such as the conduit 20, a part of the square is missing and becomes a rectangular corner or a divided corner. The correction is to correct the missing part in consideration of the concave shape. The correction method is the same as step S4 in FIG. 5. Note that the corrected cells are referred to as correction parts.

[0118] Next, the contour of the two-dimensional code is recognized (step S46), and further, based on the recognized contour, an outer circumscribed rectangle surrounding the two-dimensional code is obtained (step S47). The outer circumscribed rectangle is a rectangle whose four sides are all circumscribed to the target two-dimensional code. For example, like the set 50A of correction parts 501 - 505 shown in FIG. 13(a), if the distortion is small, the outer circumscribed rectangle C11 is substantially rectangular. On the other hand, like the set 50D of correction parts 501 - 505 shown in FIG. 13(b), if there is distortion, the outer circumscribed rectangle C12 is a parallelogram.

[0119] Next, the difference between the area of the two-dimensional code and the area of the circumscribed rectangle is obtained (step S48). Then, a predetermined threshold value (for example, 40% ± 3%) is set in advance, and it is determined whether the calculated area difference is within the threshold value. If the distortion is small, like the set 50A of the correction parts 501 - 505 shown in FIG. 13(a), the difference between the area of the set 50A and the area of the circumscribed rectangle C11 is equivalent to the area of the notch part and is within the threshold value. On the other hand, if the distortion is large, like the set 50D of the correction parts 501 - 505 shown in FIG. 13(b), the difference between the area of the two-dimensional code and the area of the circumscribed rectangle becomes larger than when there is no distortion and may exceed the threshold value. The set of correction parts that exceeds the threshold value is excluded as not being a two-dimensional code at this point. Instead of obtaining the difference between the area of the set of correction parts and the area of the circumscribed rectangle, it is also possible to calculate the ratio of the area of the two-dimensional code to the area of the circumscribed rectangle and determine whether the ratio falls within a predetermined threshold value range.

[0120] In addition, instead of obtaining the difference between the area of the set of correction parts and the area of the circumscribed rectangle, the interior angles of the circumscribed rectangle can be calculated, and it can be determined whether the interior angles fall within a predetermined threshold value range (90° ± 5°). For example, if the distortion is small, like the set 50A of the correction parts 501 - 505 shown in FIG. 13(a), all the interior angles A41 - A44 of the circumscribed rectangle C11 are approximately right angles and are within the threshold value range. On the other hand, if the distortion is large, like the set 50D of the correction parts 501 - 505 shown in FIG. 13(b), the interior angles A51 - A54 of the circumscribed rectangle C12 become larger or smaller than a right angle and may exceed the threshold value range. The set of correction parts that exceeds the threshold value range is excluded as not being a two-dimensional code at this point.

[0121] Alternatively, it may be configured to calculate the area of the circumscribed rectangle and determine whether the area falls within a predetermined threshold range. For example, as in the set 50A of the correction units 501 - 505 shown in FIG. 13(a), if the distortion is small, the area of the circumscribed rectangle C11 falls within the threshold range. On the other hand, as in the set 50D of the correction units 501 - 505 shown in FIG. 13(b), if the distortion is large, the area of the circumscribed rectangle C12 becomes large and may exceed the threshold. At this point, the set of correction units that exceeds the threshold range is excluded as not being a two-dimensional code. Note that the above threshold is, for example, ±20% with the area of the circumscribed rectangle when there is no distortion in the set of correction units at all as the central value.

[0122] Next, label recognition is performed (step S49). The processing content is the same as that of step S9 in FIG. 5. The set of correction units determined to have no label is excluded as not being a two-dimensional code at this point.

[0123] Next, for the set of correction units determined to have a label in step S49, color recognition is performed (step S50). For each color used in the cells of the two-dimensional code, a range of hue, lightness, and saturation is set, and it is determined what color the correction unit is based on whether it falls within that range. If there is even one correction unit in the set of correction units that does not fall within the set range, the set is excluded as not being a two-dimensional code at this point.

[0124] After going through the above steps, for those determined to be two-dimensional codes, a reliability is calculated as an index of whether the recognition of the two-dimensional code is accurate (step S51). The reliability is calculated, for example, using a point deduction method with a full score of 100 points. For example, if the thresholds in steps S41, S45, S48 - S50 are strict, the amount of point deduction for reliability is small, while if the thresholds are loose, the amount of point deduction for reliability becomes large.

[0125] After calculating the reliability, the thresholds in steps S41, S45, S48 - S50 are reset (step S52), and steps S41 - S51 are repeated. In step S11, at least the threshold in step S41 is increased. This is because in the first - week step S50, the threshold is strict, so there is a possibility of overlooking what is originally a two - dimensional code. In addition, if there is something that has been excluded as not being a two - dimensional code in steps S41, S45, S48 - S49, the threshold of the step where the exclusion was made is increased. This is because in the first - week steps S45, S48 - S49, the threshold is strict, so there is a possibility that what is originally a two - dimensional code has been excluded due to image distortion. Note that the point - deduction method when repeating is to deduct the total points accumulated from the first week to the last week.

[0126] When a two - dimensional code missed in step S41 is recognized, even with the same threshold, it may continue to be judged as a two - dimensional code in steps S45, S48 - S50. Therefore, devise the threshold - increasing method, the steps to repeat in the second week, the number of repetitions, and the point - deduction method for reliability. For example, only reset the threshold in step S41 and repeat steps S41 - S51. In this case, the steps that are not reset do not deduct reliability points. On the other hand, when resetting the thresholds in steps S41, S45, S48 - S50 and repeating steps S41 - S51, since the thresholds of S45, S48 - S50 are lowered too much with respect to S41, the amount of increase in the threshold is made smaller, and the number of points deducted for steps S45, S48 - S50 is reduced. Also, the threshold in step S50 has thresholds for hue, lightness, and saturation for each color number, resulting in multiple thresholds. However, since it has been judged as a two - dimensional code for things other than color up to steps S41 - S49, it is desirable to judge color early. For this reason, the initial value of the threshold in step S50 is set wide, and the amount of reliability deduction for step S50 is minimized with respect to steps S41, S45, S48 - S49. By this device, it becomes possible to deduct reliability according to the change in the threshold and the number of repetitions.

[0127] For example, the threshold value in step S41 (concave recognition) is enlarged by 3% every time the recognition program makes one round. Each time it is enlarged, the reliability calculated in step S51 is deducted by 10% of the full score (for example, if the full score is 100 points, it is deducted by 10 points). The threshold value in step S45 (square correction) is enlarged by 10% every time the recognition program makes one round. Each time it is enlarged, the reliability calculated in step S51 is deducted by 5% of the full score (for example, if the full score is 100 points, it is deducted by 5 points). The threshold value in step S48 (area difference calculation) is enlarged by 5% every time the recognition program makes one round. Each time it is enlarged, the reliability calculated in step S51 is deducted by 1% of the full score (for example, if the full score is 100 points, it is deducted by 1 point). The threshold value in step S49 (label recognition) is enlarged by 3% every time the recognition program makes one round. Each time it is enlarged, the reliability calculated in step S51 is deducted by 1% of the full score (for example, if the full score is 100 points, it is deducted by 1 point). The ranges of hue, lightness, and saturation in step S50 (color recognition) are enlarged by 1% every time the recognition program makes one round. Each time it is enlarged, the reliability calculated in step S51 is deducted by 0.1% of the full score (for example, if the full score is 100 points, it is deducted by 0.1 point).

[0128] However, when the threshold value is reset, if the reliability calculated in step S51 becomes 50 or less (step S53: YES), the repetition of steps S1 - S10 is not performed and the recognition program ends. In this case, a record with a reliability of 50 or less is recorded in the database 15. Regardless of the value of the reliability, if the operator modifies the numerical sequence of the two-dimensional code, the modification history is recorded in the database 15. The drafting unit 161 combines the reliability recorded in the database 15 with the two-dimensional code in the ortho image. Also, it can be held as an attribute of the conduit 20 and the joint 21 shown on the three-dimensional CAD drawing and the two-dimensional CAD drawing, and can also be combined and noted.

[0129] In the recognition program described so far, the reliability may be calculated based on the area of the two-dimensional code 40 or the area of the cells constituting the two-dimensional code 40. Specifically, the area of the two-dimensional code 40 here is the area of the set of correction units 501-505, and it is determined whether the area falls within a predetermined threshold range. The above threshold is, for example, ±20% with the area of the two-dimensional code 40 being used as the central value.

[0130] Also, the area of the cell here is the area of the correction units 501-505, and it is determined whether the area falls within a predetermined threshold range. The above threshold is, for example, ±20% with the area of the cell 401 of the two-dimensional code 40 being used as the central value.

[0131] In this way, from the perspective of area, the set of correction units that are considered not to be two-dimensional codes is excluded. After the second round of the recognition program, the threshold is enlarged by 10% each time the recognition program makes one round, and each time the reliability is deducted by 5% of the full score.

[0132] In addition, in the recognition program, based on the shape of the two-dimensional code 40, the centroid of the two-dimensional code 40 may be calculated, and the reliability may be calculated based on the centroid.

[0133] Specifically, for example, as shown in FIG. 14, the centroid G11 of the set 50A itself is calculated from the set 50A of the correction units 501-505. In addition, the centroid G12 of the correction unit 503 located at the center in the set 50A is calculated. Then, the length of the line segment L44 connecting the centroid G11 and the centroid G12 is calculated, and it is determined whether the length falls within a predetermined threshold range. If there is distortion in the shape of the set 50A of the correction units 501-505, the position of the centroid G11 changes, and the length of the line segment L44 may not fall within the threshold range. Therefore, the set of correction units that are considered not to be two-dimensional codes can be excluded in this way. The above threshold is, for example, 0.3 times the vertical length of the cell 401 of the two-dimensional code 40 being used ±20%.

[0134] Alternatively, instead of determining whether the length of line segment L44 is within a predetermined threshold range, it may be determined whether the angle is within the range of a predetermined threshold. The angle in this case is, for example, the angle with respect to the horizontal or vertical direction. Note that the above-mentioned threshold is, for example, 0 degrees ± 10%.

[0135] Alternatively, as shown in FIG. 15, the lengths of line segments L51 and L52 connecting the center of gravity G11 and both end portions P11 and P12 of the concave set 50A may be calculated, and it may be determined whether the lengths are within a predetermined threshold range. If there is distortion in the shape of the set 50A of the correction units 501 - 505, the position of the center of gravity G11 changes, and there is a possibility that the lengths of the line segments L51 and L52 do not fall within the threshold range. Thus, it is possible to exclude a set of correction units that is considered not to be a two-dimensional code. Alternatively, instead of the lengths of the line segments L51 and L52, it may be determined whether the angle between the line segment L51 and the line segment L52 is within a predetermined threshold range (for example, 106° ± 10%). When using the center of gravity G11 as described above, in the second and subsequent rounds of the recognition program, the threshold is increased by 3% each time the recognition program makes one round, and the confidence level is deducted by 1% of the full score each time.

[0136] (Second Modification Example of the Recognition Program) It may also be possible to calculate the confidence level based on the information read from the two-dimensional code 40. The information read from the two-dimensional code 40 is, for example, information such as the diameters, types, and materials of the conduits 20 and joints 21, and the confidence level is calculated from the compatibility of the combination of the conduits 20 and joints 21.

[0137] For example, when the network of embedded objects configured based on the information read from the two-dimensional code 40 is a combination such as a PE75 socket - PE75 90° elbow, a PE75 socket - PE75 45° elbow, or a PE75 socket - PE cast iron connecting joint 75, since this can actually be piped, it can be determined that it is a possible combination. On the other hand, for example, combinations such as a PE75 socket - welded steel pipe 75 90° bend, a PE75 socket - PE150 socket, or a PE30 service cheese - PE30 service cheese mean connecting joints with different diameters and cannot actually be piped, so it can be determined that they are unlikely combinations.

[0138] If the network configured by the information read from the two-dimensional code 40 is a possible combination, no deduction of reliability is made. However, if it is an unlikely combination, a 50% deduction from the full score is made for the reliability of both two-dimensional codes 40.

[0139] (Third Modification Example of the Recognition Program) It is also possible to calculate the reliability based on the position information acquired by the positioning device 12 and the information read from the two-dimensional code 40.

[0140] "Position information" refers to the position information of the embedded object (inside or outside the installation groove 30a, height, distance, etc.). Based on such position information, a determination of reliability is made according to whether the type (shape) of the joint 21 read from the two-dimensional code 40 is layable.

[0141] For example, when all the buried objects are within the installation trench 30a, the combination of a PE75 socket - PE75 90° elbow has the same width but different heights, and the distance between the PE75 sockets is within the length of the straight pipe, etc., it can be actually laid, so it can be judged that the buried object is in a proper position. On the other hand, when there is a joint 21 outside the installation trench 30a, the combination of a PE75 socket - PE75 90° elbow has the same width and the same height, and the distance between the PE75 sockets is longer than the length of the straight pipe, etc., it means that a non-existent straight pipe is used to connect the joints with angle changes linearly, and it cannot be actually laid, so it can be judged that the buried object is in an unlikely position.

[0142] If the network composed of the information read from the two-dimensional code 40 can be judged to be in a proper position, no deduction of the reliability score will be made. However, if it is judged to be in an unlikely position, a deduction of 50% of the full score will be made.

[0143] (The fourth modification example of the recognition program) It may also be possible to calculate the reliability by comparing the information read from the two-dimensional code 40 with the past civil engineering achievements. Specifically, the combination of the position information, year-date-time information, time information obtained by the positioning device 12 and the information read from the two-dimensional code 40 is judged for reliability based on whether it is a combination that is common (or an unlikely combination) from the perspective of probability compared with the past civil engineering achievements.

[0144] For example, combinations such as downtown - doorman, water supply point - installation of water intake wells, PE75 socket - construction worker, and Naka-ku, Nagoya City - downtown have past achievements and a high probability of combination, so it can be judged that they are common combinations from the perspective of probability. On the other hand, combinations such as PE75 socket - welder, tunneling work - loss of propulsion device, New Year's Day - weekday rate, daytime - late-night rate have no (or few) past achievements and a low probability of combination, so it can be judged that they are unlikely combinations (suspecting an error in the installation of the two-dimensional code 40).

[0145] If the combination is judged to be common from a probability standpoint, no points will be deducted from the reliability, but if the combination is judged to be unlikely, 50% of the full score will be deducted.

[0146] As described above, according to the code processing system 1 of the present embodiment, (1) In a code processing system 1 including a code (e.g., two-dimensional code 40A-40G) that is placed at a civil engineering work site and associated with information related to the civil engineering work, a photographing device 11 that takes multiple photographs of the civil engineering work site, and a recognition unit 162 for recognizing an in-image code (e.g., two-dimensional code 40A-40G in a completion drawing), which is a code that appears in an image taken by the photographing device 11, the code (two-dimensional code 40A-40G) is a two-dimensional arrangement of multiple cells 401 having a predetermined color, includes a notch 404 for detecting the position of the code area, and represents information related to the civil engineering work by a combination of colors; The present invention is characterized by comprising a recognition program which calculates the reliability of whether the recognition unit 162 has accurately recognized the code in the image (two-dimensional code 40A-40G in the completion drawing) based on at least one of the following: the positional relationship of the cells 401 (correction units 501-505) constituting the code in the image (two-dimensional code 40A-40G in the completion drawing), the number of cells 401 (correction units 501-505), the color of the cells 401 (correction units 501-505), the contour shape of the code in the image (two-dimensional code 40A-40G in the completion drawing), the color of the code in the image (two-dimensional code 40A-40G in the completion drawing), and information read from the code in the image (two-dimensional code 40A-40G in the completion drawing).

[0147] (2) In the code processing system 1 described in (1), the positional relationship is the distance between the centers of gravity G1-G5 of the cells 401 (correction units 501-505) constituting the code in the image, and it is preferable to calculate the reliability based on this distance.

[0148] (3) In the document processing system 1 described in (1) or (2), the above positional relationship refers to the angles A21 - A23 of the line segments L21 - L24 when the centroids G1 - G5 of the respective cells 401 (correction units 501 - 505) are connected by the line segments L21 - L24, and it is preferable to calculate the reliability based on the angles A21 - A23.

[0149] (4) In the document processing system 1 described in (1), based on the shape of the in-image document (2D codes 40A - 40G in the as-built drawing), obtain the circumscribed rectangle C11 (C12) that encloses the in-image document (2D codes 40A - 40G in the as-built drawing), and it is preferable to calculate the reliability based on the area of the circumscribed rectangle C11 (C12).

[0150] (5) In the document processing system 1 described in (1), based on the shape of the in-image document (2D codes 40A - 40G in the as-built drawing), obtain the circumscribed rectangle C11 (C12) that encloses the in-image document (2D codes 40A - 40G in the as-built drawing), and it is preferable to calculate the reliability based on the difference between the area of the in-image document (2D codes 40A - 40G in the as-built drawing) and the area of the circumscribed rectangle C11 (C12).

[0151] (6) In the document processing system 1 described in (1), based on the shape of the in-image document (2D codes 40A - 40G in the as-built drawing), obtain the circumscribed rectangle C11 (C12) that encloses the in-image document (2D codes 40A - 40G in the as-built drawing), and it is preferable to calculate the reliability based on the interior angles A41 - A44 (A51 - A54) of the circumscribed rectangle C11 (C12) (step S41).

[0152] (7) In the document processing system 1 described in (6), based on the shape of the in-image document (2D codes 40A - 40G in the as-built drawing), calculate the interior angles (angles A61 - A64 (angles A71 - A74)) of the in-image document (2D codes 40A - 40G in the as-built drawing), and it is preferable to calculate the reliability based on whether the interior angles (angles A61 - A64 (angles A71 - A74)) are within a predetermined threshold.

[0153] (8) In the document processing system described in (1), it is preferable to calculate the centroid G11 of the in-image document (the two-dimensional codes 40A - 40G in the as-built drawing) based on the shape of the in-image document, and calculate the reliability based on the centroid G11.

[0154] (9) In the document processing system 1 described in (1), it is preferable to calculate the area of the in-image document (the area of the set of correction parts 501 - 505) based on the shape of the in-image document, and calculate the reliability based on the area.

[0155] (10) In the document processing system 1 described in (1), it is preferable to calculate the area of each cell (the area of correction parts 501 - 505) that constitutes the in-image document (the two-dimensional codes 40A - 40G in the as-built drawing) based on the shape of the in-image document, and calculate the reliability based on the area.

[0156] (11) In the document processing system 1 described in (1), it is preferable to recognize the number of corners of the in-image document (the number of corners of correction parts 501 - 505) based on the shape of the in-image document, and calculate the reliability based on the number of corners (for example, whether the number of corners is 8).

[0157] (12) In the document processing system described in (1), it is preferable to calculate the reliability based on the positioning device used at the civil engineering site and based on the position information acquired by the positioning device and the information read from the in-image document.

[0158] (13) In the document processing system described in (1), it is preferable to compare the information read from the in-image document (the two-dimensional codes 40A - 40G in the as-built drawing) with the past civil engineering achievements and calculate the reliability.

[0159] (14) In the document processing system 1 according to any one of (1) to (9), it is preferable to include a document generation unit 164 that creates a construction report document R describing the results of civil engineering work based on the information associated with the document (two-dimensional codes 40A-40G), and that the reliability calculated by the recognition program is described in the construction report document R.

[0160] (15) In the document processing system 1 according to (1), the recognition program determines whether the in-image document is a document according to the threshold value of each item among the number of cells constituting the in-image document, the shape of the cells, the color of the cells, the positional relationship of the cells, the shape of the in-image document, and the color of the in-image document. When resetting the threshold value in the item related to the determination of whether it is a document among each item, a deduction (first deduction) of the reliability is performed only in the item where the resetting is performed. When resetting the threshold value in the item other than the item related to the determination of whether it is a document among each item, in the item where the resetting is performed, a deduction of the reliability is performed in a small amount with respect to the first deduction for the items other than the item related to the determination of whether it is a document.

[0161] (16) In the document processing system 1 according to (1), it is preferable that the recognition program extracts cells from the image captured by the imaging device 11 (for example, step S1), and calculates the reliability by determining whether the set of cells is a document (for example, steps S6-S9).

[0162] (17) In the document processing system 1 according to (16), it is preferable that the recognition program determines whether the set of cells is the document based on the shape of the set of cells extracted from the image captured by the imaging device 11 (for example, step S6).

[0163] In the document processing system 1 described in (18), (15) or (16), the recognition program determines whether it is a cell by extracting cells constituting the document in the image according to the number of cells, the shape of the cells, the color of the cells, and the threshold of the positional relationship of the cells. When resetting the threshold in the item related to the determination of whether it is a cell among each item, a deduction of reliability (the first deduction) is performed only in the item where the resetting is performed. When resetting the threshold in an item other than the item related to the determination of whether it is a cell and the item related to the determination of whether it is a document among each item, in the item where the resetting is performed, a deduction of the reliability is performed in a small amount with respect to the first deduction for items other than the item related to the determination of whether it is a cell. This is preferable.

[0164] According to the above-described document processing system 1, since the reliability of whether the recognition of the document in the image (two-dimensional codes 40A - 40G in the completion drawing) is accurately performed based on the recognition program is calculated, by using this reliability in, for example, the construction report material R, it is possible to reduce the labor for checking the mistakes in the construction report material R and for the inspection of the construction report material. Specifically, for example, when information (such as the cost generated according to the work content of the civil engineering work) is described in the construction report material R based on the information associated with the document (two-dimensional codes 40A - 40G), the reliability of whether the corresponding document (two-dimensional codes 40A - 40G) is accurately recognized is also noted. By doing so, when checking and inspecting the content of the construction report material R, it is only necessary to focus on checking those with low reliability, those with a reliability of 50 or less where the recognition program has aborted the recognition, and those corrected by the operator, so that the labor can be reduced.

[0165] Note that the above embodiments are merely illustrative and do not limit the present invention in any way. Therefore, the present invention can naturally be improved and modified in various ways without departing from its gist. For example, the form of the two-dimensional code is not limited to the two-dimensional code 40 formed by the five cells shown in this embodiment, and it may be a two-dimensional code formed by six cells or a two-dimensional code formed by seven cells. As the two-dimensional code formed by six cells or seven cells, for example, the two-dimensional code disclosed in Japanese Patent Application Laid-Open No. 2020-160626 is used.

[0166] Also, although it has been described that the absolute coordinates acquired by the positioning device 12 are used for specifying the position of the buried object etc., it may be possible to use relative coordinates with respect to the reference position based on the three-dimensional point cloud data, three-dimensional mesh data, orthoimage, three-dimensional CAD drawing, two-dimensional CAD drawing generated by the drawing unit 161. For example, if an intersection of a road etc. is shown on the drawing, relative coordinates are calculated with the intersection as the reference position.

[0167] Also, the construction report material creation device 10 may be used as a portable terminal at the site of civil engineering work.

[0168] Also, in this embodiment, the construction report material R is generated by collecting the drawing D, the work report W, and the settlement statement S as one material, but the drawing D, the work report W, and the settlement statement S may be generated as separate materials.

[0169] Also, in this embodiment, a gas conduit is described as an example of the buried object, but the buried object is not limited to this, and it may be a sewage pipe, an optical cable, a water pipe, etc.

[0170] When the buried object is a sewage pipe, the plurality of tags include tags for identifying the results of the flow test performed after the laying of the sewage pipe, and the construction report material includes a work report showing the results of the flow test.

[0171] When the buried object is an optical cable, the plurality of tags include tags for identifying the results of the communication test conducted after the laying of the optical cable, and the construction report materials include a work report indicating the results of the communication test.

[0172] When the buried object is a water pipe, the plurality of tags include tags for identifying the results of the leakage test conducted after the laying of the water pipe, and the construction report materials include a work report indicating the results of the leakage test.

Explanation of symbols

[0173] 1 Tag processing system 11 Photographing device 20 Conduit (an example of a buried object) 21 Joint (an example of a buried object) 40A - 40G 2D code (an example of a tag) 162 Recognition unit 163 Reading unit 164 Material generation unit

Claims

1. A token associated with information regarding the civil engineering work, which is arranged and used at the site of the civil engineering work, a photographing device that takes a plurality of photographs of the site of the civil engineering work, a recognition unit for recognizing an in-image token that is the token appearing in the image photographed by the photographing device, in a token processing system comprising: the token has a plurality of cells having a predetermined color two-dimensionally arranged, includes a notch for detecting the position of the area of the token, and represents the information by a combination of colors; a recognition program for calculating a reliability as to whether the recognition of the in-image token by the recognition unit is accurately performed based on at least any one of the positional relationship of the cells constituting the in-image token, the number of the cells, the color of the cells, the shape of the in-image token, the color of the in-image token, and the information read from the in-image token; A token processing system characterized by the above.

2. In the token processing system according to Claim 1, the positional relationship is the distance between the centers of gravity of the respective cells constituting the in-image token, and the reliability is calculated based on the distance; A token processing system characterized by the above.

3. In the token processing system according to Claim 1, the positional relationship is the angle of the line segment when the centers of gravity of the respective cells are connected by a line segment, and the reliability is calculated based on the angle; A token processing system characterized by the above.

4. In the token processing system according to Claim 1, an outer circumscribed rectangle surrounding the in-image token is obtained based on the shape of the in-image token, and the reliability is calculated based on the area of the outer circumscribed rectangle; A token processing system characterized by the above.

5. In the token processing system according to Claim 1, an outer circumscribed rectangle surrounding the in-image token is obtained based on the shape of the in-image token, and the reliability is calculated based on the difference between the area of the in-image token and the area of the outer circumscribed rectangle; A token processing system characterized by the above.

6. In the token processing system according to Claim 1, an outer circumscribed rectangle surrounding the in-image token is obtained based on the shape of the in-image token, and the reliability is calculated based on the interior angles of the outer circumscribed rectangle; A token processing system characterized by the above.

7. In the token processing system according to Claim 1, the interior angles of the in-image token are calculated based on the shape of the in-image token, and the reliability is calculated based on whether the interior angles are within a predetermined threshold value. A permit processing system characterized by

8. In the permit processing system according to Claim 1, calculating the center of gravity of the permit in the image based on the shape of the permit in the image, calculating the reliability based on the center of gravity, A permit processing system characterized by

9. In the permit processing system according to Claim 1, calculating the area of the permit in the image based on the shape of the permit in the image, calculating the reliability based on the area, A permit processing system characterized by

10. In the permit processing system according to Claim 1, calculating the area of each cell constituting the permit in the image based on the shape of the permit in the image, calculating the reliability based on the area, A permit processing system characterized by

11. In the permit processing system according to Claim 1, recognizing the number of corners of the permit in the image based on the shape of the permit in the image, calculating the reliability based on the number of corners, A permit processing system characterized by

12. In the permit processing system according to Claim 1, equipped with a positioning device used at the civil engineering site, calculating the reliability based on the position information acquired by the positioning device and the information read from the permit in the image, A permit processing system characterized by

13. In the permit processing system according to Claim 1, comparing the information read from the permit in the image with the past performance of civil engineering works and calculating the reliability, A permit processing system characterized by

14. In the permit processing system according to any one of Claims 1 to 13, comprising a document generation unit that creates a construction report document describing the results of the civil engineering work based on the information associated with the permit, the reliability calculated by the recognition program is described in the construction report document, A permit processing system characterized by

15. In the permit processing system according to Claim 1, the recognition program in each item of the number of cells constituting the permit in the image, the shape of the cell, the color of the cell, the positional relationship of the cells, the shape of the permit in the image, and the color of the permit in the image, determines whether the permit in the image is a permit according to the threshold value of each item, when resetting the threshold value in the item related to the determination of whether it is a permit among the above items, a point deduction (first point deduction) of the reliability is performed only in the item where the resetting is performed, Among the above items, when resetting the threshold value in the item related to the determination of whether it is a document voucher and the items other than the item related to the determination of whether it is a document voucher, in the item where the resetting is performed, with respect to the first point deduction, a point deduction for the reliability is performed in a small amount for the point deduction other than the item related to the determination of whether it is a document voucher. A document voucher processing system characterized by the above.

16. In the document voucher processing system according to Claim 1, the recognition program extracts the cells from the image captured by the imaging device, and calculates the reliability by determining whether the set of the cells is the document voucher. A document voucher processing system characterized by the above.

17. In the document voucher processing system according to Claim 16, the recognition program determines whether the set of the cells is the document voucher based on the shape of the set of the cells extracted from the image captured by the imaging device. A document voucher processing system characterized by the above.

18. In the document voucher processing system according to Claim 15 or 16, the recognition program determines whether it is a cell by extracting the cells constituting the document voucher in the image according to the thresholds of the number of cells, the shape of the cells, the color of the cells, and the positional relationship of the cells. Among the above items, when resetting the threshold value in the item related to the determination of whether it is a cell, a point deduction (first point deduction) for the reliability is performed only in the item where the resetting is performed. Among the above items, when resetting the threshold value in the items other than the item related to the determination of whether it is a cell and the item related to the determination of whether it is a document voucher, in the item where the resetting is performed, with respect to the first point deduction, a point deduction for the reliability is performed in a small amount for the point deduction other than the item related to the determination of whether it is a cell. A document voucher processing system characterized by the above.

Images