Reinforcing-bar arrangement measurement system and reinforcing-bar arrangement measurement method

Abstract

To reduce errors in measurement in a reinforcing-bar arrangement measurement system.SOLUTION: In a reinforcing-bar arrangement measurement system according to an embodiment, a position storage unit receives a process of specifying at least each reinforcing-bar position along a longitudinal direction of a reinforcing bar, in a captured image including a reference object indicating an actual size, and at least temporarily stores each specified reinforcing-bar position. A designation unit designates a range of a predetermined length, based on actual size indicated by the reference object for each reinforcing-bar region indicated at each reinforcing-bar position. According to an identification element characteristic for each of reinforcing-bar classifications, a determination unit determines to which classification a partial image within the range designated by the designation unit belongs.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a bar arrangement measurement system and a bar arrangement measurement method. [Background technology]

[0002] In civil engineering and architectural works for structures using reinforced concrete, there is a process of placing rebars at intervals. Conventionally, after this process, the number and spacing of rebars to be placed were checked visually. However, in recent years, there has been a demand for improved quality of measurements and labor saving at civil engineering sites.

[0003] In this regard, technology is being developed that photographs the placed rebar (rebar finished form) and measures the rebar diameter, rebar spacing, etc. based on the data. For example, in civil engineering works using reinforced concrete, the person carrying out the rebar placement work photographs the rebar diameter, rebar spacing, etc. of the rebar finished form and converts it into image data. For example, a technology has been proposed that detects and measures rebars using a region detection model from image data obtained by photographing the rebar finished form (for example, Patent Document 1). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent Publication No. 2021-039009 Summary of the Invention [Problem to be solved by the invention]

[0005] When photographing the finished shape of rebar, there is a problem that the way in which knots and ribs of the rebar are captured is not consistent. Even if the accuracy of detecting rebars in the photographed image is high, if there is a problem with how knots and ribs are captured, the rebar diameter measured based on the detected rebars may contain errors. Also, the rebars in the photographed images are cylindrical in shape. This means that they may be depicted as if they have different thicknesses depending on the distance between the photographing means and the target.

[0006] The present invention has been made to solve the above-mentioned problems, and has an object to reduce measurement errors in a reinforcement bar measurement system. [Means for solving the problem]

[0007] The reinforcing bar measurement system of the embodiment includes a position storage unit, a designation unit, a discrimination unit, and an acquisition unit. The position storage unit performs a process of identifying at least each reinforcing bar position along the longitudinal direction of the reinforcing bar in a captured image including a reference object showing the actual size, and at least temporarily stores each identified reinforcing bar position. The designation unit designates a range of a predetermined length for each reinforcing bar area shown at each reinforcing bar position based on the actual size shown on the reference object. The discrimination unit determines which classification a partial image within the range designated by the designation unit belongs to according to a characteristic identification element for each classification of reinforcing bar. The acquisition unit acquires a reinforcing bar diameter using a discrimination result of the discrimination unit. In the reinforcing bar measurement system of the embodiment, the designation unit may designate the range by the predetermined length in the image and a predetermined width in a direction perpendicular to the length direction based on a reference object in the captured image.

[0008] The reinforcement bar measurement system of the embodiment may have an image processing unit that converts the line of sight of an image captured from one viewpoint so that the vertical and horizontal directions of the image correspond to the vertical and horizontal directions of the reinforcement bar.

[0009] In the bar arrangement measurement system according to the embodiment, the discrimination unit may discriminate the classification by image classification using machine learning.

[0010] In the embodiment of the reinforcement bar measurement system, the position of the reinforcing bar may be identified by identifying the reinforcing bar parts and non-reinforcing bar parts in the converted image using machine learning.

[0011] In the reinforcement bar arrangement system of the embodiment, the line of sight direction may be converted by the image processing unit according to an input specifying one range or the entire range of the captured image. Effect of the Invention

[0012] According to the embodiment, it is possible to reduce errors in measuring reinforcement in the reinforcement measurement system. [Brief description of the drawings]

[0013] [Figure 1] FIG. 1 is a schematic block diagram showing a bar arrangement measurement system according to a first embodiment. [Diagram 2] An example of a captured image. [Figure 3A] An example of a reference object showing actual size. [Figure 3B] An example of a partial image [Figure 4] 3 is a schematic flowchart according to the first embodiment. [Diagram 5] FIG. 11 is a schematic block diagram showing a bar arrangement measurement system according to a second embodiment. [Figure 6] An example of an image before perspective transformation (projection transformation). [Figure 7] An example of an image after perspective transformation (projection transformation). [Figure 8] 11 is a schematic flowchart according to a second embodiment. [Figure 9] FIG. 11 is a schematic block diagram showing a bar arrangement measurement system according to a third embodiment. [Figure 10] FIG. 13 is a schematic block diagram showing a bar arrangement measurement system according to a fourth embodiment. [Figure 11A] First example of rod tape placement. [Figure 11B] A second example of rod tape placement. [Figure 11C] Third example of rod tape arrangement. [Figure 12] An example of a binary image of a rod tape image. [Figure 13] 13 is a schematic flow chart of rod tape placement detection. [Figure 14] An example showing rod tape detection results. [Figure 15] An example showing before and after frame reduction processing [Figure 16] 13 is an example showing the process of rod tape detection for the vertical and horizontal directions in a cross pattern. [Figure 17]13 is a schematic flowchart according to a fourth embodiment. [Figure 18] 13 is a diagram showing an example of a modification of the second embodiment. [Figure 19] FIG. 13 is a schematic block diagram showing a bar arrangement measurement system according to a fifth embodiment. [Figure 20] FIG. 13 is a conceptual diagram showing a scanning direction in the fifth embodiment. [Figure 21] 13 is a schematic flowchart according to a fifth embodiment. [Figure 22] An example of a binary image after perspective transformation (projection transformation). [Diagram 23] FIG. 11 is a conceptual diagram showing a vertical white pixel region. [Figure 24] FIG. [Diagram 25] A conceptual diagram showing vertical scanning and ratio judgment. [Figure 26] FIG. 1 is a conceptual diagram showing lateral scanning. [Figure 27] A schematic diagram showing the pixel replacement state after horizontal scanning and determination. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] The bar arrangement measuring system according to each embodiment will be described with reference to Figs. 1 to 27.

[0015] [First embodiment] The overall configuration of the bar arrangement measurement system 100 according to the first embodiment will be described with reference to Fig. 1 to Fig. 4. Note that the bar arrangement measurement system in Fig. 1 is just an example, and does not exclude the inclusion of other configurations, and can be embodied in various forms.

[0016] (System Overview) As shown in FIG. 1, the reinforcing bar measurement system 100 includes a photographing unit 110, a storage unit 120, an image processing unit 130, a designation unit 140, a discrimination unit 150, an acquisition unit 160, a control unit C, and a display unit D. The photographing unit 110 is an imaging unit provided in, for example, a tablet terminal. The imaging data obtained by the imaging unit is stored in the storage unit 120 as image data by operating a touch panel or a save button. The storage unit 120 also performs a process of identifying at least each reinforcing bar position along the longitudinal direction of the reinforcing bar in the photographed image including a reference object showing the actual size, and stores each identified reinforcing bar position at least temporarily. The image processing unit 130 executes a process of generating and displaying the photographed data. The designation unit 140 designates a range of a predetermined length for each reinforcing bar area shown in each temporarily stored reinforcing bar position based on the actual size shown on the reference object. The discrimination unit 150 determines which classification the partial image within the range designated by the designation unit belongs to according to a characteristic identification element for each classification of reinforcing bar. The acquisition unit 160 acquires the reinforcing bar diameter using the discrimination result of the discrimination unit. Note that the images shown in the following figures are examples.

[0017] (Camera Division 110) The photographing unit 110 has an optical system (not shown), i.e., a lens, an aperture, a shutter, etc. The photographing unit 110 may have a focusing function (autofocus, etc.). The photographing unit 110 also has an image sensor (not shown), which controls the optical system to generate image data. Examples of the image sensor include image sensors such as a CCD (Charge Coupled Device) and a CMOS (Complementary Metal Oxide Semiconductor). The photographing unit 110 also has a signal processing circuit that performs A / D conversion of the image data.

[0018] (Storage unit 120) The storage unit 120 stores imaging data, image data, various programs, and designated rebar positions (described later). Any device can be used as the storage unit 120, including HDD, SSD, and flash memory. The storage unit 120 also includes a temporary storage device that stores at least temporary data. The main storage device may be included in the storage unit 120, or may be configured to use an external storage means.

[0019] (Image processing unit 130) The image processing unit 130 generates image data that can be displayed on any display means (for example, display means D) based on the imaging data stored in the storage unit 120. Furthermore, when an instruction to display image data is received via an operation unit (not shown), the image processing unit 130 executes a process to display the captured image based on the image data according to the instruction.

[0020] (Specification part 140) <Specifying rebar position> In this embodiment, it is possible for the user to specify the position of the reinforcing bar on the displayed captured image via a pointing device, etc. When this specification is received, the specification unit 140 stores the reinforcing bar position in the storage unit 120 based on the position specification.

[0021] An example will be described. Image data of the reinforcement state is displayed on the display means D by arbitrary control. When an image ID1 based on the image data is displayed, and a user performs an operation to specify two arbitrary points C1 and C2 in the image via an input unit or the like (not shown) as shown in FIG. 2, the specification unit 140 receives the two specified points and stores them in the storage unit 120.

[0022] The designation unit 140 specifies the area between the designated points C1 and C2 as the rebar position. Furthermore, the designation unit 140 stores in the storage unit 120 not only the line segment between the designated points, but also a rebar area including a predetermined range in the short direction perpendicular to the longitudinal direction centered on the line segment. Here, even if the designated points do not cover the entire rebar as shown in FIG. 2, the designation unit 140 may be configured to specify the rebar position including the extension line of the line segment connecting the designated points (a straight line passing through two designated points). Furthermore, if the designated points are misaligned with the rebar in the captured image, for example, if a predetermined angle is not formed between the line segment connecting the designated points and the center line of the rebar in the captured image, a warning may be output, a position correction may be suggested, or a correction process may be executed. In this example, the designation unit 140 acquires the center line of the rebar based on pixel values ​​on the line segment connecting the designated points or in the vicinity of the line segment. Furthermore, the designation unit 140 determines whether the angle or position between this center line and the line segment connecting the designated points is misaligned beyond a predetermined range. If it exceeds the range, the designation unit 140 executes the above-mentioned correction process. The designation of the rebar positions may be performed for all rebars, and the rebar diameter acquisition process described below may be performed accordingly. However, the present invention is not limited to this configuration, and may be configured to designate the rebar positions for a portion of all rebars in the image. The designation of a portion may be, for example, the designation of one rebar each in the vertical and horizontal directions. In this case, the rebar diameter acquisition process described below is also performed according to the designated rebar positions. This corresponds to aggregating the individual estimation results on the assumption that rebars of the same diameter are lined up. In addition, since the present embodiment does not perform projective transformation as in the other embodiments, a message or the like may be displayed to prompt the user to designate the rebar position of the column or row in which the rod tape is arranged.

[0023] <Specify measurement range> As shown in FIG. 2, if a reference object showing actual dimensions such as a rod tape is present in the captured image, this can be used to measure the diameter of the rebar. As shown in FIG. 3A, in this embodiment, by specifying a range of the rod tape (see FIG. 3A; symbol R), it is possible to obtain the actual dimensions corresponding to the specified range. Here, as an example, the specification unit 140 obtains a length of 500 mm in the longitudinal direction. This obtained length is converted to a length in the image based on the coordinates, pixels, etc. on the image and stored. Note that, as an example of the range specification, a range specification by a user via a pointing device, etc. is received.

[0024] Furthermore, the designation unit 140 extracts the length of the acquired actual dimension of 500 mm from the stored reinforcing bar area as the reinforcing bar diameter acquisition target range (hereinafter referred to as "partial image"). Note that the designated range is not limited to the entire length of the rod tape. For example, it is possible to designate a part of the length indicated by the rod tape as the designated range, without designating the entire range of the length of the rod tape in the image.

[0025] (Discrimination unit 150) The discrimination unit 150 discriminates the rebar diameter from the partial image by deep learning. For example, the discrimination unit 150 discriminates the rebar diameter from the partial image by an image classification model (which may be an algorithm) by deep learning. An example of an image classification model is a configuration in which a labeled data set (partial image, etc.) is input and a label corresponding to the content of the image is output. In this embodiment, a partial image of a rebar is input and a label indicating the standard of the rebar diameter is output (discriminated).

[0026] Reinforcing bars are classified according to their nominal diameter, such as "D**". In the deep learning of this embodiment, deep learning is performed in advance on a predetermined length of reinforcing bar images for each nominal diameter classification. The actual dimensions of the reinforcing bar images in this deep learning correspond to the actual dimensions of the partial images extracted above.

[0027] The discrimination unit 150 discriminates the nominal diameter classification of the partial image by discriminating the feature amount in the partial image in which a reinforcing bar region of a predetermined length is extracted, using the image classification model, etc. The feature amount corresponds to, for example, the number of nodes of the reinforcing bar for each nominal diameter classification included in a predetermined actual size range (for example, "maximum average interval of nodes in mm").

[0028] (Acquisition part 160) The acquisition unit 160 acquires the reinforcing bar diameter using the discrimination result of the discrimination unit 150. For example, when the discrimination unit 150 discriminates the nominal diameter classification, the acquisition unit 160 acquires the reinforcing bar diameter for each nominal diameter classification stored in advance in the storage unit 120 or the like according to the discrimination result. As an example, the storage unit 120 stores a numerical value of the reinforcing bar diameter in association with each discriminated nominal diameter classification.

[0029] For convenience of explanation, the processing of each unit in the above description has been described as being executed under the control of the control unit C shown in FIG. 1. The control unit C may be configured to include a single or multiple circuits, such as a central processing unit (CPU), a graphics processing unit (GPU), or an application specific integrated circuit (ASIC). The control unit C realizes its function by reading and executing, for example, a reinforcement bar measurement program stored in a memory. Note that instead of storing a program in a memory, the program may be directly built into the circuit. In this case, the control unit C realizes its function by reading and executing a program built into the circuit as the control unit C. Note that the control unit C is not limited to being configured as a single circuit, and may be configured as a single processor by combining multiple independent circuits to realize its function. Furthermore, multiple components in the embodiment (for example, at least two or more of the photographing unit 110, the storage unit 120, the image processing unit 130, the designation unit 140, the correction unit 150, and the acquisition unit 160) may be integrated into a single processor to realize its function.

[0030] [Operation] 4 is a diagram showing a process flow in this embodiment. The operation of the bar arrangement measurement system 100 will be explained below along with step numbers (S101 to S106).

[0031] (S101) When an imaging operation is performed, the imaging unit 110 in the bar arrangement measurement system 100 controls the optical system and generates imaging data based on the information input to the imaging element. The storage unit 120 temporarily stores the imaging data. The image processing unit 130 converts the imaging data into image data and displays it on the display means D.

[0032] (S102) With the image data displayed on the display means D, the user designates the rebar position via a pointing device or the like, as shown in Fig. 2. Upon receiving this designation, the designation unit 140 stores in the storage unit 120 a rebar area including the longitudinal direction of the rebar and a predetermined range in the transverse direction perpendicular to the longitudinal direction, centered on the line segment between the designated points, based on the position designation.

[0033] (S103) Next, the user specifies the range of the rod tape via the input unit or the like, and the specification unit 140 acquires the length corresponding to the actual size (e.g., 500 mm) corresponding to the specified range. The acquired length corresponding to the actual size is converted to the length in the image based on the coordinates, pixels, etc. on the image, and stored.

[0034] (S104) Furthermore, the designation unit 140 extracts, from the reinforcing bar region stored in S102, a length corresponding to the actual length of 500 mm acquired in S103 as a partial image of the reinforcing bar diameter.

[0035] (S105) The discrimination unit 150 discriminates the reinforcing bar diameter from the partial image by, for example, an image classification model using deep learning, etc. In this embodiment, a partial image of the reinforcing bar is input, and a label indicating the standard of the reinforcing bar diameter is output.

[0036] (S106) The acquisition unit 160 acquires the reinforcing bar diameter using the discrimination result of the discrimination unit 150. For example, when the discrimination unit 150 discriminates the nominal size classification, the acquisition unit 160 acquires the reinforcing bar diameter associated with each nominal size classification stored in advance in the storage unit 120 or the like according to the discrimination result.

[0037] The order of specifying the rebar position in S102 and the rod tape in S103 may be reversed.

[0038] [effect] According to this embodiment, it is possible to reduce measurement errors of rebars in the rebar measurement system. At least a part of the rebars photographed is drawn in the photographed image at an inclination different from the direction in which they are actually arranged. Therefore, there is a possibility that the deviation between the photographed image and the actual plane may become large. Furthermore, in the photographed image, the nodes and ribs of the rebars may not be constant, which may lead to erroneous measurement of the rebar diameter. In addition, since the rebars have a cylindrical shape, there is a possibility that the thickness may differ on the image depending on the distance between the photographing means and the target. In contrast, in this embodiment, the rebar diameter is not measured based on the rebars in the photographed image, for example, based on pixels, but the rebar diameter is estimated by image classification based on the characteristics of the rebars included in a predetermined range of the actual rebars. This makes it possible to eliminate the problems caused by the photographing conditions as described above and avoid erroneous measurement of the rebar diameter.

[0039] [Second embodiment] A bar arrangement measurement system 200 according to the second embodiment will be described with reference to Fig. 5 to Fig. 10. As shown in Fig. 5, the bar arrangement measurement system 200 includes a photographing unit 210, a storage unit 220, an image processing unit 230, a designation unit 240, a discrimination unit 250, an acquisition unit 260, a control unit C, and a display means D. In the following description, descriptions that overlap with those of the first embodiment will be omitted.

[0040] (Image processing unit 230) The image processor 230 in the second embodiment also converts the captured image data into image data. Furthermore, the image processor 230 converts the line of sight of the image so that the vertical and horizontal directions of the image data correspond to the vertical and horizontal directions of the reinforcing bars that are the subject of the image capture, that is, the vertical and horizontal directions when the reinforcing bars are actually installed and viewed from the vertical direction.

[0041] An example will be described. Image data of the reinforcement state is displayed on the display means D by arbitrary control. With image ID1 based on the image data displayed, when the user performs an operation to specify an area R surrounded by any four points C1 to C4 in the image via an input unit or the like (not shown) as shown in Fig. 6, the control unit C of the reinforcement measurement system 200 accepts the area R as the specified range.

[0042] The image processing unit 230 performs projective transformation on the image data of the specified range to correct the distortion. As shown in the example of Fig. 7, in the corrected image of the specified range, the directions of the vertical and horizontal rebars in the reinforcement arrangement substantially correspond to the vertical and horizontal directions of the image data. "Substantially corresponding directions" means that the center lines of the rebars in the specified range are roughly aligned with the vertical and horizontal directions of the image data. For example, this is the case when the deviation angle of each center line with respect to the vertical and horizontal directions of the image data is about ±3°.

[0043] [Operation] 8 is a diagram showing a process flow in this embodiment. The operation of the bar arrangement measurement system 200 will be explained below along with step numbers (S201 to S207).

[0044] (S201) When an imaging operation is performed, the imaging unit 210 in the bar arrangement measurement system 200 controls the optical system and generates imaging data based on the information input to the imaging element. The storage unit 220 temporarily stores the imaging data. The image processing unit 230 converts the imaging data into image data and displays it on the display means D.

[0045] (S202) With the image data displayed on the display means D, when the user performs an operation to specify an area surrounded by any four points in the image via an input unit or the like (not shown) as shown in Figure 6, the control unit C of the reinforcement bar measurement system 200 accepts that area as the specified range.

[0046] (S203) The image processing unit 230 performs projective transformation on the image data of the specified range to correct the distortion. As shown in the example of Fig. 7, in the corrected image of the specified range, the directions of the vertical and horizontal rebars in the reinforcement arrangement substantially correspond to the vertical and horizontal directions of the image data.

[0047] (S204) Furthermore, the image processing unit 230 receives a rebar position specification from the user via the input unit, etc., for the image data after projective transformation (keystone correction, etc.). When this specification is received, the specification unit 240 stores in the storage unit 220 a rebar area including a predetermined range in the short direction perpendicular to the longitudinal direction, centered on the line segment between the specified points, based on the position specification.

[0048] (S205) Upon receiving a range of the rod tape designated by the user via an input unit or the like, the designation unit 240 acquires the length corresponding to the actual size (e.g., 500 mm) corresponding to the designated range. The acquired length corresponding to the actual size is converted to the length in the image based on the coordinates, pixels, etc. on the image and stored.

[0049] Furthermore, the designation unit 240 extracts the acquired length of 500 mm of the actual dimension from the stored reinforcing bar area as a partial image of the reinforcing bar diameter.

[0050] (S206) The discrimination unit 250 discriminates the rebar diameter from the partial image by, for example, an image classification model using deep learning, etc. In this embodiment, a partial image of the rebar is input, and a label indicating the standard of the rebar diameter is output.

[0051] (S207) The acquisition unit 260 acquires the reinforcing bar diameter using the discrimination result of the discrimination unit 250. For example, when the discrimination unit 250 discriminates the nominal diameter classification, the acquisition unit 260 acquires the reinforcing bar diameter associated with each nominal diameter classification stored in advance in the storage unit 220 or the like according to the discrimination result.

[0052] [effect] According to this embodiment, it is possible to avoid mistakes in measuring rebars in the rebar measurement system. For example, by trying to align the vertical and horizontal directions of the image with the line of sight of the image by projective transformation or the like before image classification, it is possible to significantly reduce mistakes in obtaining rebar diameters by image classification. In addition, by performing projective transformation, it is possible to narrow down the learning model to the two axes of rebar arrangement, vertical and horizontal, improving detection accuracy.

[0053] In other words, when a structure such as reinforcing bars is photographed using a normal imaging method, the vertical and horizontal directions of the image tend not to correspond to the actual vertical and horizontal arrangement of the structure, even though the structure is actually aligned vertically and horizontally. By setting the four corners of the measurement area and performing a projective transformation, it is possible to correct the image to retain the information of the original structure. Furthermore, the specified position of the reinforcing bar position does not have to correspond to the column or row where the rod tape is arranged.

[0054] [Third embodiment] A reinforcement bar measurement system 300 according to the third embodiment will be described with reference to Fig. 3B and Fig. 9. As shown in Fig. 9, the reinforcement bar measurement system 300 includes a photographing unit 310, a storage unit 320, an image processing unit 330, a designation unit 340, a discrimination unit 350, an acquisition unit 360, a control unit C, and a display means D. In the following description, descriptions that overlap with those of the first embodiment will be omitted.

[0055] <Specify measurement range> The designation unit 140 in the first embodiment was configured to acquire a length of 500 mm in the longitudinal direction of the rod tape (a reference object showing actual dimensions) as an example. The designation unit 340 in the third embodiment acquires the length not only in the longitudinal direction but also in the lateral direction. In the lateral direction, a length of 30 mm, for example, is acquired (see FIG. 3B). The length of 500 mm x 30 mm is converted to a length in the image based on the coordinates, pixels, etc. on the image and stored.

[0056] The designation unit 340 also extracts the length on the image corresponding to the acquired rectangular area of ​​actual dimensions 500 mm x 30 mm from the stored reinforcing bar area as a partial image of the reinforcing bar diameter. This value of 30 mm is a value that takes into consideration the value of the maximum standard in the range of reinforcing bar standards that are considered to be frequently used, but other values ​​may be used.

[0057] [effect] According to this embodiment, it is possible to reduce measurement errors of rebars in the rebar measurement system. For example, information on the rebar diameter in the short side direction is direct information on the rebar diameter, and by utilizing information in the short side direction as well as the long side direction, it is possible to estimate the rebar diameter in a state that is more in line with the actual characteristics of the partial image of the rebar. This makes it possible to eliminate the problems caused by the shooting conditions as described above and reduce erroneous measurements of the rebar diameter.

[0058] [Fourth embodiment] A reinforcement bar measurement system 400 according to the fourth embodiment will be described with reference to Fig. 10 to Fig. 13. As shown in Fig. 10, the reinforcement bar measurement system 400 includes a photographing unit 410, a storage unit 420, an image processing unit 430, a designation unit 440, a discrimination unit 450, an acquisition unit 460, a control unit C, and a display means D. In the following description, descriptions that overlap with the first to third embodiments will be omitted.

[0059] (Specification part 440) As in the first embodiment, the designation unit 440 extracts the length of the acquired actual dimension of 500 mm from the stored reinforcing bar area as a partial image of the reinforcing bar diameter. Here, the designation unit 440 in the fourth embodiment detects and classifies rod tapes as described below to acquire the actual dimension length corresponding to the partial image.

[0060] After detecting the rod tape, the designation unit 440 classifies the arrangement pattern of the rod tape. For example, the designation unit 440 distinguishes between three types of patterns: a pattern in which one rod tape is arranged vertically or horizontally on the image, a pattern in which two rod tapes are arranged vertically and horizontally in a cross shape, and a pattern in which two rod tapes are arranged vertically and horizontally without overlapping (see FIGS. 11A to 11C).

[0061] An overview of an example of the above-mentioned processing by the designation unit 440 will be described. The designation unit 440 detects the rod tape from the image data by deep learning. That is, each is detected by an object detection model (which may be an algorithm) by deep learning.

[0062] Next, the designation unit 440 uses an image classification model (which may be an algorithm) based on deep learning to determine the classification of the rod tape arrangement pattern from the detected rod tape image (see FIG. 12). One example of an image classification model is a configuration in which a labeled data set (partial image, etc.) is input and a label corresponding to the content of the image is output. In this embodiment, an image of a rod tape is input and one of the above three label patterns is output.

[0063] When the classification of the arrangement pattern is acquired by outputting the label, the designation unit 440 uses the classification to extract four points of the circumscribing rectangle of the rod tape image. An example will be described in detail below with reference to the drawings.

[0064] A state in which two rod tapes are overlapped in a cross shape as shown in Fig. 11B will be described as an example. The designation unit 440 detects the shape of two rod tapes overlapped in a cross shape from the image data based on an object detection model using deep learning or the like.

[0065] Next, the designation unit 440 extracts an area image (rod tape image) including the detected rod tape from the captured image. Furthermore, the designation unit 440 converts the area image into a binary image expressed in binary so that it is divided into a part detected as a rod tape and other parts other than the detected part (see FIG. 12). In the example of FIG. 12, the area of ​​the detected rod tape is converted to white, and the other areas (other parts) are converted to black.

[0066] Next, the designation unit 440 determines the classification of the arrangement pattern of the rod tape from the detected rod tape image (see FIG. 12) using an image classification model or the like. In this case, a classification like that shown in FIG. 11B, in which two rod tapes are arranged overlapping each other in a cross shape, is output. This estimation will be described in detail below.

[0067] The above estimation is processed, for example, in the order shown in FIG. 13. First, the designation unit 440 performs a first classification, which labels the number of outer edges of white areas (hereinafter simply referred to as "frames") detected in the binary image of the area image including the rod tape. In this example, the number of detected frames is first classified as either one or two. Those classified as having two frames in this first classification become "patterns in which two are arranged vertically and horizontally without overlapping" as shown in FIG. 11C.

[0068] Next, the designation unit 440 performs a second classification on the patterns as shown in Figures 11A and 11B, which are determined to have one frame in the first classification. That is, the designation unit 40 designates a rectangular frame that surrounds the white area of ​​the rod tape image. This rectangular frame is designated so that it includes all the white area within its area and is adjacent or close to the white area. The reason for the rectangular frame is that the rod tape has a rectangular shape.

[0069] Next, the designation unit 440 determines the proportion of the white area within the rectangular frame, i.e., the rod tape portion, as the second classification. In this case, the cross pattern will have more black areas (less proportion of white areas) compared to a pattern of one rod tape. The designation unit 440 can classify these two patterns according to the proportion of white areas. In one example, the designation unit 440 is configured to determine the classification according to this proportion using a threshold value. If it is below the threshold value, it will be a pattern of one rod tape, and if it exceeds the threshold, it will be a cross pattern.

[0070] In any case, including the case where the classification of the cross pattern is determined as shown in Fig. 11B, the designation unit 440 specifies four points that are the four corners of the white area, i.e., the outer edge of the rod tape, for each arrangement pattern. At this time, in the case of a cross, four points are specified separately in the vertical and horizontal directions with respect to the reinforcement.

[0071] As described above, in the fourth embodiment, the rod tape is specified. In specifying the rectangular frame, the specification unit 440 specifies a rectangular frame that is in contact with the outside of the white area, and obtains the four points of the four corners. At this time, the rectangular frame tends to be slightly larger than the white area. Therefore, the specification unit 440 may be configured to reduce the rectangular frame by a predetermined ratio (see FIG. 14). This reduction ratio is, for example, about 0.5% to 2%. FIG. 15 shows an example of reducing by about 1%.

[0072] Next, a case where a cross pattern is estimated in the second classification will be described. When a cross pattern is determined, the designation unit 440 scans the rod tape image vertically and horizontally (see FIG. 16). As a result of the scan, the designation unit 440 identifies a pixel group with a high proportion of white pixels extending vertically as a vertical rod tape region. Similarly, the designation unit 440 identifies a pixel group with a high proportion of white pixels extending vertically as a vertical rod tape region.

[0073] The designation unit 440 can also determine the ratio by threshold processing or the like to specify the region. However, this configuration is not limited. For example, in this vertical scanning, the portion in the image where the vertical rod tape exists has a identifiable number of white regions compared to when other portions are scanned. Therefore, the designation unit 440 can specify the adjacent portion with a high ratio of white pixels as the vertical rod tape region when the vertical rod tape region is scanned in the vertical scanning, thereby specifying the vertical rod tape region without setting a threshold.

[0074] The specified range is not limited to the length of the rod tape. For example, it is possible to specify a part of the length of the rod tape as the specified range, instead of specifying the entire range of the length of the rod tape in the image.

[0075] Other processing in the fourth embodiment is similar to that in the other embodiments.

[0076] [Operation] Fig. 17 is a diagram showing a process flow in the fourth embodiment. The operation of the reinforcement bar measurement system 400 will be described below along with step numbers (S401 to S406). Note that, although a combination of the second embodiment and the fourth embodiment will be described below, the fourth embodiment is not limited to this, and may be an embodiment other than the second embodiment, or may be combined with another embodiment in addition to the second embodiment.

[0077] (S401) When an imaging operation is performed, the imaging unit 410 in the bar arrangement measurement system 400 controls the optical system and generates imaging data based on the information input to the imaging element. The storage unit 420 temporarily stores the imaging data. The image processing unit 430 converts the imaging data into image data and displays it on the display means D.

[0078] (S402) With the image data displayed on the display means D, when the user performs an operation to specify an area surrounded by any four points in the image via an input unit or the like (not shown) as shown in Figure 6, the control unit C of the reinforcement bar measurement system 400 accepts that area as the specified range.

[0079] (S403) The image processing unit 430 performs projective transformation on the image data of the specified range to correct the distortion. As shown in the example of Fig. 7, in the corrected image of the specified range, the directions of the vertical and horizontal rebars in the reinforcement arrangement substantially correspond to the vertical and horizontal directions of the image data.

[0080] (S404) Furthermore, the image processing unit 430 receives a rebar position specification from the user via the input unit, etc., for the image data after projective transformation (keystone correction, etc.). When this specification is received, the specification unit 440 stores in the storage unit 420 a rebar area including a predetermined range in the short direction perpendicular to the longitudinal direction, centered on the line segment between the specified points, based on the position specification.

[0081] (S405) The designation unit 440 detects each rod tape from the image data using an object detection model (which may be an algorithm) based on deep learning.

[0082] (S406) The designation unit 440 converts the region image into a binary image expressed in binary so as to be separated into a portion detected as a rod tape and a portion other than the detected portion (see FIG. 12).

[0083] (S407) The designation unit 440 performs a first classification of labeling based on the number of frames detected in the binary image of the area image including the rod tape. In this example, the number of frames detected is first classified as either 1 or 2.

[0084] (S408-1) When the number of frames is classified as two in this first classification (S407; A), designation unit 440 identifies the pattern as one in which two frames are arranged without overlapping, as shown in FIG. 11C.

[0085] (S408-2) The designation unit 440 further performs a second classification on patterns such as those shown in FIGS. 11A and 11B, in which the number of frames is determined to be one in the first classification (S407;B).

[0086] (S409) The designation unit 440 determines the ratio of the white area in the rectangular frame, i.e., the rod tape portion, to the second classification. The designation unit 440 further classifies these two patterns based on this ratio. The second classification is determined, for example, based on a threshold value.

[0087] (S410-1) If the result of this second classification is equal to or less than the threshold value (S409; C), the pattern is determined to be one rod tape pattern as shown in FIG. 11A.

[0088] (S410-2) If the threshold value is exceeded in this second classification (S409; D), it is determined to be a cross pattern as shown in FIG. 11B.

[0089] (S411) In all three cases, including the case where the classification of the cross pattern is determined as shown in Fig. 11B, the designation unit 440 specifies four points that are the four corners of the white area, i.e., the outer edge of the rod tape, for each arrangement pattern. In this case, in the case of a cross, four points are specified separately in the vertical and horizontal directions with respect to the reinforcement. In this case, the designation unit 440 may be configured to reduce the rectangular frame by a predetermined ratio (see Fig. 14).

[0090] (S412) Next, the designation unit 440 acquires the actual size (e.g., 500 mm) corresponding to the designated range. The acquired actual size is converted to a length in the image based on the coordinates on the image, pixels, etc., and stored.

[0091] (S413) Furthermore, the designation unit 440 extracts the length of 500 mm of the actual dimension acquired in S412 from the reinforcing bar area stored in S404 as a partial image of the reinforcing bar diameter.

[0092] (S414) The discrimination unit 450 discriminates the rebar diameter from the partial image by, for example, an image classification model using deep learning, etc. In this embodiment, a partial image of the rebar is input, and a label indicating the standard of the rebar diameter is output.

[0093] (S415) The acquisition unit 460 acquires the reinforcing bar diameter using the discrimination result of the discrimination unit 450. For example, when the discrimination unit 450 discriminates the nominal diameter classification, the acquisition unit 460 acquires the reinforcing bar diameter for each nominal diameter classification stored in advance in the storage unit 420 or the like according to the discrimination result.

[0094] The order of specifying the reinforcing bar position in S404 and the order of specifying the rod tape in S405 to S411 may be reversed.

[0095] [effect] According to this embodiment, it is possible to reduce errors in measuring rebars in the rebar measurement system. That is, if the rod tape is manually specified before image classification, differences will occur between the feature values ​​in the partial image and the feature values ​​in the image to be compared, which may lead to errors in obtaining the rebar diameter. In this regard, by specifying the rod tape as in the above embodiment, it is possible to properly obtain a partial image that corresponds to the actual size based on the rod tape. As a result, it is possible to avoid obtaining the erroneous rebar diameter as described above.

[0096] With the image data displayed on the display means D, when the user performs an operation to specify an area surrounded by any four points in the image via an input unit or the like (not shown) as shown in Figure 6, the control unit C of the reinforcement bar measurement system 200 accepts that area as the specified range.

[0097] [Variation 1] In the second embodiment, when an image ID1 based on image data is displayed, and the user performs an operation to specify an area R surrounded by any four points C1 to C4 in the image via an input unit or the like (not shown), the control unit C of the reinforcement bar measurement system 200 accepts the area R as the specified range, as shown in Fig. 6. In this modified example, instead of the above configuration, the following configuration is adopted.

[0098] The image processor 230X detects markers arranged in association with the reinforcement (see FIG. 18: symbols M1 to M4). The image processor 230X accepts an area R surrounded by four markers M1 to M4 as in the example of FIG. 18 as a specified range.

[0099] [Fifth embodiment] The overall configuration of a bar arrangement inspection system 500 according to the fifth embodiment will be described with reference to Fig. 19 to Fig. 21. Note that the bar arrangement inspection system in Fig. 19 is just one example, and does not exclude the inclusion of other configurations, and can be embodied in various forms.

[0100] (System Overview) As shown in FIG. 19, the reinforcing bar inspection system 500 includes a photographing unit 510, a storage unit 520, an image processing unit 530, a discrimination unit 570, a correction unit 580, a control unit C, and a display unit D. The photographing unit 510 is an imaging unit provided in, for example, a tablet terminal. The imaging data obtained by the imaging unit is stored as image data in the storage unit 520 by operating a touch panel or a save button. The image processing unit 530 converts the line of sight of the image so that the vertical and horizontal directions of the image correspond to the vertical and horizontal directions of the reinforcing bar. The discrimination unit 570 detects the reinforcing bar area from the converted image data. The correction unit 580 corrects the reinforcing bar part detected by the discrimination unit 570 and the other parts based on the statistical information of the image. Note that the image shown in the following figure is an example.

[0101] (Storage unit 520) The storage unit 520 stores imaging data, image data, binary image data, various programs, area detection models, and the like.

[0102] (Image processing unit 530) As in the second embodiment and its modified example, the image processing unit 530 converts the line of sight direction of the image so that it corresponds to the length and width directions when the actually installed reinforcing bar is viewed from the vertical direction.

[0103] An example will be described. Image data of the reinforcement state is displayed on the display means D by arbitrary control. With image ID1 based on the image data displayed, when the user performs an operation to specify an area R surrounded by any four points C1 to C4 in the image via an input unit or the like (not shown) as shown in Fig. 6, the control unit C of the reinforcement inspection system 100 accepts the area R as the specified range. Alternatively, instead of this configuration, a configuration may be used in which the area R surrounded by four markers M1 to M4 as in the example of Fig. 18 is accepted as the specified range, as in Modification 1.

[0104] (Identification unit 570) The identification unit 570 detects reinforcing bars from the image data by deep learning. The identification unit 570 detects each of the surface reinforcing bars in the image data by an object detection model (or an algorithm) by deep learning.

[0105] (revised part 580) The correction unit 580 corrects the reinforcing bar portion and the other portion detected by the identification unit 570 based on statistical information of the image. As one specific example, the correction unit 580 determines a detection error in the image based on the statistical information and executes the correction. Note that the detection described here is an example of "identification."

[0106] <Example 1: Vertical direction judgment> The correction unit 580 judges a detection error for each of the rebars detected from the image data, that is, for a vertical rebar portion that is substantially equivalent to the vertical direction in the image. For example, the correction unit 580 includes a determination unit 581, which scans the image vertically in units of a predetermined number of pixels, and determines a first pixel indicating a rebar portion for each pixel corresponding to the scanning direction (see FIG. 20). Specifically, the determination may be performed based on whether the pixel falls within one or more ranges of pixel values ​​stored in advance. In other words, the image to be corrected may be a color image, a grayscale image, or a binary image, and this determination is performed based on the pixel values ​​of the corresponding image. In other words, the first pixel is a pixel having a specific single or multiple pixel values.

[0107] The modifying unit 580 obtains the proportion of the first pixel in each of the vertical scanning directions. The modifying unit 580 also determines whether the obtained proportion exceeds a threshold. This threshold may be, for example, a value stored in advance in the storage unit 520, or may be a value set by the user through an input operation each time. When the threshold is exceeded, the modifying unit 580 may replace all pixels existing in that scanning direction with the first pixel. For example, the threshold of the proportion is 50%, and when the obtained proportion of the first pixel exceeds 50%, the modifying unit 580 replaces all pixels existing in that scanning direction with the first pixel.

[0108] In this way, the correction unit 580 determines whether the reinforcing bar portion detected from the image is a false detection or not. Note that the threshold value for the ratio of the first pixels is 45%, for example. The lower limit is 20%, and the upper limit is 55%.

[0109] <Second example: Horizontal judgment> The correction unit 580 judges a detection error for each rebar detected from the image data, for a horizontal rebar portion that substantially corresponds to the horizontal direction in the image. As described above, the identification unit 151 scans the image horizontally in units of a predetermined number of pixels, and identifies a first pixel for each pixel corresponding to the scanning direction. Specifically, the pixel may be identified based on whether it falls within one or more ranges of pixel values ​​stored in advance. In other words, the first pixel is a pixel having a specific single or multiple pixel values.

[0110] The modifying unit 580 obtains the proportion of the first pixel in each of the horizontal scanning directions. The modifying unit 580 also determines whether the obtained proportion exceeds a threshold value. This threshold value may be, for example, a value stored in advance in the storage unit 520, or may be a value set by the user through an input operation each time. When the threshold value is exceeded, the modifying unit 580 may replace all pixels existing in that scanning direction with the first pixel. Examples of the proportion threshold value are the same as those described above.

[0111] <Third example: Diagonal direction judgment> Even if the direction of the reinforcing bar part in the image after detection by projective transformation does not correspond to the vertical and horizontal directions of the image, the correction unit 580 may identify the extension direction of the center line of the reinforcing bar part, obtain the proportion of the first pixel in that direction, determine the detection error, and convert it to the first pixel in the scanning direction.

[0112] [Operation] 21 is a diagram showing the flow of processing in this embodiment. The operation of the bar arrangement measurement system 500 will be explained below along with step numbers (S501 to S507).

[0113] (S501) When an imaging operation is performed, the imaging unit 510 in the bar arrangement inspection system 500 controls the optical system and generates imaging data based on the information input to the imaging element. The storage unit 520 temporarily stores the imaging data. The image processing unit 530 converts the imaging data into image data and displays it on the display means D.

[0114] (S502) With the image data displayed on the display means D, when the user performs an operation to specify an area surrounded by any four points in the image via an input unit or the like (not shown) as shown in FIG. 6, the control unit C of the reinforcement inspection system 500 accepts the area as the specified range.

[0115] (S503) The image processing unit 530 performs projective transformation on the image data of the specified range to correct the distortion. As shown in the example of Fig. 7, in the corrected image of the specified range, the directions of the vertical and horizontal rebars in the reinforcement arrangement substantially correspond to the vertical and horizontal directions of the image data.

[0116] (S504) The identification unit 540 detects reinforcing bars from the image data using deep learning.

[0117] (S505) A determination unit 581 in the correction unit 580 vertically scans the image in units of a predetermined number of pixels, and determines a first pixel that indicates a reinforcing bar portion for each pixel corresponding to the scanning direction (see FIG. 20).

[0118] (S506) The correction unit 550 obtains, for example, the proportion of the first pixel in each of the vertical scanning directions.

[0119] (S507) The modification unit 580 then determines whether the calculated ratio exceeds a threshold value. If the ratio does not exceed the threshold value (S507; No), the modification unit 580 ends the process in the scanning direction. Alternatively, as the reverse process of the next step S508, the modification unit 580 may be configured to replace all pixels in the scanning direction with pixel values ​​other than that of the first pixel.

[0120] (S508) If it is determined in S507 that the ratio exceeds the threshold value (S507; Yes), the modifying unit 580 replaces all pixels existing in that scanning direction with the first pixel.

[0121] Similarly, for the horizontal operation, if the ratio does not exceed the threshold (S507; No), the modifying unit 580 ends the processing in that scanning direction. If the ratio exceeds the threshold (S507; Yes), the modifying unit 580 replaces all pixels in that horizontal scanning direction with the first pixel. In this case, the modifying unit 580 can also be configured to replace all pixels in that scanning direction with pixel values ​​other than the first pixel as the reverse process.

[0122] The correction unit 580 performs this scanning and ratio determination on the entire image in each of the vertical and horizontal directions at a predetermined interval, and the replacement process with the first pixel. Then, the measurement process is performed.

[0123] [effect] According to this embodiment, it is possible to reduce the number of errors in detecting reinforcing bars in a reinforcing bar inspection system. For example, before detecting reinforcing bars in an image using deep learning, it is possible to significantly reduce false detections by aligning the vertical and horizontal directions of the image using projective transformation or the like with the line of sight of the image. For example, this makes it easier to detect the reinforcement bars on the surface when reinforcement bars are arranged in several overlapping layers. In addition, by performing projective transformation, it is possible to narrow down the learning model to the two axes of reinforcing bar arrangement, vertical and horizontal, improving detection accuracy.

[0124] In other words, when a structure such as reinforcement is photographed using normal imaging means, the vertical and horizontal directions of the image tend not to correspond to those of the actual structure, even though it is aligned vertically and horizontally. By setting the four corners of the measurement area and performing a projective transformation, it is possible to correct the image to retain the information of the original structure.

[0125] In this embodiment, after rebar detection, the detected image is scanned in a predetermined direction to identify a first pixel that indicates a rebar portion. When the first pixel is identified, the ratio of the first pixel in the scanning direction is identified, and if the ratio exceeds a predetermined ratio, all pixels in the scanning direction are replaced with the first pixel. This makes it possible to determine that a rebar portion in the scanning direction is a false positive if the ratio is below a predetermined ratio. Furthermore, it is possible to complement pixels that are actually rebar portions but could not be detected as rebar portions due to detection errors.

[0126] The effects of the above-described embodiment make it possible to prevent the need for re-photographing or an increase in manual image processing operations, and to reduce the labor required for reinforcement inspection work compared to conventional reinforcement inspection systems.

[0127] [Variation 2] The overall configuration of a bar arrangement inspection system 500 according to a modified example (modified example 2) of the fifth embodiment will be described with reference to Figs. 22 to 27. In modified example 2, an image in which a reinforcing bar portion is detected by a discrimination unit 570 is converted into a binary image. The details will be described below.

[0128] (Correction part 580x; binary image conversion) The correction unit 580x converts the image into a binary image expressed in binary so that the reinforcing bar portion in the image detected by the discrimination unit 570 from the captured image is separated from the non-relevant portion. In a modified example of the image data as shown in FIG. 7, the correction unit 580x converts the image into a binary image. For example, as shown in FIG. 22, the correction unit 580x converts the detected reinforcing bar area to white and the other area to black. That is, the discrimination unit 570 detects the reinforcing bar area from the image by deep learning, and converts the pixels indicating the reinforcing bar area to white. The correction unit 580x also converts the pixels indicating the area other than the reinforcing bar area in the image to black. The binary image is generated in this manner.

[0129] (Modification 580x; Scan) The correction unit 580x identifies the extension direction of the rebar area by any detection method such as edge detection, line segment detection, or center line detection for the rebar part. The correction unit 580x judges the detection error along the extension direction of the identified rebar part. For example, the correction unit 580x includes a specification unit 581, which scans for each extension direction in units of a predetermined number of pixels, and specifies a first pixel indicating the rebar part for each pixel corresponding to the scanning direction (see FIG. 23). Specifically, the pixel is specified based on whether it falls within a range of one pixel value stored in advance. As a first example, the vertical direction judgment is performed as follows. The correction unit 580x judges the detection error for the vertical rebar part that is substantially equivalent to the vertical direction in the white pixel area WPA (black pixel area when black pixels are considered to be the rebar area) in the binary image from which noise has been removed. For example, correction portion 580x includes identification portion 451, and identification portion 581x scans the image vertically in units of a predetermined number of pixels (see FIG. 24), and determines the proportion of the first pixel indicating the rebar portion for each pixel corresponding to the scanning direction.

[0130] For pixels between each detected reinforcing bar region, the scanning direction is determined based on the extension direction of the reinforcing bar regions at both ends of the adjacent reinforcing bar regions as shown in Fig. 23. For pixels between each reinforcing bar region, scanning is performed according to the determined scanning direction.

[0131] The correction unit 580x obtains the proportion of the first pixel in each scanning direction. The correction unit 580x also judges whether the obtained proportion exceeds a threshold. This judgment is the same as in the fifth embodiment (see FIG. 25). In this way, the correction unit 580x judges whether the reinforcing bar portion detected from the image is a false detection or not. Note that 45% is an example of the threshold for the proportion of the first pixel. The lower limit is 20% and the upper limit is 55%.

[0132] As a second example, the correcting unit 580x performs the horizontal direction determination as follows. The correcting unit 580x determines a detection error for a horizontal rebar portion that substantially corresponds to the horizontal direction within a white pixel region (a black pixel region when black pixels are regarded as a rebar region) in a binary image from which noise has been removed. For example, the correcting unit 580x includes a specifying unit 581x, which scans the image in the horizontal direction in units of a predetermined number of pixels (see FIG. 26), and determines the proportion of the first pixel indicating the rebar portion for each pixel corresponding to the scanning direction.

[0133] Furthermore, the correction unit 580x judges whether the calculated ratio exceeds a threshold value. This threshold value may be, for example, a value stored in advance in the storage unit 520, or may be a value set by the user through an input operation each time. When the threshold value is exceeded, the correction unit 580x may replace all pixels existing in the scanning direction with the first pixel (see FIG. 27). The ratio threshold value is, for example, 50% (see FIG. 25).

[0134] [effect] According to this embodiment, it is possible to reduce the number of reinforcing bar detection errors in a reinforcing bar inspection system. For example, after detecting reinforcing bar parts in an image using deep learning, the image is converted into a binary image of reinforcing bar areas and other areas. Scanning is performed based on this binary image.

[0135] In other words, by converting into a binary image, an image is generated in which the reinforcing bar area is extracted before the subsequent detection error determination and completion, making it possible to improve the accuracy of detection error determination and completion.

[0136] In this embodiment, after rebar detection, the detected image is scanned in a predetermined direction to identify a first pixel that indicates a rebar portion. When the first pixel is identified, the ratio of the first pixel in the scanning direction is identified, and if the ratio exceeds a predetermined ratio, all pixels in the scanning direction are replaced with the first pixel. This makes it possible to determine that a rebar portion in the scanning direction is a false positive if the ratio is below a predetermined ratio. Furthermore, it is possible to complement pixels that are actually rebar portions but could not be detected as rebar portions due to detection errors.

[0137] The effects of the above-described embodiment make it possible to prevent the need for re-photographing or an increase in manual image processing operations, and to reduce the labor required for reinforcement inspection work compared to conventional reinforcement inspection systems.

[0138] [Combination of each embodiment] The above-mentioned embodiments and their modifications can be combined in any way. For example, according to a combination of the modification of the second embodiment, the fourth embodiment, and the fifth embodiment, after the image data is acquired by the photographing unit, the present system is involved in each process, including projective transformation (keystone correction, etc.) in the modification of the second embodiment (modification 1), rebar detection in the fifth embodiment and its modifications (modification 2), rod tape detection in the fourth embodiment, and measurement of the rebar diameter. This makes it possible to achieve significant labor savings. Furthermore, it is possible to avoid human error, and by preventing re-doing of rebar inspection and measurement due to human error, it is possible to perform many inspections in a short period of time.

[0139] Although the embodiments of the present invention have been described, the above embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their modifications are included in the scope and spirit of the invention, and are included in the invention and its equivalents described in the claims. [Explanation of symbols]

[0140] 100, 200, 300, 400, 500 Reinforcement Measuring System 110, 210, 310, 410, 510 Photography Department 120, 220, 320, 420, 520 storage section 130, 220, 330, 430, 530 Image processing unit 140, 240, 340, 440 designated part 150, 250, 350, 450 discrimination section 160, 260, 360, 460 acquisition part 570 Identification Unit 580, 580x correction section 581, 581x specific part C control section D Display means

Claims

1. A position storage unit that performs a process of identifying each reinforcing bar position along at least the longitudinal direction of the reinforcing bar in a captured image including a reference object showing actual size, and at least temporarily stores each identified reinforcing bar position; A designation unit that designates a range of a predetermined length based on the actual dimensions shown on the reference object for each reinforcing bar area shown at each reinforcing bar position; A discrimination unit that discriminates which classification a partial image within the range designated by the designation unit belongs to according to a characteristic discrimination element for each classification of reinforcing bars; An acquisition unit that acquires a reinforcing bar diameter using the discrimination result of the discrimination unit; A reinforcement measurement system equipped with the above.

2. The designation unit designates the range based on the predetermined length in the image and a predetermined width in a direction perpendicular to the length direction, based on a reference object in the captured image.

2. The reinforcing bar measuring system according to claim 1 .

3. The present invention further includes an image processing unit that converts the line of sight of the image based on the captured image so that the vertical and horizontal directions of the image correspond to the vertical and horizontal directions of the reinforcement.

2. The reinforcing bar measuring system according to claim 1 .

4. The discrimination unit discriminates the classification by image classification using machine learning.

2. The reinforcing bar measuring system according to claim 1 .

5. The designation unit is Detecting reference objects in the captured image and classifying their arrangement patterns; Detecting feature points of a reference object for each of the classifications; By detecting these characteristic points, the range of the reference object is identified.

3. The reinforcing bar measuring system according to claim 2 .

6. A method for measuring reinforcement, comprising the steps of: A process of performing a process of identifying each reinforcing bar position along at least the longitudinal direction of the reinforcing bar in a captured image including a reference object showing actual size, and at least temporarily storing each identified reinforcing bar position; A step of designating a range of a predetermined length for each reinforcing bar area shown at each reinforcing bar position based on the actual dimensions shown on the reference object; A step of determining which classification a partial image within the range designated by the designation unit belongs to according to a characteristic identification element for each classification of reinforcing bars; Obtaining a reinforcing bar diameter using the discrimination result of the discrimination unit; A reinforcement measurement method comprising the steps of:

Images