Location measuring device, location measuring method, and location measuring program

Abstract

To accurately pinpoint the position of the edges of the object being inspected. [Solution] The position measuring device includes: a division unit that divides the point cloud data of a region corresponding to the end of the object to be inspected into a first region, which is a region obtained by dividing it into a first region at a predetermined first interval in the direction of the central axis; a first point cloud data deletion unit that deletes the point cloud data in the first region when the number of point cloud data in the first region falls below a threshold; a second point cloud data deletion unit that deletes the point cloud data remaining after deletion by the first point cloud data deletion unit when the distance between adjacent point cloud data in the direction of the central axis is greater than or equal to a predetermined distance; and an end identification unit that performs plane fitting on the point cloud data remaining after deletion by the second point cloud data deletion unit to identify the position of the end of the object to be inspected.

Description

Technical Field

[0001] The present invention relates to a position measuring device, a position measuring method, and a position measuring program.

Background Art

[0002] Conventionally, steel pipes have been transported by attaching hanging fixtures or the like to predetermined locations on the steel pipes and passing wires through the hanging fixtures. Therefore, it is necessary to confirm that the hanging fixtures are attached to the predetermined locations, and the distance between the hanging fixtures and the end faces of the steel pipes has been measured manually. However, due to measurement variations among measurers and misreading of measurement values, automation has been desired. As an example of such an automated process, in a method of detecting accessories such as hanging fixtures using a matching process, a technique is known in which the matching process is performed with high accuracy by including the surface of the steel pipe in addition to the accessories, and the attachment position of the accessories is specified (for example, Patent Document 1). The matching process in Patent Document 1 is a process of determining the position on the steel pipe where the object corresponding to the surface shape model exists by matching the point cloud data of the steel pipe acquired by a three-dimensional shape meter (laser scanner) with the surface shape model.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] The technology described in Patent Document 1 above is a method for detecting accessories and identifying their mounting positions, assuming that the ends of the steel pipe's point cloud data are accurately determined. Even when using the technology described in Patent Document 1, it is difficult to identify the mounting positions of accessories if the ends of the steel pipe's point cloud data cannot be accurately determined. The method is for detecting accessories, and does not disclose how to accurately measure the position of the ends.

[0005] The present invention has been made in view of the above circumstances, and provides a position measuring device, a position measuring method, and a position measuring program that can accurately identify the position of the end of an object to be inspected. [Means for solving the problem]

[0006] To achieve the above objective, a position measuring device according to a first aspect of the present invention is a position measuring device for determining the position of the end of an object to be inspected, comprising: a 3D shape acquisition unit that measures the surface shape of the object to be inspected and acquires point cloud data of the object to be inspected; a central axis calculation unit that uses a model that mimics the outer shape of the object to be inspected to fit the point cloud data and calculates the direction of the central axis in the point cloud data based on the central axis of the obtained model; a division unit that divides the point cloud data of the region corresponding to the end of the object to be inspected into predetermined first intervals in the direction of the central axis, and acquires point cloud data included in each first region, which is a region obtained by dividing into each first interval; and The system includes: a first point cloud data deletion unit that deletes point cloud data in the first region when the number of point cloud data in the first region falls below a threshold determined using the number of point cloud data in the second region, which is a region extracted at a second interval in the direction of the central axis in the region corresponding to the body of the object under inspection; a second point cloud data deletion unit that deletes point cloud data in the point cloud data remaining after deletion by the first point cloud data deletion unit when the distance between adjacent point cloud data in the direction of the central axis is greater than or equal to a predetermined distance; and an end identification unit that performs plane fitting on the point cloud data remaining after deletion by the second point cloud data deletion unit to identify the position of the end of the object under inspection.

[0007] According to the position measuring device of the first aspect of the present invention, a three-dimensional shape measuring unit measures the surface shape of the object to be inspected and acquires point cloud data of the object to be inspected. A central axis calculation unit fits a model that mimics the outer shape of the object to be inspected to the point cloud data and calculates the direction of the central axis in the point cloud data based on the central axis of the obtained model. A division unit divides the point cloud data of the region corresponding to the end of the object to be inspected at predetermined first intervals in the direction of the central axis, and acquires point cloud data for each first region obtained by dividing at each first interval. A first point cloud data deletion unit deletes the point cloud data in the first region if the number of point cloud data in the first region falls below a threshold determined using the number of point cloud data in the second region, which is a region extracted at second intervals in the direction of the central axis in the region corresponding to the body of the object to be inspected. The second point cloud data deletion unit deletes point cloud data from the point cloud data remaining after deletion by the first point cloud data deletion unit if the distance between adjacent point cloud data in the direction of the central axis is greater than a predetermined distance. The end identification unit performs plane fitting on the point cloud data remaining after deletion by the second point cloud data deletion unit to identify the position of the end of the object under inspection. This allows for accurate identification of the position of the end of the object under inspection.

[0008] A position measuring device according to a second aspect of the present invention, in a position measuring device according to a first aspect, the threshold is determined using the number of point cloud data included in the second region, the length of the first interval, and the length of the second interval, wherein the length of the second interval is longer than the length of the first interval.

[0009] In the position measuring device according to the third aspect of the present invention, the threshold value is determined according to the following formula, in the position measuring device according to the second aspect.

[0010] a×L×x / y

[0011] Note that a is a coefficient, x is the length of the first interval, y is the length of the second interval, and L is the number of point cloud data points included in the second region.

[0012] In the position measuring device according to the fourth aspect of the present invention, a is 0.7 to 1, as in the position measuring device according to the third aspect.

[0013] A fifth aspect of the present invention is a position measurement method for determining the position of an end of an object to be inspected, comprising: measuring the surface shape of the object to be inspected; acquiring point cloud data of the object to be inspected; fitting a model that mimics the outer shape of the object to be inspected to the point cloud data; calculating the direction of the central axis in the point cloud data based on the central axis of the obtained model; dividing the point cloud data of the region corresponding to the end of the object to be inspected into predetermined first intervals in the direction of the central axis; and acquiring the point cloud data included in each first region, which is a region obtained by dividing into each first interval. The computer then performs a planar fitting on the deleted point cloud data to determine the position of the end of the object under inspection. This is done when the number of point cloud data points in the first region falls below a threshold determined using the number of point cloud data points in the second region, which is a region extracted at a second interval in the direction of the central axis in the region corresponding to the body of the object under inspection. If the distance between adjacent point cloud data points in the direction of the central axis is greater than or equal to a predetermined distance, the computer deletes the corresponding point cloud data.

[0014] A position measurement program according to a sixth aspect of the present invention measures the surface shape of an object to be inspected, acquires point cloud data of the object to be inspected, fits the point cloud data to a model that mimics the outer shape of the object to be inspected, calculates the direction of the central axis in the point cloud data based on the central axis of the obtained model, divides the point cloud data of the region corresponding to the end of the object to be inspected into predetermined first intervals in the direction of the central axis, acquires the point cloud data included in each first region obtained by dividing into each first interval, and the point cloud data included in the first region This position measurement program causes a computer to perform the following actions: if the number of points in the first region falls below a threshold determined using the number of point cloud data points included in a second region, which is a region extracted at a second interval in the direction of the central axis in the region corresponding to the torso of the object under inspection, the program deletes the point cloud data in the first region; if the distance in the direction of the central axis of adjacent point cloud data points after deletion is greater than or equal to a predetermined distance, the program deletes the corresponding point cloud data; and then performs plane fitting on the deleted point cloud data to determine the position of the end of the object under inspection. [Effects of the Invention]

[0015] According to the present invention, the position of the end of the object to be inspected can be accurately determined. [Brief explanation of the drawing]

[0016] [Figure 1] This is a diagram showing the schematic configuration of a position measuring device. [Figure 2] This is a diagram showing the configuration of the computing unit. [Figure 3] (A) A diagram illustrating a method for splitting 3D point cloud data, and (B) A diagram illustrating a method for removing noise. [Figure 4] This is a flowchart of the position measurement process. [Figure 5] This figure shows the data after noise reduction processing in the example. [Figure 6]This is a diagram showing the data after noise removal processing in an embodiment.

[0017] Hereinafter, while appropriately referring to the accompanying drawings, an embodiment of the present invention will be described by taking the case where the inspection object body is a spiral steel pipe and the accessory is a suspension fitting as an example.

[0018] <Overview of this embodiment> Accessories such as suspension fittings are attached to the spiral steel pipe. Although the operator welds and attaches these accessories based on the drawing, another operator is checking whether the attachment position and posture of the accessories are within the tolerance range of the drawing. Currently, measurement is performed using a convexity or the like, and the results are input into the system, but incorrect results may be input.

[0019] To address this problem, three-dimensional shape measurement (laser scanner) is used to obtain point cloud data, and point cloud data processing is performed to measure the attachment position of the accessory. Since the attachment position of the accessory is output as the distance from the pipe end, the detection accuracy of the pipe end is directly related to the attachment position accuracy of the accessory. However, when measuring a steel pipe with strong mirror properties using a laser scanner, the intensity of diffused light decreases and the variation in point cloud data increases. Therefore, in order to perform accurate pipe end detection, noise data removal becomes an issue. As a countermeasure, a method of applying a diffusion spray and wiping it off after measurement can be considered, but due to uneven coating or the like, noise data cannot be completely removed. In this embodiment, in order to detect the pipe end with high accuracy, a processing method for removing noise data at the pipe end is used.

[0020] <System configuration> FIG. 1 is a diagram schematically showing the schematic configuration of a position measuring device according to an embodiment of the present invention.

[0021] As shown in Figure 1, the position measuring device 100 according to this embodiment is a device for inspecting the mounting position of an accessory (hanging bracket) S2 attached to an object to be inspected (spiral steel pipe with hanging bracket) S, which comprises an object to be inspected body (spiral steel pipe) S1 and an accessory (hanging bracket) S2 attached to the outer surface of the object to be inspected body S1 so as to protrude in the normal direction. Generally, spiral steel pipes have an outer diameter of up to 2500 mm and a length of up to 70 m. Also, generally, a pair of hanging brackets are attached radially opposite to each other at a position about 1 to 2 m away from the end face of the spiral steel pipe.

[0022] The position measuring device 100 according to this embodiment comprises an optical three-dimensional shape measuring device 1 and a calculation device 2.

[0023] The 3D shape measuring device 1 is a device that measures the surface shape of an object S under inspection by projecting and receiving light onto the object S under inspection. Specifically, the 3D shape measuring device 1 of this embodiment is a phase-shift type 3D shape measuring device that measures the distance to the surface of the object S under inspection based on the phase difference between the projected light (such as laser light) and the light reflected and received by the object S under inspection, and measures the surface shape of the object S under inspection by scanning the direction of light projection. Because the phase-shift type 3D shape measuring device has a wide measurement range, it is possible to acquire 3D point cloud data relatively easily even for large objects S under inspection, such as when the object S1 under inspection is a spiral steel pipe. The specific configuration of the phase-shift type 3D shape measuring device is publicly known, so a detailed explanation is omitted here. As the phase-shift type 3D shape measuring device 1, for example, the Leica Geosystems laser scanner "Scan Station" or the FARO laser scanner "Focus Laser Scanner" can be suitably used.

[0024] However, the 3D shape measuring device 1 used in the present invention is not necessarily limited to a phase-shift type 3D shape measuring device. Depending on the measurement environment, it is also possible to apply a TOF (Time of Flight) type, stereo camera type, or pattern projection type 3D shape measuring device.

[0025] In this embodiment, as shown in Figure 1, a single 3D shape measuring device 1 is moved to a total of four positions: a position where the end face SE of the object to be inspected S is approximately at the center of the measurement range (positions indicated by reference numerals 1a and 1b), and a position where the accessory S2 of the object to be inspected S is approximately at the center of the measurement range (positions indicated by reference numerals 1c and 1d). The surface shape of the object to be inspected S is measured at each of these positions. At least the measurement ranges of the 3D shape measuring device 1 at adjacent positions in the longitudinal direction (left-right direction in Figure 1) of the object to be inspected S overlap. That is, the measurement range of the 3D shape measuring device 1 at position 1a and the measurement range of the 3D shape measuring device 1 at position 1c overlap. Also, the measurement range of the 3D shape measuring device 1 at position 1b and the measurement range of the 3D shape measuring device 1 at position 1d overlap. Then, before measuring the surface shape of the object S under inspection, a calibration piece (for example, a reference sphere) is placed in the overlapping portion, and the position of the calibration piece is measured by the 3D shape measuring device 1 at each of the positions 1a to 1d.

[0026] <Configuration of the arithmetic unit> Next, we will describe the configuration of the arithmetic unit. Figure 2 shows the functional configuration of the arithmetic unit 2.

[0027] As shown in Figure 2, the system includes a 3D shape acquisition unit 110, a central axis calculation unit 112, a division unit 114, a first point cloud data deletion unit 116, a second point cloud data deletion unit 118, an end identification unit 120, and an accessory identification unit 122.

[0028] The 3D shape acquisition unit 110 acquires 3D point cloud data of the object to be inspected, obtained by measuring the surface shape of the object to be inspected.

[0029] Specifically, the arithmetic unit 2 receives input of the surface shape of the object under inspection S measured by the 3D shape measuring device 1 at each position 1a to 1d, and the position of the calibration piece measured by the 3D shape measuring device 1 at each position 1a to 1d. The 3D shape acquisition unit 110 then synthesizes the surface shape of the object under inspection S measured by the 3D shape measuring device 1 at position 1a and the surface shape of the object under inspection S measured by the 3D shape measuring device 1 at position 1c, based on the position of the calibration piece measured by the 3D shape measuring device 1 at position 1a and the position of the calibration piece measured by the 3D shape measuring device 1 at position 1c, so that the positions of these calibration pieces match. This synthesizes the surface shape of the object under inspection S, including the range from the end face SE of the object under inspection S to the accessory S2 (the upper accessory S2 in Figure 1), and generates (acquires) 3D point cloud data of the surface of the object under inspection S. Similarly, the 3D shape acquisition unit 110, based on the position of the calibration piece measured by the 3D shape measuring device 1 at position 1b and the position of the calibration piece measured by the 3D shape measuring device 1 at position 1d adjacent to position 1b, synthesizes the surface shape of the object under inspection S measured by the 3D shape measuring device 1 at position 1b and the surface shape of the object under inspection S measured by the 3D shape measuring device 1 at position 1d, so that the positions of these calibration pieces match, and generates (acquires) 3D point cloud data of the surface of the object under inspection S, including the range from the end face SE of the object under inspection S to the accessory S2 (the lower accessory S2 in Figure 1).

[0030] The central axis calculation unit 112 processes the 3D point cloud data acquired by the 3D shape acquisition unit 110 to fit the point cloud data using a model that mimics the external shape of the object under inspection, and calculates the direction of the central axis in the 3D point cloud data based on the central axis of the obtained model. In other words, the central axis calculation unit 112 uses a model that mimics the external shape of the object under inspection, enlarges or reduces it in the length and diameter directions to match it with the 3D point cloud data, and calculates the direction of the central axis in the 3D point cloud data based on the central axis of the obtained model. Here, the central axis refers to the axis passing through the center of a cross-section perpendicular to the longitudinal direction of the object under inspection or the model.

[0031] Here, the computing device 2 has a model pre-configured that mimics the external shape of the object S under inspection. Specifically, the model that mimics the external shape is a model that replicates a shape similar to the general external shape of the object under inspection, such as a cylinder or a rectangular prism. In other words, in this embodiment, a cylindrical model that mimics the external shape of a spiral steel pipe is used, so the axis perpendicular to the circular cross-section is set as the central axis of the model.

[0032] In this embodiment, the central axis calculation unit 112 moves the 3D point cloud data acquired by the 3D shape acquisition unit 110 so that its central axis aligns with one of the orthogonal coordinate axes, such as the X-axis. More specifically, the central axis calculation unit 112 performs a fitting process to fit the 3D point cloud data using a cylinder, which is a model that mimics the outer shape of the object S under inspection. The central axis calculation unit 112 then translates and rotates the cylinder and the 3D point cloud data so that the central axis of the fitted cylinder aligns with one of the orthogonal coordinate axes, such as the X-axis. The central axis of the cylinder is then calculated as the direction of the central axis in the 3D point cloud data.

[0033] The division unit 114 divides the 3D point cloud data of the region that is thought to be the edge of the object under inspection from the 3D point cloud data at predetermined first intervals in the direction of the central axis, and acquires the point cloud data contained in each first region, which is a region obtained by dividing at each first interval.

[0034] Specifically, the 3D point cloud data is extracted using the result of fitting a model that mimics the outer shape of the object under inspection, obtained by the central axis calculation unit 112, to the 3D point cloud data (the result of cylindrical fitting). Then, as shown in Figure 3(A), the object is divided into predetermined first intervals (for example, x (mm)) in the direction of the central axis, and the 3D point cloud data contained in each first region, which is the region obtained by dividing the object into first intervals, is acquired.

[0035] The first point cloud data deletion unit 116 deletes the corresponding 3D point cloud data in the first region if the number of 3D point cloud data in the first region falls below a threshold determined using the number of point cloud data in the second region, which is a region extracted at a second interval in the direction of the central axis in the region corresponding to the torso of the object being inspected.

[0036] Here, the threshold is determined using the number of 3D point cloud data points included in the second region, the length of the first interval, and the length of the second interval, where the length of the second interval is longer than the length of the first interval. Specifically, the threshold is determined according to the following formula.

[0037] a×L×x / y

[0038] Note that a is a coefficient, x is the length of the first interval, y is the length of the second interval, and L is the number of 3D point cloud data points included in the second region.

[0039] Thus, the first point cloud data deletion unit 116, based on the result of dividing the 3D point cloud data for each first region by the division unit 114, determines that the number of 3D point cloud data in the first region is less than or equal to a threshold, and removes the corresponding 3D point cloud data in the first region, as noise that is not at the end of the object being inspected. At this time, the number of data points used as the threshold varies depending on the measurement conditions, so as shown in Figure 3(A) above, the threshold is set based on the number of 3D point cloud data in the second region set outside the region that is thought to be the end of the object being inspected. Specifically, the threshold is determined using the above formula. If the coefficient a is too large, non-noise parts will be removed (especially in steel pipes with beveled ends). On the other hand, if the coefficient a is too low, noise will remain. For this reason, it is necessary to set the coefficient a within an appropriate range. For example, the value of the coefficient a should be 0.7 to 1.

[0040] The second point cloud data deletion unit 118 deletes 3D point cloud data from the 3D point cloud data remaining after deletion by the first point cloud data deletion unit 116 if the distance in the direction of the central axis of adjacent point cloud data is greater than a predetermined distance.

[0041] Specifically, as shown in Figure 3(B), in the first region remaining after deletion by the first point cloud data deletion unit 116, if the distance in the direction of the central axis of adjacent point cloud data is greater than or equal to a predetermined distance (for example, b (mm)), the 3D point cloud data corresponding to the area outside the edge (the elliptical portion in Figure 3(B)) is deleted.

[0042] More specifically, the 3D point cloud data contained in the first region closest to the end of the pipe is projected along the central axis. After projection, if the distance between adjacent points is greater than or equal to a predetermined distance, the 3D point cloud data consisting of the outer point and the point further out of that adjacent point is removed as noise. The predetermined distance can be set according to the desired detection accuracy; for example, if you want to detect with an accuracy of 0.1 mm, you should set the predetermined distance to 0.1 mm.

[0043] The end-specification unit 120 performs plane fitting on the 3D point cloud data after deletion by the second point cloud data deletion unit 118 to identify the position of the end of the object under inspection.

[0044] Specifically, after deletion by the second point cloud data deletion unit 118, plane fitting is performed on the 3D point cloud data of the remaining first region to identify the position of the plane as the position of the end of the object under inspection.

[0045] The accessory identification unit 122 identifies the distance from the end of the object under inspection as the mounting position of the accessory.

[0046] Specifically, the accessory identification unit 122 accurately matches the two coordinate systems by aligning the position of the plane fitted by the end identification unit 120 with the position of the surface shape model of the end face of the body S1 under inspection (for example, by translating the fitted plane to make them match).

[0047] Then, the accessory identification unit 122 extracts partial 3D point cloud data of the surface of the object under inspection S from the 3D point cloud data of the surface of the object under inspection S, based on the position of the accessory S2 in the surface shape model of the object under inspection S with a coordinate system that matches, and which may include the accessory S2 and a part of the main body S1 of the object under inspection S.

[0048] The accessory identification unit 122 then matches the extracted partial 3D point cloud data of the surface of the object under inspection S with a partial surface shape model of the object under inspection S that includes the accessory S2 and a part of the main body S1 of the object under inspection.

[0049] More specifically, in this embodiment, the computing device 2 matches the partial 3D point cloud data of the surface of the object under inspection S with the partial surface shape model of the surface of the object under inspection S, using the PPF features of the partial 3D point cloud data of the surface of the object under inspection S and the PPF features of the partial surface shape model of the surface of the object under inspection S.

[0050] Then, based on the matching results described above, the accessory identification unit 122 identifies the mounting position of the accessory S2 in the 3D point cloud data of the surface of the object S under inspection. Since the end is accurately measured using the position measuring device 100, it is possible to accurately identify the mounting position of the accessory S2.

[0051] The computing unit 2 consists of, for example, a computer on which programs or applications that perform the processing of each of the above parts are installed. Specifically, the computing unit 2 can be configured by implementing a known point cloud processing library on a computer, such as the open-source "PCL (Point Cloud Library)" or MVTec's "HALCON". The above point cloud processing libraries can handle not only point cloud data but also surface data (data composed of cylinders, planes, triangular meshes, etc.), and can perform various calculations related to point cloud data and surface data, such as preprocessing such as smoothing and thinning, extraction of point cloud data based on coordinates and distances, coordinate transformation, matching, fitting, dimensional measurement of point cloud data, and generation of three-dimensional surfaces.

[0052] Next, the position measurement process performed by the computing device 2 will be explained with reference to the flowchart shown in Figure 4. Here, it is assumed that 3D point cloud data of the surface of the object under inspection S has been acquired by measuring the surface shape of the object under inspection S using the 3D shape measuring device 1. Specifically, the 3D shape measuring device 1 is moved to four positions 1a to 1d, and the surface shape of the object under inspection S and the surface shape of the calibration piece are measured at each position. These measurement results are input to and stored in the computing device 2 via a storage medium such as an SD card.

[0053] In step S100, the 3D shape acquisition unit 110 acquires 3D point cloud data of the object to be inspected, obtained by measuring the surface shape of the object to be inspected.

[0054] In step S102, the central axis calculation unit 112 performs a process to fit the 3D point cloud data acquired by the 3D shape acquisition unit 110 to the point cloud data using a model that mimics the outer shape of the object to be inspected, and calculates the direction of the central axis in the 3D point cloud data based on the central axis of the obtained model.

[0055] In step S104, the division unit 114 extracts 3D point cloud data from the 3D point cloud data, specifically the region that appears to be the edge of the object being inspected.

[0056] In step S106, the division unit 114 divides the extracted 3D point cloud data at predetermined first intervals in the direction of the central axis, and acquires the point cloud data contained in each first region, which is a region obtained by dividing the data at each first interval.

[0057] In step S108, the first point cloud data deletion unit 116 obtains the number of point cloud data points included in the second region, which is a region extracted at a second interval in the direction of the central axis in the region corresponding to the torso of the object under inspection.

[0058] In step S110, the first point cloud data deletion unit 116 deletes the corresponding 3D point cloud data in the first region if the number of 3D point cloud data in the first region falls below a threshold determined using the number of point cloud data in the second region, which is a region extracted at a second interval in the direction of the central axis in the region corresponding to the torso of the object under inspection.

[0059] In step S112, the second point cloud data deletion unit 118 deletes the 3D point cloud data remaining after deletion by the first point cloud data deletion unit 116 if the distance in the direction of the central axis of adjacent point cloud data is greater than or equal to a predetermined distance.

[0060] In step S114, the end identification unit 120 performs plane fitting on the 3D point cloud data after deletion by the second point cloud data deletion unit 118 to identify the position of the end of the object under inspection.

[0061] In step S116, the accessory identification unit 122 identifies the distance from the end as the mounting position of the accessory.

[0062] <Examples> Table 1 shows the number of points removed for each coefficient a for three types of steel pipes. Only the 14mm wall thickness steel pipe has a beveled end; the 22mm and 9mm wall thickness steel pipes have right-angled ends. The distance range x was 1mm, and the reference range y was set to 300mm, from 200mm to 500mm from the pipe end. In Table 1, for the 22mm wall thickness steel pipe with a right-angled end, the number of points removed does not change when coefficient a is 0.7 or higher, and for the 9mm wall thickness steel pipe, the number of points removed does not change when coefficient a is 0.2 or higher. This indicates that noise can be removed by setting it above a certain threshold. From these results, it is considered that a coefficient a of 0.7 or higher is appropriate.

[0063] On the other hand, for steel pipes with a wall thickness of 14 mm and beveled ends, the number of points removed increases as the coefficient a increases. This is because the bevel reduces the number of points within the distance range x. In this case, setting the coefficient a too large may remove non-noise data. Figure 5 shows the data after noise reduction processing when the coefficient a=1 for steel pipes with tapered ends. Black dots indicate points that have not been removed, and white dots indicate points that have been removed. Near the pipe ends, it can be seen that the steel pipe data has not been removed, and only the noise data has been removed. On the other hand, Figure 6 shows the data after noise reduction processing when the coefficient a=2 for steel pipes with beveled ends. In Figure 6, it can be seen that the data near the pipe ends has been removed. In other words, it was confirmed that if the coefficient a is 1 or less, the steel pipe data is not removed. From the above, it is considered that a coefficient a of 0.7 to 1 is appropriate when setting an appropriate coefficient a for all steel pipes.

[0064] [Table 1]

[0065] As described above, the position measuring device of this embodiment divides the point cloud data of the region corresponding to the end of the object under inspection from the point cloud data at predetermined first intervals in the direction of the central axis, and acquires the point cloud data contained in each first region obtained by dividing at the first intervals. The position measuring device deletes the point cloud data in the first region if the number of point cloud data contained in the first region falls below a threshold determined using the number of point cloud data contained in the second region, which is a region extracted at second intervals in the direction of the central axis from the region corresponding to the body of the object under inspection from the point cloud data. The position measuring device deletes the point cloud data in the corresponding first region if the distance in the direction of the central axis of adjacent point cloud data in the deleted point cloud data is greater than a predetermined distance, and performs plane fitting on the deleted point cloud data to identify the position of the end of the object under inspection. This makes it possible to accurately identify the position of the end of the object under inspection. Furthermore, the detection accuracy of the end of the object under inspection is improved, and the mounting position of accessories can be measured with high precision.

[0066] The embodiments of the present invention described above can be realized by a computer executing a program. Furthermore, a computer-readable recording medium on which the program is recorded, and a computer program product such as the program itself, can also be applied as embodiments of the present invention. Examples of recording media that can be used include flexible disks, hard disks, optical disks, magneto-optical disks, CD-ROMs, magnetic tapes, non-volatile memory cards, ROMs, and the like. The present invention may also be applied to program products.

[0067] Furthermore, although the explanation used the example of a spiral steel pipe as the object under inspection, it is not limited to this. For example, the object under inspection could be a long, rectangular plate or column.

[0068] The following additional information is disclosed regarding the embodiments described above.

[0069] [Note 1] A position measuring device for determining the position of the end of an object to be inspected, A 3D shape acquisition unit measures the surface shape of the object to be inspected and acquires point cloud data of the object to be inspected, A central axis calculation unit that uses a model that mimics the external shape of the object to be inspected to fit the point cloud data and calculates the direction of the central axis in the point cloud data based on the central axis of the obtained model, A division unit divides the point cloud data of the region corresponding to the end of the object under inspection from the point cloud data at predetermined first intervals in the direction of the central axis, and for each first region obtained by dividing at the first interval, it acquires the point cloud data included in the first region. A first point cloud data deletion unit deletes the point cloud data in the first region if the number of point cloud data in the first region falls below a threshold determined using the number of point cloud data in the second region, which is a region extracted at a second interval in the direction of the central axis in the region corresponding to the torso of the object being inspected. A second point cloud data deletion unit deletes point cloud data from the point cloud data remaining after deletion by the first point cloud data deletion unit if the distance between adjacent point cloud data in the direction of the central axis is greater than or equal to a predetermined distance. An end-identification unit performs plane fitting on the point cloud data after deletion by the second point cloud data deletion unit to identify the position of the end of the object under inspection, A position measuring device having the following features. [Note 2] The threshold is determined using the number of point cloud data included in the second region, the length of the first interval, and the length of the second interval. The position measuring device according to Appendix 1, wherein the length of the second interval is longer than the length of the first interval. [Note 3] The aforementioned threshold is determined according to the following formula in the position measuring device described in Appendix 2. a×L×x / y Note that a is a coefficient, x is the length of the first interval, y is the length of the second interval, and L is the number of point cloud data points included in the second region. [Note 4] a is a position measuring device as described in Appendix 3, where a is 0.7 to 1. [Note 5] A method for determining the position of the end of an object to be inspected, The surface shape of the object to be inspected is measured, and point cloud data of the object to be inspected is acquired. Using a model that mimics the external shape of the object under inspection, the point cloud data is fitted, and the direction of the central axis in the point cloud data is calculated based on the central axis of the obtained model. Of the point cloud data, the point cloud data of the region corresponding to the end of the object to be inspected is divided at predetermined first intervals in the direction of the central axis, and for each first region obtained by dividing at the first interval, the point cloud data included in the first region is acquired. If the number of point cloud data points in the first region falls below a threshold determined using the number of point cloud data points in the second region, which is a region extracted at a second interval in the direction of the central axis in the region corresponding to the torso of the object being inspected, then the corresponding point cloud data points in the first region are deleted. If, among the point cloud data after deletion, the distance in the direction of the central axis of adjacent point cloud data is greater than or equal to a predetermined distance, the corresponding point cloud data is deleted. Planar fitting is performed on the point cloud data after the deletions are made to determine the position of the end of the object being inspected. A method of position measurement performed by a computer. [Note 6] The surface shape of the object to be inspected is measured, and point cloud data of the object to be inspected is acquired. Using a model that mimics the external shape of the object under inspection, the point cloud data is fitted, and the direction of the central axis in the point cloud data is calculated based on the central axis of the obtained model. Of the point cloud data, the point cloud data of the region corresponding to the end of the object to be inspected is divided at predetermined first intervals in the direction of the central axis, and for each first region obtained by dividing at the first interval, the point cloud data included in the first region is acquired. If the number of point cloud data points in the first region falls below a threshold determined using the number of point cloud data points in the second region, which is a region extracted at a second interval in the direction of the central axis in the region corresponding to the torso of the object being inspected, then the corresponding point cloud data points in the first region are deleted. If, among the point cloud data after deletion, the distance in the direction of the central axis of adjacent point cloud data is greater than or equal to a predetermined distance, the corresponding point cloud data is deleted. Planar fitting is performed on the point cloud data after the deletions are made to determine the position of the end of the object being inspected. A location measurement program that instructs a computer to perform this task. [Explanation of symbols]

[0070] 1 3D shape measuring device 2 Arithmetic unit 100 Position Measuring Device 110 3D shape acquisition section 112 Central axis calculation section 114 Division 116 First point cloud data deletion unit 118 Second point cloud data deletion unit 120 End identification part 122 Accessory Identification Section S. Item under inspection (spiral steel pipe with hanging bracket) S1 Main body of the object to be inspected (spiral steel pipe) S2 Accessories (hanging hardware) SE end face

Claims

1. A position measuring device for determining the position of the end of an object to be inspected, A 3D shape acquisition unit measures the surface shape of the object to be inspected and acquires point cloud data of the object to be inspected, A central axis calculation unit that uses a model that mimics the external shape of the object under inspection to fit the point cloud data and calculates the direction of the central axis in the point cloud data based on the central axis of the obtained model, A division unit divides the point cloud data of the region corresponding to the end of the object under inspection from the point cloud data at predetermined first intervals in the direction of the central axis, and for each first region obtained by dividing at each first interval, it acquires the point cloud data included in the first region. A first point cloud data deletion unit deletes point cloud data in the first region if the number of point cloud data in the first region falls below a threshold determined using the number of point cloud data in the second region, which is a region extracted at a second interval in the direction of the central axis in the region corresponding to the torso of the object under inspection, from the point cloud data. A second point cloud data deletion unit deletes point cloud data from the point cloud data remaining after deletion by the first point cloud data deletion unit if the distance between adjacent point cloud data in the direction of the central axis is greater than or equal to a predetermined distance. An end-identification unit performs plane fitting on the point cloud data after deletion by the second point cloud data deletion unit to identify the position of the end of the object under inspection, A position measuring device having the following features.

2. The threshold is determined using the number of point cloud data included in the second region, the length of the first interval, and the length of the second interval. The position measuring device according to claim 1, wherein the length of the second interval is longer than the length of the first interval.

3. The position measuring device according to claim 2, wherein the threshold is determined according to the following formula. a × L × x / y Note that a is a coefficient, x is the length of the first interval, y is the length of the second interval, and L is the number of point cloud data included in the second region.

4. The position measuring device according to claim 3, wherein a is 0.7 to 1.

5. A method for determining the position of the end of an object to be inspected, The surface shape of the object to be inspected is measured, and point cloud data of the object to be inspected is acquired. Using a model that mimics the external shape of the object under inspection, the point cloud data is fitted, and the direction of the central axis in the point cloud data is calculated based on the central axis of the obtained model. Of the point cloud data, the point cloud data of the region corresponding to the end of the object to be inspected is divided at predetermined first intervals in the direction of the central axis, and for each first region obtained by dividing at the first interval, the point cloud data included in the first region is acquired. If the number of point cloud data points in the first region falls below a threshold determined using the number of point cloud data points in the second region, which is a region extracted at a second interval in the direction of the central axis in the region corresponding to the torso of the object being inspected, then the corresponding point cloud data points in the first region are deleted. If, among the point cloud data after deletion, the distance in the direction of the central axis of adjacent point cloud data is greater than or equal to a predetermined distance, the corresponding point cloud data is deleted. Planar fitting is performed on the point cloud data after the deletions are made to determine the position of the end of the object being inspected. A method of position measurement performed by a computer.

6. The surface shape of the object to be inspected is measured, and point cloud data of the object to be inspected is acquired. Using a model that mimics the external shape of the object under inspection, the point cloud data is fitted, and the direction of the central axis in the point cloud data is calculated based on the central axis of the obtained model. Of the point cloud data, the point cloud data of the region corresponding to the end of the object to be inspected is divided at predetermined first intervals in the direction of the central axis, and for each first region obtained by dividing at the first interval, the point cloud data included in the first region is acquired. If the number of point cloud data points in the first region falls below a threshold determined using the number of point cloud data points in the second region, which is a region extracted at a second interval in the direction of the central axis in the region corresponding to the torso of the object being inspected, then the corresponding point cloud data points in the first region are deleted. If, among the point cloud data after deletion, the distance in the direction of the central axis of adjacent point cloud data is greater than or equal to a predetermined distance, the corresponding point cloud data is deleted. Planar fitting is performed on the point cloud data after the deletions are made to determine the position of the end of the object being inspected. A location measurement program that instructs a computer to perform this task.

Images