Method and system for estimating normal vector of mold and attachment

Abstract

To provide a method for estimating a normal vector of an attachment attached to an entry frame of a molded product take-out machine and a normal vector of a mold attached to a molding machine.SOLUTION: A normal vector of a take-out head 60 is estimated through computation by using a coordinate / depth determination section 98 and a normal vector computation section 99 on a basis of a depth data or a coordinate data on three mounting members such as mounting bolts 82 or three suction pads obtained from an image. At least three or more points of extending portions (36A to 36D, 35A to 35) that extend in a direction identical with an opening / closing direction of a fixed mold 31 and movable molds (32 and 34) and that are a part of the fixed mold 31 or the movable molds (32 and 34) or a part of a surrounding component are determined from the image. A normal vector to the mold is estimated through computation using the coordinate / depth determination section 98 and the normal vector computation section 99 on a basis of depth data or coordinate data on the determined three or more extending portions.SELECTED DRAWING: Figure 12

Description

[Technical field]

[0001] The present invention relates to a method and system for estimating a normal vector of an attachment when attached to an entry frame of a molded product removal machine and a normal vector of a mold when attached to a molding machine. [Background technology]

[0002] JP 2002-120175 A (Patent Document 1) discloses an invention related to teaching of a molded product removal machine. Conventionally, when removing a molded product from the mold of a molding machine, an operator would observe the relationship between the removal head (attachment) attached to the entrance frame of the molded product removal machine and the mold, and while moving the head, set up a teaching program to pre-program the path of the entrance frame and the posture of the attachment so that the removal head would not collide with the mold. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2002-120175 A Summary of the Invention [Problem to be solved by the invention]

[0004] When performing the above-mentioned work, there is a problem that the take-out head collides with the mold, damaging the mold. This problem occurs due to the worker's awareness that the installation state of the mold installed in the molding machine and the installation state of the components of the take-out head do not always change. However, the actual situation is that the installation state of the take-out head is not always constant. Under such circumstances, in order to create or modify a teaching program so that the mold and the take-out head do not collide, the worker needs to have a high level of skill and observational skills. If the take-out head is inclined with respect to the mold, there is a concern that, for example, the take-out head may come into contact with the mold when entering the mold, or the mold may be damaged when the faces are matched. Therefore, it is necessary to grasp the inclination of the take-out head and the mold, and adjust the inclination when entering the mold or when the faces are matched.

[0005] To automate teaching and improve accuracy, it is necessary to know the mounting state of the mold and attachments. The mounting state of the mold and attachments can be recognized by knowing the normal vector of the attachment and the normal vector of the mold when it is mounted on the molding machine.

[0006] An object of the present invention is to provide a method and system for estimating the normal vector of an attachment when attached to the entry frame of a molded product removal machine and the normal vector of a mold when attached to a molding machine. [Means for solving the problem]

[0007] The present invention is a method for estimating the normal vector of an attachment attached to the entry frame of a molded product removal machine and the normal vector of a mold attached to a molding machine. In the present invention, a vector perpendicular to each mating surface of the fixed mold and the movable mold of a mold attached to a molding machine is defined as the normal vector of the mold, and a vector perpendicular to a virtual plane that extends in the extension direction of the entry frame and is parallel to the mating surface when the attachment is inserted between the fixed mold and the movable mold is defined as the normal vector of the attachment. In this specification, the movable mold also includes an intermediate mold that is inserted between a general fixed mold and a movable mold. Then, an image including the mating surface of the fixed mold or the peripheral parts around the mating surface of the movable mold is captured using an imaging device capable of capturing an image including depth data or coordinate data of a subject, and at least three or more extension parts that are part of the fixed mold or the movable mold or part of the peripheral parts and extend in a direction that coincides with the opening and closing direction of the fixed mold and the movable mold are identified from the image, and the normal vector of the mold is estimated based on the depth data or coordinate data of the identified three or more extension parts.

[0008] In this specification, "depth data" refers to data including coordinate data of point positions obtained from point cloud data obtained from a three-dimensional imaging device. In other words, depth data refers to information on the direction (specified by an angle) and distance r of an object as seen from the imaging device (camera) when the object is viewed from the imaging device. And "coordinate data" refers to the coordinates P (x, y, z) of an object in an x, y, z Cartesian coordinate system with the imaging device (camera) as the origin. And depth data can be converted into coordinate data.

[0009] In the present invention, an image including three or more mounting fixtures that extend in a direction perpendicular to the entry frame and attach the attachment to the entry frame or three or more contact fixtures that contact the molded product is captured using an imaging device capable of capturing an image including depth data or coordinate data of a subject, and the normal vector of the attachment is estimated based on the depth data or coordinate data of the three mounting fixtures or three contact fixtures from the image. A three-dimensional imaging device can be used as the imaging device that outputs the depth data and coordinate data. Also, a known mathematical method can be used as a method for estimating the normal vector from data of three or more points based on the depth data or coordinate data. In the present invention, the imaging device collects image data including depth data or coordinate data obtained from the mold and attachment that are actually used, and estimates the normal vector of the attachment and the normal vector of the mold, so that the normal vector that has a close relationship to the actual mounting state of the mold and attachment can be estimated. Therefore, it becomes possible to easily realize automation of teaching, posture control of the attachment, and the like.

[0010] The present invention may be specified as a method for estimating a normal vector of a mold in a state where the mold is attached to a molding machine. In this case, too, a vector perpendicular to each mating surface of the fixed mold and the movable mold of the mold in a state where the mold is attached to the molding machine is defined as the normal vector of the mold. Then, an image including the mating surface of the fixed mold or the peripheral parts around the mating surface of the movable mold is captured using an imaging device capable of capturing an image including depth data or coordinate data of a subject. At least three or more extension parts that are part of the fixed mold or the movable mold or part of the peripheral parts and extend in a direction that coincides with the opening and closing direction of the fixed mold and the movable mold are identified from the image, and the normal vector of the mold is estimated based on the depth data or coordinate data of the identified three or more extension parts.

[0011] The present invention may also be specified as a method for estimating the normal vector of an attachment attached to the entry frame of a molded product removal machine. In this case, the normal vector of the attachment is defined as a vector perpendicular to a virtual plane that extends in the direction in which the entry frame extends and is parallel to the mating surface when the attachment is inserted between the fixed mold and the movable mold. Then, an image including three or more mounting fixtures that extend in a direction perpendicular to the entry frame and attach the attachment to the entry frame or three or more contact fixtures that contact the molded product is captured using an imaging device that can capture an image including depth data or coordinate data of a subject. Depth data or coordinate data of the three mounting fixtures or three contact fixtures is obtained from this image, and the normal vector of the attachment is estimated based on this depth data or coordinate data.

[0012] The extending portion may be a guide pin of a mold, a fixed mold, an edge surface of a mold, a tie bar, or the like.

[0013] The mounting tool is, for example, a mounting bolt or an edge surface of a mounting bracket (mounting plate), and the contact tool is, for example, a suction pad.

[0014] The imaging device is a three-dimensional imaging device positioned so as to be capable of capturing an image including one of the mating surfaces of the movable mold and the fixed mold in the open state, and also capable of capturing an image including the removal surface of the attachment when the attachment is outside the mold.

[0015] By using the normal vector of the mold and the normal vector of the attachment estimated using the normal vector estimation method of the present invention, the posture of the attachment when inserting the attachment into the mold can be determined so that the removal operation can be performed in a posture where the normal vector of the mold and the normal vector of the attachment are aligned.

[0016] Furthermore, by using the normal vector of the attachment estimated using the normal vector estimation method of the present invention and an image from an imaging device, the normal vector of the attachment can be defined as a vector along one dimension of the three-dimensional dimensions, and the maximum three-dimensional dimension of the attachment can be accurately measured.

[0017] The invention can be understood as a system for estimating a normal vector of an attachment attached to an entry frame of a molded product removal machine and a normal vector of a mold attached to a molding machine. The system of the present invention includes a processor, and defines a vector perpendicular to each of the mating surfaces of the fixed mold and the movable mold of the mold attached to the molding machine as the normal vector of the mold, and defines a vector perpendicular to a virtual plane that extends in the extension direction of the entry frame and is parallel to the mating surface when the attachment is inserted between the fixed mold and the movable mold as the normal vector of the attachment. The processor obtains depth data or coordinate data of at least three or more extension parts that are part of the fixed mold or the movable mold or part of the peripheral parts and extend in a direction that coincides with the opening and closing direction of the fixed mold and the movable mold, identified from an image including the mating surface of the fixed mold or the peripheral parts around the mating surface of the movable mold captured using a first imaging device capable of capturing an image including depth data or coordinate data of a subject, and estimates the normal vector of the mold based on the depth data or coordinate data. The processor further obtains depth data or coordinate data of three mounting fixtures or three contact fixtures identified from an image including three or more mounting fixtures extending in a direction perpendicular to the approach frame and attaching attachments to the approach frame, or three or more contact fixtures in contact with the molded product, captured using a second imaging device capable of capturing an image including depth data or coordinate data of the subject, and estimates the normal vector of the attachment based on the depth data or coordinate data. [Brief description of the drawings]

[0018] [Figure 1]1 is a perspective view of a molded product manufacturing system including an orthogonal three-axis robot having motion axes in the X, Y, and Z directions to which the normal vector estimation method of the present invention can be applied. [Diagram 2] FIG. 2 is a left side view of the molded product manufacturing system. [Diagram 3] FIG. 2 is a front view of the molded product manufacturing system. [Figure 4] FIG. 1 is a plan view of a molded product manufacturing system. [Diagram 5] 1 is a block diagram showing the configuration of a maximum direction measurement system constructed within the control system of a molded product removal machine in order to apply the normal vector estimation method of the present invention. FIG. [Figure 6] FIG. 13 shows the maximum dimensions of the extraction head. [Figure 7] 1 is an image display for explaining a method for measuring the maximum dimension. [Figure 8] 13 is an image display for explaining a method for measuring maximum sensitivity. [Figure 9] FIG. 11 is a block diagram showing the configuration of a measurement system used when measuring a maximum dimension using one three-dimensional imaging device as an imaging device. [Figure 10] 13A to 13C are diagrams used to explain a case where maximum dimensions in the X direction and Z direction are measured using point cloud data. [Figure 11] 13A to 13C are diagrams used to explain a case where the maximum dimension in the Y direction is measured using point cloud data. [Figure 12] 1 is a block diagram showing a configuration of a normal vector estimation system used when the normal vector estimation method of the present invention is applied to a molded product manufacturing system. [Figure 13] FIG. 11 is a diagram used to explain an example of a method for estimating a normal vector of a pickup head. [Figure 14] FIG. 11 is a diagram used to explain an example of a method for estimating a normal vector of a pickup head. [Figure 15] FIG. 1A is a diagram for explaining depth data obtained by a three-dimensional imaging device, and FIG. 1B is a diagram for explaining a case where three points on an imaging target are imaged by the three-dimensional imaging device. [Figure 16] 10 is a flowchart showing a flow of estimating a normal vector from vibration data by calculation. [Figure 17] 11 is a diagram used to explain an example of a method for estimating a normal vector of a mold. FIG. [Figure 18] 11 is a diagram used to explain an example of a method for estimating a normal vector of a mold. FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0019] Hereinafter, an embodiment of the normal vector estimation method of the present invention will be described in detail with reference to the drawings. Figs. 1 to 4 are a perspective view, a left side view, a front view, and a plan view of a molded product manufacturing system 1 including an orthogonal three-axis robot having motion axes in the X, Y, and Z directions to which the normal vector estimation method of the present invention can be applied and the maximum dimension measurement method can be applied. Fig. 1 shows the X, Y, and Z directions used in this embodiment. The molded product manufacturing system 1 is configured by combining a resin molding machine 3 and a molded product remover 5 as an orthogonal three-axis robot. The molded product remover 5 is a traverse type molded product remover, and its base is supported by a fixed platen 30 of the resin molding machine 3.

[0020] A fixed mold 31 is fixed to a fixed platen 30 of the resin molding machine 3, and a movable mold 34 is fixed to a movable platen 33. The intermediate mold 32 is also considered to be a movable mold. Four tie bars 35A to 35D are arranged between the fixed platen 30 and the movable platen 33 to guide the movement of the movable platen 33. The four tie bars 35A to 35D are arranged at equal intervals, and a virtual center line passing through the centers of the four tie bars 35A to 35D passes through the centers (nozzle centers) of the fixed mold 31 and the movable mold (32, 34). The intermediate mold 32 and the movable mold 34 are guided by guide pins 36A to 36D, respectively. The four guide pins 36A to 36D are also arranged at equal intervals, and a virtual center line passing through the centers of the four guide pins 36A to 36D also passes through the centers (nozzle centers) of the fixed mold 31 and the movable mold 34.

[0021] The molded product remover 5 includes a traverse shaft 53, a first traveling body 55, a pull-out shaft 57, a runner lifting unit 58, and a molded product suction lifting unit 59. The traverse shaft 53 has a cantilever beam structure extending in the X-axis direction horizontally perpendicular to the longitudinal direction of the molding machine 3. The first traveling body 55 is supported by the traverse shaft 53, and moves forward and backward in the X-axis direction along the traverse shaft 53 using an AC servo motor included in the servo mechanism as a driving source. The pull-out shaft 57 is provided on the first traveling body 55, and extends in the Y-axis direction parallel to the longitudinal direction of the molding machine. The pull-out shaft 57 supports the runner lifting unit 58 and the molded product suction lifting unit 59 so as to be movable in the Y direction using an AC servo motor included in the servo mechanism as a driving source. The runner lifting unit 58 has a structure including a lifting frame 58B that moves up and down in the Z direction on a traveling body 58A movably supported on the pull-out shaft 57. The traveling body 58A is driven by an AC servo motor to move in the Y direction. The lifting frame 58B is raised and lowered in the vertical direction (Z direction) by a drive source. The lifting frame 58B is provided with a chuck 58C as an attachment for holding the runner to be discarded.

[0022] The traveling body included in the lifting unit 59 for suction of the molded product 59A is driven by an AC servo motor to move in the Y direction on the pulling shaft 57. The lifting unit 59 for suctioning the molded product includes a lifting frame 59B which is raised and lowered in the vertical direction (Z direction) by a drive source, an inversion unit 59C which serves as a posture control device and rotates about the axis of the lifting frame 59B, and a take-out head 60 provided on the inversion unit 59C. In this embodiment, the normal vector of the take-out head 60 is estimated when the maximum dimension of the take-out head 60 is measured as an attachment.

[0023] In this embodiment, in order to experimentally find a suitable installation location, eight imaging devices C1 to C8 are installed on each part of the molded product remover 5 and molding machine 3, and on a stand 7 placed to the side of the molding machine 3. Three-dimensional cameras (three-dimensional imaging devices) are used as these imaging devices C1 to C8. In this embodiment, the imaging devices C1 to C8 that can provide a suitable image can be selected to obtain the required image.

[0024] Fig. 5 is a block diagram showing the configuration of a maximum dimension measuring system built in the control system of the molded product removal machine 5 in order to implement a method for measuring the three-dimensional shape dimensions of an attachment. Fig. 6 is a diagram showing the maximum dimensions of the removal head 60. Figs. 7 and 8 are image displays for explaining a method for measuring the maximum dimensions.

[0025] In the method for measuring the maximum dimension using the normal vector estimation method of this embodiment, the maximum dimension in the X direction, the maximum dimension in the Y direction, and the maximum dimension in the Z direction of the take-out head 60 are measured based on images of the take-out head 60 taken by at least one or more imaging devices (C1 to C8) before the take-out head 60 starts working, with the take-out head 60 as an attachment attached to the lifting frame 59B as the entry frame of the molded product take-out machine 5 as an orthogonal three-axis robot. Depending on the measurement method used, it is preferable that the three-dimensional position coordinates of the imaging device when taking the image by at least one or more imaging devices, the three-dimensional position coordinates of the taken take-out head (attachment) 60, the angle of view of the imaging device, and the mounting attitude of the take-out head (attachment) 60 to the lifting frame 59B (entry frame) are determined so that the image obtained by the imaging device contains information necessary for measuring the three-dimensional shape dimensions of the take-out head 60. The more information such as coordinate information, the angle of view of the imaging device, and the mounting attitude is available in advance, the easier it is to perform calculations when measuring the maximum dimension based on the image, and the less correction calculations are required for the image data due to differences in the mounting position and mounting attitude of the imaging device. Therefore, in order to obtain the angle of view information of the imaging device, the method of estimating the normal vector of the attachment of the present invention can be used. In other words, if the normal vector of the extraction head 60 is known, the angle of view of the imaging device can be matched to the angle of view of the normal vector.

[0026] 5, an imaging system 71 including one or more imaging devices is used to measure the maximum dimension by an operator, the imaging system 71 including a first imaging device C11 for obtaining a first image of the take-out head 60 in a position where the maximum dimension in the X direction and the maximum dimension in the Z direction of the take-out head 60 can be measured, a second imaging device C12 for obtaining a second image of the take-out head 60 in a position where the maximum dimension in the Y direction and the maximum dimension in the Z direction of the take-out head 60 can be measured, and an image display device 72 having a screen 73 with an orthogonal coordinate axis indicator for displaying the first image or the second image. Here, the screen 73 with an orthogonal coordinate axis indicator is a gauge screen having two orthogonal axes (GZ-GX, GZ-GY).

[0027] In this embodiment, the lift frame 59B The controller 10 uses an X-direction drive source 74 and an X-direction movement amount measuring device 75 for moving the drive source 74 in the X direction, a Y-direction drive source 76 and a Y-direction movement amount measuring device 77 for moving the drive source 74 in the Y direction, and a Z-direction drive source 78 and a Z-direction movement amount measuring device 79 for moving the drive source 74 in the Z direction. The operator performs the following operations using an operation unit 97 consisting of operation switches and the like provided on the controller. In the following operations, image display on the screen 73 of the image display device 72 is performed by an image control unit 96 in a control device 90 configured in a control unit of the molded product extractor 5. The calculation of the maximum dimension is performed by a dimension calculation unit 95 based on the outputs of the X-direction movement amount measuring device 75 to the Z-direction movement amount measuring device 79. Then, operation commands to the X-direction drive source 74 to the Z-direction drive source 78 are output from a drive control unit 94 in accordance with the operation from the operation unit 97. The control device 90 further includes a teaching unit 91 for performing teaching operations described later, a data storage unit 92 for storing teaching data, and a usability determination unit 93.

[0028] Specifically, the first imaging device C11 may be the imaging device C1 or C2 shown in FIG. 1 to FIG. 4 if the attitude of the take-out head 60 is not changed. The first imaging device C11 photographs the take-out head 60 as an attachment from the front. The second imaging device C12 may be the imaging device C6 to C8 shown in FIG. 1 to FIG. 4 if the attitude of the take-out head 60 is not changed. The imaging devices C1 to C5 shown in FIG. 1 to FIG. 4 may be used if the attitude of the take-out head 60 is changed. The first imaging device C11 photographs the take-out head 60 as an attachment from the front. The second imaging device C12 photographs the take-out head 60 from the side. When the first and second imaging devices C11 and C12 are arranged in front or on the side of the take-out head 60, the normal vector estimation method of the present invention may be used.

[0029] More specifically, as shown in Fig. 7, the directions of the orthogonal coordinate axes GZ and GX of the indicator G with orthogonal coordinate axes on the screen 73 are aligned with the X direction and the Z direction in the first image from the first imaging device C11. To achieve this state, the image on the screen 73 and the image of the indicator G are viewed. Then, while driving the Z direction drive source 78 consisting of a servo motor or the like to move the take-out head 60 in the Z direction, the maximum dimension in the Z direction is calculated from the distance [see Zs-Ze in Fig. 6] measured by the Z direction movement amount measuring device 79 consisting of an encoder or the like from when the suction pad 60A, which is the outermost end of the take-out head 60 in one direction in the Z direction in the first image, passes through a reference line (GX in this example) on the screen (the measured value at this time is Zs) until the suction pad 60B, which is the outermost end in the other direction in the Z direction, passes through the reference line (GX in this example) (the measured value at this time is Ze) [Figs. 7(A) to 7(D)].

[0030] In addition, while the directions in which the orthogonal coordinate axes GZ and GY of the indicator G with orthogonal coordinate axes on the screen 73 extend are aligned with the X and Z directions in the first image, the X-direction drive source 74 is driven to move the take-out head 60 in the X direction, and the maximum dimension in the X direction is measured from the distance [see Xs-Xe in FIG. 6] measured by the X-direction movement amount measuring device from when the suction pad 60C, which is the outermost end in one direction in the X direction in the first image, passes through a reference line (GZ in this example) on the screen (the measured value at this time is designated as Xs) until the suction pad 60D, which is the outermost end in the other direction in the X direction, passes through the reference line (GZ) (the measured value at this time is designated as Xe).

[0031] Furthermore, as shown in FIG. 8, while the direction in which the Cartesian coordinate axes GZ-GY of the indicator G with Cartesian coordinate axes on the screen 73 extend is aligned with the Y direction and Z direction in the second image (side image), the Y direction drive source 76 is driven to move the take-out head 60 in the Y direction, and the maximum dimension in the Y direction is measured from the distance measured by the Y direction movement measuring device [see Ys-Ye in FIG. 6] from when the outermost end 60E in one direction of the Y direction in the second image passes through the reference line (GZ in this example) on the screen (the measured value at this time is Ys) until the outermost end 60F in the other direction of the Y direction passes through the reference line (GZ) (the measured value at this time is Ye).

[0032] According to this method for measuring the maximum dimension, the maximum dimensions in the X-, Y-, and Z-directions can be measured with simple equipment and simple operations. The take-out head 60 mounted on the lifting frame 59B of the molded product take-out machine 5 is often subject to replacement of parts mounted for repairs with parts different from the design specifications, or to changes in the position of the parts mounted for maintenance. For example, the take-out head 60 mounted on the lifting frame 59B as the entry frame is configured by combining accessories including air tubes or wiring that power the take-out head 60. Therefore, the position and position of these air tubes or wiring may change every time the take-out head 60 is replaced. In addition, there are also cases where an operator mounts a wrong take-out head on the lifting frame 59B. Even in such a case, if the maximum dimensions in the X-, Y-, and Z-directions of the take-out head 60 are measured while mounted on the lifting frame 59B, it is possible to determine the change in the shape of the take-out head 60 to be actually used before actually starting the take-out operation using the take-out head 60. As a result, it is possible to detect in advance whether the take-out head 60 will collide with a part or the like on its path of movement, or whether the wrong take-out head is attached.

[0033] (Other measurement systems) FIG. 9 is a block diagram showing the configuration of a measurement system used when measuring the maximum dimension using one three-dimensional imaging device C13 as an imaging device. In this measurement system, the maximum dimension is measured using a three-dimensional imaging device C13 as an imaging device, which measures the surface of an object and outputs a number of points having three-dimensional coordinates as point cloud data. From image data obtained by imaging the take-out head 60 from the front using the three-dimensional imaging device C13, a first point cloud data acquisition unit 101A acquires first point cloud data capable of measuring the maximum dimension in the X direction and the maximum dimension in the Z direction of the take-out head 60. Also, from image data obtained by imaging the take-out head 60 from the side using the three-dimensional imaging device C13, a second point cloud data acquisition unit 101B acquires second point cloud data capable of measuring the maximum dimension in the Y direction and the maximum dimension in the Z direction of the take-out head 60. When the three-dimensional imaging device C13 is disposed in front or on the side of the take-out head 60, the normal vector estimation method of the present invention can be used. When one three-dimensional imaging device is used, the imaging device C1 in Fig. 1 is used as the three-dimensional imaging device, and after capturing a front image, the take-out head 60 is rotated 90 degrees using the inversion unit 59C as the attitude changing device to obtain a side image. Of course, two three-dimensional imaging devices may be used to obtain both a front image and a side image.

[0034] A maximum dimension determination unit 100 in the control device 90' ​​measures the maximum dimension in the Y direction based on the coordinates of a point located at one outermost end in the Z direction and the coordinates of a point located at the other outermost end from the first point cloud data. It also measures the maximum dimension in the X direction based on the coordinates of a point located at one outermost end in the X direction and the coordinates of a point located at the other outermost end from the first point cloud data. It also measures the maximum dimension in the Z direction based on the coordinates of a point located at one outermost end in the Y direction and the coordinates of a point located at the other outermost end from the second point cloud data.

[0035] The point cloud data obtained by the three-dimensional imaging device is a data file of three-dimensional coordinates of a large number of points that are automatically measured on the surface of an object. That is, each point constituting the point cloud contains three-dimensional coordinate information. When the take-out head 60 shown in FIG. 10(A) is imaged by the three-dimensional imaging device, point cloud data as shown in FIG. 10(B) is obtained. Each point contains three-dimensional coordinate information. As shown in FIG. 10(C)(D), the maximum dimension determination unit 100 obtains the coordinate values ​​of the point with the largest and smallest X-direction coordinates among the points having the same depth in the X direction (same Y coordinate value) from the point cloud data, and the difference between these coordinate values ​​is set as the maximum dimension Xm in the X direction. Also, the point cloud data obtains the coordinate values ​​of the point with the largest and smallest Z-direction coordinates among the points having the same depth in the Z direction (same Y coordinate value), and the difference between these coordinate values ​​is set as the maximum dimension Zm in the Z direction. Next, as shown in Fig. 11, from the point cloud data acquired by the second point cloud data acquisition unit 101B, the coordinate values ​​of the point with the largest and smallest Y coordinates among points having the same depth in the Y direction (same X coordinate value) are obtained, and the difference between these coordinate values ​​is set as the maximum dimension Ym in the Y direction. In this way, the three-dimensional maximum dimensions Xm, Ym, and Zm of the extraction head 60 can be obtained from the point cloud data. After the maximum dimensions are obtained, they can be used in the same manner as in the first embodiment.

[0036] The method of measuring the maximum dimension based on an image may use a known measurement technique, but the above example can minimize measurement errors caused by differences in the skill level of the worker. By using the photographed data, the maximum dimension can be automatically measured from the difference between the image of the regular extraction head (attachment) and the photographed image, or the maximum dimension can be measured using AI technology that excels in image recognition. As a result, it is possible to automate the measurement.

[0037] (Estimation of normal vectors of attachments and molds) Fig. 12 is a block diagram showing the configuration of a normal vector estimation system used when the normal vector estimation method of the present invention is applied to the above-mentioned molded product manufacturing system. In the block diagram of Fig. 12, the same reference numerals are attached to the components common to the components of the block diagram of Fig. 5, and elements unnecessary for estimating the normal vector are not shown. In the block diagram of Fig. 12, a coordinate or depth determination unit 98 and a normal vector calculation unit 99 are provided in a control device 90, and a posture control device drive unit 59D that controls a reversal unit 59C as a posture control device based on the normal vector estimation result is also shown. At least the coordinate or depth determination unit 98 and the normal vector calculation unit 99 in the control device 90 are each configured using a processor.

[0038] This normal vector estimation system estimates the normal vector of the take-out head 60 as an attachment attached to the lift frame 59B as the entry frame of the molded product take-out machine 5 and the normal vector of the molds 31, 32, 34 attached to the molding machine 3. In this embodiment, a vector perpendicular to each mating surface of the fixed mold 31 and the movable mold (32, 34) of the molds attached to the molding machine 3 becomes the normal vector of the mold. Also, a vector perpendicular to a virtual plane that extends in the extension direction of the lift frame 59B as the entry frame and is parallel to the mating surface when the take-out head 60 as an attachment is inserted between the fixed mold 31 and the movable mold (32, 34) becomes the normal vector of the take-out head 60 as an attachment.

[0039] In this embodiment, as shown in Fig. 13 and Fig. 14, an image including the attachment bolt 82 as the attachment at three or more points for attaching the take-out head (attachment) 60 to the lift-up frame 59B, which is the entry frame, or the suction pads 60A to 60D as the contacting devices at three or more points for contacting the molded product is taken using an imaging device C11 (three-dimensional imaging device) capable of taking an image including depth data or coordinate data of the subject, which extends in a direction perpendicular to the lift-up frame 59B as the entry frame (pulling direction) and includes the attachment at three or more points. As the first imaging device C11, the imaging device C1 or C2 shown in Figs. 1 to 4 can be used. The first imaging device C11 takes an image of the take-out head 60 as the attachment from the front or back. The image data taken by the three-dimensional imaging device includes the depth data of the subject.

[0040] As shown in Fig. 14(A), the four mounting bolts 82 seen on the back of the take-out head 60 are the mounting references of the take-out head. Therefore, a virtual plane PS1 formed by connecting the four mounting bolts 82 and perpendicular to the four mounting bolts 82 is a plane perpendicular to the normal vector of the take-out head 60. Also, as shown in Fig. 14(B), a virtual plane PS2 formed by connecting the centers of the suction pads 60A to 60D and perpendicular to the suction pads 60A to 60D is a plane perpendicular to the normal vector of the take-out head 60. Then, based on the depth data or coordinate data of the three mounting bolts 82 or the three suction pads from these images, a coordinate or depth determination unit 98 and a normal vector calculation unit 99 are used to calculate and estimate the normal vector of the take-out head 60. The coordinate or depth determination unit 98 acquires depth data or coordinate data of the four mounting bolts 82 or the four suction pads 60A to 60D from information (data stored in the data storage unit 92) obtained from an image captured by the three-dimensional imaging device used as the imaging device C11. Of course, this data will depend on the specifications of the three-dimensional imaging device used. In this embodiment, the mounting bolt 82 is used as an example of a mounting fixture, but the mounting fixture also includes the mounting plate 81. In the case of a mounting bracket (mounting plate), the multiple edge surfaces can be regarded as three or more mounting fixtures.

[0041] The normal vector calculation unit 99, which includes a processor, calculates the normal vector according to a known calculation method for obtaining a normal vector from a plane in space. FIG. 15(A) is a diagram used to explain the relationship between depth data and coordinate data. In FIG. 15(A), point O is the position of the three-dimensional imaging device, and point P is, for example, a point on the front or back of the extraction head. In this case, the depth data refers to information on the direction (specified by angles θ and φ) and distance r of the extraction head as seen from the three-dimensional imaging device when point P of the extraction head is seen from the three-dimensional imaging device. The coordinate data x, y, and z of point P can be expressed as x=rsinθcosφ, y=rsinθsinφ, and z=rcosθ.

[0042] As shown in FIG. 15(B), for example, when the coordinate data of three points P, Q, R on the front of the extraction head are P(x1, y1, z1), Q(x2, y2, z2), and R(x3, y3, z3), the coordinate data x, y, and z of these three points can be obtained by inserting the depth data P(r1, φ1, θ1), Q(r2, φ2, θ2), and R(r3, φ3, θ3) of points P, Q, and R obtained by the three-dimensional imaging device into the above formula [x=rsinθcosφ, y=rsinθsinφ, z=rcosθ]. If the plane in the xyz orthogonal coordinate space is given by the plane equation Ax+By+Cz+D=0, where A, B, C, and D are real numbers, then the normal vector n of the front of the pick-up head is n=(A, B, C), so by substituting the coordinate data of P, Q, and R into the plane equation and solving the simultaneous equations, A, B, and C can be obtained. Figure 16 shows the flow of this conversion and calculation.

[0043] Another way to calculate the normal vector is to let the basis vectors of the x-axis, y-axis, and z-axis be i, j, and k, respectively. The normal vector a is expressed by the formula a=Ai+Bj+Ck. The unit normal vector n of magnitude 1 is n=(Ai+Bj+Ck) / (A 2 +B 2 +C 2 ) 1 / 2The normal vector of the virtual surface PS1 or PS2 may be calculated using this calculation method. Of course, other calculation methods may also be used. The above calculations are performed by a normal vector calculation unit 99.

[0044] In this embodiment, when the normal vector of the mold is obtained, an image including the mating surface of the fixed mold 31 or the peripheral parts around the mating surface of the movable mold (32, 34) is captured using a second imaging device C12 (three-dimensional imaging device) capable of capturing an image including depth data or coordinate data. At least one of the imaging devices C2 to C8 provided so that the angle of view can be adjusted at the positions shown in Figs. 1 to 4 can be used as the second imaging device C12. In this embodiment, at least three or more extension parts (see guide pins 36A to 36D of the mold, edge surface 34A of the fixed mold 31 or the movable mold 34, tie bars 35A to 35D, etc. in Figs. 17 and 18) which are part of the fixed mold 31 or the movable mold 34 or part of the peripheral parts and extend in a direction that coincides with the opening / closing direction (pulling direction) of the fixed mold 31 and the movable mold (32, 34) are identified from the image captured by the second imaging device consisting of a three-dimensional imaging device. This can be determined from the image displayed on the screen 73 of the image display device 72. As described above, the image data captured by the three-dimensional imaging device and stored in the data storage unit 92 also includes depth data and coordinate data of the subject, so the coordinate or depth determination unit 98 determines the depth data or coordinate data of the three or more identified extension parts. Then, based on the depth data or coordinate data determined by the coordinate or depth determination unit 98, the normal vector calculation unit 99 calculates and estimates the normal vector of the mold.

[0045] In the example of Fig. 17, depth data or coordinate data of the four guide pins 36A to 36D or tie bars 35A to 35D are obtained from the image, and a virtual surface PS3 or PS4 perpendicular to the four guide pins 36A to 36D or tie bars 35A to 35D is obtained based on this data. The guide pins 36A to 36D or tie bars 35A to 35D are extensions that do not change when the mold is attached, and the mating surface of the mold attached using these is substantially parallel to the three-dimensional virtual surface PS3 or PS4 obtained from the coordinate data or depth data of the guide pins 36A to 36D or tie bars 35A to 35D. Therefore, in this way, the normal vector of the mold can be obtained using the above-mentioned calculation formula.

[0046] In the example of FIG. 18, a three-dimensional imaging device is attached to the position of the imaging device C5 shown in FIG. 1, and this three-dimensional imaging device is advanced into the mold to obtain three-dimensional image data including four intersections of the edge surface 34A (four edges surrounding the mating surface of the mold) of the movable mold 34. Then, a three-dimensional virtual surface PS5 including these four intersections is obtained from the coordinate data or depth data of the four intersections obtained accompanying this image data. The mating surface of the mold is substantially parallel to the three-dimensional virtual surface PS5. Then, a normal vector in the drawing direction can be obtained from these data. In this example, the normal vector is obtained from information on the part related to the mating surface of the mold, so that the normal vector that may change depending on the mounting state of the mold can be accurately estimated.

[0047] (Usability, Teaching) The maximum dimension and the normal vector measured as described above can be used in any manner. In the above embodiment, before teaching is performed using the operation of the teaching unit 91, the usability determination unit 93 determines whether the take-out head 60 can be used in the take-out work using the take-out head 60 from the information on the three-dimensional shape and dimensions of the take-out head 60, and if the take-out head 60 is unusable, an alarm is output from the usability determination unit 93. In other words, it is possible to determine whether the take-out head 60 can enter the open mold without colliding with the mold from the maximum dimension, and it is possible to determine whether the take-out head can reliably take out the molded product from the degree of mismatch between the normal vector of the take-out head and the normal vector of the mold. In other words, when teaching is performed, it is determined in advance that there is a possibility of collision with surrounding objects during the movement of the take-out head 60 or that the molded product cannot be reliably taken out, and if the take-out head 60 is not usable, an alarm is issued.

[0048] If the problem is with the mounting posture of the take-out head 60, the take-out head 60 can be re-mounted based on the alarm, or a command can be given to the posture control device drive unit 59D to operate the reversal unit 59C as a posture control device, thereby changing the posture of the take-out head 60 to the appropriate posture. [Industrial Applicability]

[0049] According to the present invention, image data including depth data or coordinate data obtained from the actually mounted mold and attachment are collected by an imaging device, and the normal vector of the attachment and the normal vector of the mold are estimated, so that the normal vector that is closely related to the actual mounting state of the mold and attachment can be estimated. Therefore, it becomes possible to easily realize automation of teaching, posture control of the attachment, etc. [Explanation of symbols]

[0050] 1 Molding manufacturing system 3 Molding machine 31 Fixed mold 32 Intermediate mold 34 Movable mold 5 Molded product removal machine 59B Lifting frame (entrance frame) 6 Extraction head (attachment) 71 Imaging Device System 72 Image display devices 73 screens 74 X-axis direction drive source 75 X-axis direction movement measuring section 76 Y-axis direction drive source 77 Y-axis direction movement measurement unit 78 Z-axis direction drive source 79 Z-axis direction movement measuring section C11 First imaging device C12 Second imaging device 90 Control device 91 Teaching Department 92 Data storage unit 93 Usability determination section 96 Image control section 97 Control section 98 Coordinate or depth determination unit 99 Normal vector calculation section

Claims

1. A method for estimating a normal vector of an attachment when attached to an entry frame of a molded product removal machine and a normal vector of a mold when attached to a molding machine, comprising: A vector perpendicular to each mating surface of the fixed mold and the movable mold of the mold when the mold is mounted on the molding machine is defined as a normal vector of the mold, and a vector perpendicular to a virtual plane that extends in the extension direction of the entry frame and is parallel to the mating surfaces when the attachment is inserted between the fixed mold and the movable mold is defined as a normal vector of the attachment, an image including the mating surface of the fixed mold or peripheral components around the mating surface of the movable mold is captured using a first imaging device capable of capturing an image including depth data or coordinate data of a subject, at least three or more extension portions which are part of the fixed mold or the movable mold or part of the peripheral components and extend in a direction that coincides with the opening and closing direction of the fixed mold and the movable mold are identified from the image, and a normal vector of the mold is estimated based on the depth data or coordinate data of the identified three or more extension portions; A method for estimating a normal vector, comprising: using a second imaging device capable of capturing an image including depth data or coordinate data of a subject, capturing an image including three or more mounting fixtures extending in a direction perpendicular to the entry frame and attaching the attachment to the entry frame, or three or more contact fixtures that come into contact with a molded product; and estimating the normal vector of the attachment based on the depth data or coordinate data of the three mounting fixtures or three contact fixtures obtained from the image.

2. A method for estimating a normal vector of a mold when attached to a molding machine, comprising the steps of: A vector perpendicular to each mating surface of the fixed mold and the movable mold of the mold in a state where the mold is mounted on the molding machine is defined as a normal vector of the mold; A method for estimating a normal vector, comprising the steps of: capturing an image including the mating surface of the fixed mold or peripheral parts surrounding the mating surface of the movable mold using an imaging device capable of capturing an image including depth data or coordinate data of a subject; identifying from the image at least three or more extension portions which are part of the fixed mold or movable mold or part of the peripheral parts and extend in a direction which coincides with the opening and closing direction of the fixed mold and the movable mold; and estimating a normal vector of the mold based on the depth data or coordinate data of the identified three or more extension portions.

3. A method for estimating a normal vector of an attachment attached to an entry frame of a molded product removal machine, comprising: A vector perpendicular to a virtual plane that extends in the extension direction of the entry frame and is parallel to the mating surface when the attachment is inserted between the fixed mold and the movable mold is defined as a normal vector of the attachment; A method for estimating a normal vector, comprising: using an imaging device capable of capturing an image including depth data or coordinate data of a subject, capturing an image including three or more mounting fixtures extending in a direction perpendicular to the entry frame and attaching the attachment to the entry frame, or three or more contact fixtures that come into contact with a molded product; and estimating the normal vector of the attachment based on the depth data or coordinate data of the three mounting fixtures or three contact fixtures obtained from the image.

4. The method for estimating a normal vector according to claim 1 or 2, wherein the extended portion is a guide pin of the mold, an edge surface of the fixed mold or a movable mold, or a tie bar.

5. The method for estimating a normal vector according to claim 1 or 3, wherein the mounting fixture is a mounting bolt or a mounting metal fitting.

6. The method for estimating a normal vector according to claim 1 or 3, wherein the contact tool is a suction pad.

7. The method for estimating a normal vector according to claim 1 , wherein the imaging device is a three-dimensional imaging device.

8. A method for determining the attitude of an attachment, characterized in that, using the normal vector of the mold and the normal vector of the attachment estimated using the normal vector estimation method of claim 1, a posture of the attachment when the attachment is inserted into the mold is determined so that a removal operation can be performed in a posture in which the normal vector of the mold and the normal vector of the attachment are aligned.

9. A method for measuring the maximum three-dimensional dimension of an attachment by using a normal vector of the attachment estimated using the normal vector estimation method described in claim 3 and the image of the imaging device, and defining the normal vector of the attachment as a vector along a one-dimensional dimension of the three-dimensional dimension.

10. A system for estimating a normal vector of an attachment attached to an entry frame of a molding machine and a normal vector of a mold attached to a molding machine, A processor is included. A vector perpendicular to each mating surface of the fixed mold and the movable mold of the mold when the mold is mounted on the molding machine is defined as a normal vector of the mold, and a vector perpendicular to a virtual plane that extends in the extension direction of the entry frame and is parallel to the mating surfaces when the attachment is inserted between the fixed mold and the movable mold is defined as a normal vector of the attachment, The processor, determining depth data or coordinate data of at least three or more extension portions that are part of the fixed mold or the movable mold or part of the peripheral components and extend in a direction that coincides with the opening and closing direction of the fixed mold and the movable mold, the extension portions being identified from an image including the mating surface of the fixed mold or the peripheral components around the mating surface of the movable mold, captured using a first imaging device capable of capturing an image including depth data or coordinate data of a subject; and estimating a normal vector of the mold based on the depth data or coordinate data; A normal vector estimation system characterized by determining depth data or coordinate data for three mounting fixtures or three contact fixtures identified from an image including three or more mounting fixtures extending in a direction perpendicular to the entry frame and attaching the attachment to the entry frame, or three or more contact fixtures in contact with the molded product, captured using a second imaging device capable of capturing an image including depth data or coordinate data of a subject, and estimating a normal vector of the attachment based on the depth data or coordinate data.

Images