JP6650420B2 - Alignment detection device and alignment detection method - Google Patents

Description

  An object of the present invention is to detect a workpiece having an outer peripheral portion in which a convex portion and a concave portion are provided in a shape that is rotationally symmetrical about a symmetry axis and are provided periodically and repeatedly, and detect a misalignment of the workpiece with respect to the rotating portion. It relates to detection technology.

  As a device for inspecting the appearance of a work rotationally symmetric about a symmetry axis, for example, a work inspection device described in Patent Document 1 is known. In this work inspection device, the work is held by the holder connected to the motor. Then, the work is imaged by a plurality of cameras while rotating the work by the motor, and the appearance of the work is inspected based on the captured images.

JP 2012-63268 A

  The apparatus described in Patent Literature 1 uses a gear as a workpiece, and inspects the gear for any scratches or defects. In this device, the holder has a shaft portion that is passed through a through hole that penetrates the work in the axial direction, and a clamp mechanism that clamps the work coaxially with the shaft portion. The rotation of the rotating shaft of the motor causes the shaft and the workpiece to rotate integrally. Here, if the outer peripheral surface of the work is a smooth surface, even if the symmetry axis of the work held by the clamp mechanism does not coincide with the rotation axis of the motor, the captured image is processed. It is possible to perform the above inspection with relatively high accuracy. On the other hand, when a work in which a concave portion and a convex portion are periodically repeated with respect to the outer peripheral surface, such as a gear, is to be inspected, it is difficult to perform the inspection by simple image processing. Therefore, in order to inspect such a work with high accuracy, it is desired to accurately grasp the misalignment of the work with respect to the motor. However, there has not been a technique for detecting the above-described misalignment with high accuracy. In addition to the detection accuracy, a reduction in the time required for detecting the misalignment is also desired at the same time.

  SUMMARY OF THE INVENTION The present invention has been made in view of the above-described problems, and has a short period of time when a workpiece having an outer peripheral portion in which a convex portion and a concave portion are provided in a shape that is rotationally symmetric about a symmetry axis and are provided periodically and cyclically with respect to a rotating portion is short. It is another object of the present invention to provide a misalignment detection technique for detecting with high accuracy.

A first aspect of the present invention is to detect a misalignment of a work with respect to a rotating part, which rotates a work having an outer peripheral portion provided with a convex portion and a concave portion periodically and repetitively provided in a rotationally symmetric shape about a symmetry axis. An image pickup unit for picking up an image of a part of a work corresponding to an angle φ ( = 90 ° ) of an outer peripheral part of a work rotated by a rotating unit, and an outer peripheral part of the work included in the work partial image. An edge waveform deriving unit that obtains an edge waveform indicating a shape, and from the edge waveform derived by the edge waveform deriving unit, obtains position information indicating a position of a convex portion or position information indicating a position of a concave portion, based on the position information. And a misalignment calculating unit for calculating a misalignment of the work with respect to the rotating unit.
According to a second aspect of the present invention, there is provided a method for rotating a work having an outer peripheral portion provided with a convex portion and a concave portion periodically and repeatedly formed in a rotationally symmetrical shape about a symmetry axis, the work being misaligned with respect to the rotating portion. An image pickup unit for picking up an image of a work part corresponding to an angle φ (90 ° ≦ φ <360 °) of an outer peripheral portion of a work rotated by a rotating unit; From the edge waveform deriving unit for obtaining an edge waveform indicating the shape of the outer peripheral portion of the workpiece to be obtained, and from the edge waveform derived by the edge waveform deriving unit, position information indicating the position of the convex portion or position information indicating the position of the concave portion is obtained. And a misalignment calculating unit for calculating a misalignment of the work with respect to the rotating unit based on the position information, wherein the imaging unit has a first imaging region and a second imaging region for imaging the outer peripheral portion of the work at different positions. While the work is being rotated by the rotating unit by the angle θ (where φ / 2 ≦ θ <φ), the first imaging region captures the first work local image for the angle θ and the second imaging region It is characterized in that a work partial image is obtained by capturing a second work local image for θ.

A third aspect of the present invention is a method for detecting misalignment, comprising: holding a workpiece having an outer peripheral portion provided with a convex portion and a concave portion which are rotationally symmetrical about a symmetry axis and provided periodically and repeatedly. A first step of holding the work by the rotating unit, a second step of rotating the work around the rotation axis of the rotating unit by rotating the work holding unit holding the work by 90 ° by the rotating unit, and a second step of rotating the work by the rotating unit. A third step of obtaining a workpiece partial image for an angle φ ( = 90 ° ) of the outer peripheral portion of the work, a fourth step of obtaining an edge waveform indicating the shape of the outer peripheral portion of the work included in the workpiece partial image, and an edge waveform And a fifth step of calculating the misalignment of the workpiece with respect to the rotating unit on the basis of the above.
Further, a fourth aspect of the present invention is a method for detecting misalignment, comprising: holding a workpiece having an outer peripheral portion having a shape which is rotationally symmetric about an axis of symmetry and in which convex portions and concave portions are provided periodically and repeatedly. A first step of holding the work by the unit, a second step of rotating the work around the rotation axis of the rotating unit by rotating the work holding unit holding the work by an angle θ by the rotating unit, and a second step of rotating the work by the rotating unit. A third step of acquiring, by the imaging unit, a workpiece partial image corresponding to an angle φ of the outer periphery of the workpiece, a fourth step of obtaining an edge waveform indicating the shape of the outer periphery of the workpiece included in the workpiece partial image, and A fifth step of calculating the center misalignment of the work with respect to the rotating part by using an angle θ of 90 ° or more and less than 360 °; an angle θ of φ / 2 or more and less than φ; Different positions from each other The image forming apparatus has a first imaging region and a second imaging region for imaging the outer peripheral portion of the work, and in the third step, the first imaging region has an angle θ of the first work while the work is rotated by the angle θ by the rotating unit. It is characterized in that a partial image of a work is obtained by capturing a local image and capturing a second work local image corresponding to the angle θ in the second imaging region.

  When the work is misaligned with respect to the rotating unit, an image obtained by imaging the outer peripheral portion of the work while rotating the work by the rotating unit includes information reflecting the misalignment. Therefore, as will be described in detail later, an edge waveform indicating the shape of the outer peripheral portion of the work is obtained based on an image of one or more rounds of the outer peripheral portion of the work, and the misalignment of the work with respect to the rotating portion is determined based on the edge waveform. It is effective to use a detection method of calculating (comparative example) to increase the detection accuracy. However, in the above detection method, it is necessary to rotate the work at least one turn or more, which is one obstacle in shortening the time required for detecting the misalignment.

  Here, as will be described in detail later, when an edge waveform for one round is analyzed, four characteristic portions (two extreme values + two inflection points) are present in the edge waveform in association with the misalignment. These four feature portions appear at 90 ° intervals, and one feature portion is always included in the edge waveform corresponding to 90 ° of the outer peripheral portion of the work. Thus, in the present invention, the workpiece is rotated by 90 ° or more and less than 360 ° to capture a workpiece partial image for an angle φ (90 ° ≦ φ <360 °) of the outer peripheral portion of the workpiece, and based on the workpiece partial image. An edge waveform partially indicating the shape of the outer peripheral portion of the work is obtained, and the center deviation of the work with respect to the rotating portion is calculated based on the edge waveform.

  As described above, according to the present invention, an edge waveform is obtained from a work partial image, and a work misalignment with respect to the rotating unit is calculated based on the edge waveform. It can be detected with high accuracy.

FIG. 1 is a diagram showing an entire configuration of an inspection device equipped with a first embodiment of a misalignment detection device according to the present invention. FIG. 2 is a block diagram illustrating an electrical configuration of the inspection device illustrated in FIG. 1. FIG. 3 is a perspective view illustrating a configuration of a work holding unit. 3 is a flowchart illustrating a work inspection operation by the inspection device illustrated in FIG. 1. It is a figure which shows an inspection operation | movement typically. 5 is a flowchart illustrating a comparative example with respect to the misalignment detection method according to the present invention. FIG. 7 is a diagram showing an example of various waveforms obtained in the comparative example shown in FIG. 4 is a flowchart illustrating a process of detecting a misalignment which is a first embodiment of the misalignment detection method according to the present invention. FIG. 9 is a diagram schematically illustrating an example of various waveforms acquired by the method of detecting a misalignment illustrated in FIG. 8 and a technique for deriving the misalignment. It is a figure showing composition of a work holding unit equipped with a 2nd embodiment of a misalignment detection device concerning the present invention. It is a top view which shows typically the positional relationship between a line sensor and a workpiece | work in the workpiece holding unit in 2nd Embodiment. 9 is a graph illustrating a function indicating a work misalignment state derived based on images captured in a first imaging region and a second imaging region in the second embodiment. It is a flowchart which shows the detection process of the misalignment which is 2nd Embodiment of the misalignment detection method which concerns on this invention. It is a figure showing the combination of the appearance mode of a characteristic part, and a center gap derivation method. It is a top view which shows typically the positional relationship between a line sensor and a workpiece | work in the workpiece holding unit in 3rd Embodiment. It is a graph which shows the function which shows the misalignment state of the work derived based on the image picturized by the 1st imaging field and the 2nd imaging field in a 3rd embodiment.

  FIG. 1 is a diagram showing an entire configuration of an inspection device equipped with a first embodiment of a misalignment detection device according to the present invention. FIG. 2 is a block diagram showing an electrical configuration of the inspection apparatus shown in FIG. This inspection apparatus 100 is an apparatus for inspecting the appearance of a work W having an outer peripheral portion in which a convex portion and a concave portion are provided in a shape rotationally symmetric about an axis of symmetry, such as a gear and an impeller, in which a convex portion and a concave portion are provided periodically. It has a loading unit 1, a work holding unit 2, an imaging unit 3, an unloading unit 4, and a control unit 5. In this case, the work W is a mechanical component having a gear Wb provided on the shaft Wa as shown in FIG. 1 and is formed by, for example, forging or casting. After the parts are manufactured, the work W is transferred to the loading unit 1 by an external transfer robot or an operator.

  The loading unit 1 is provided with a work accommodating section (not shown) such as a table and a stocker. Then, when the work W is temporarily stored in the work storage unit by an external transfer robot or the like, the work detection sensor 11 (FIG. 2) provided in the work storage unit detects the work W and sends a signal to that effect to the apparatus. The information is transmitted to the control unit 5 that controls the entire system. The loading unit 1 is provided with a loader 12 (FIG. 2), which receives an uninspected work W stored in a work storage unit in response to an operation command from the control unit 5, and Transport to

  FIG. 3 is a perspective view showing the configuration of the work holding unit. The work holding unit 2 is equipped with holding tables 21A and 21B for holding the work W transported by the loader 12. These holding tables 21A and 21B have the same configuration, and can hold and hold a part of the shaft Wa of the work W in a posture where the gear Wb is in a horizontal state. Hereinafter, the configuration of the holding table 21A will be described with reference to FIG. 3, while the holding table 21B has the same configuration as the holding table 21A.

  In the holding table 21A, as shown in FIG. 3, a chuck mechanism 22, a horizontal positioning mechanism 23, a rotating mechanism 24, and a vertical positioning mechanism 25 are vertically stacked. The chuck mechanism 22 includes movable members 221 to 223 that are substantially L-shaped in a side view, and a moving unit 224 that moves the movable members 221 to 223 in a radial manner in response to a movement command from the control unit 5. ing. A projecting member 225 protrudes from the upper end surface of each of the movable members 221 to 223, and the upper end surface and the projecting member 225 can be engaged with the step portion of the shaft Wa. For this reason, the moving unit 224 moves the movable members 221 to 223 closer to each other in response to a grip command from the control unit 5, so that the center axis of the chuck mechanism 22 (reference numeral AX <b> 2 in FIG. 5) and the axis of the shaft Wa are aligned. And the work W can be held. On the other hand, the moving unit 224 moves the movable members 221 to 223 away from each other in response to the release command from the control unit 5, so that the unloading work W is loaded by the loading unit 1 and the inspected work W is moved by the unloading unit 4. Unloading can be performed.

  The chuck mechanism 22 configured as described above is supported by the horizontal positioning mechanism 23. The horizontal positioning mechanism 23 has a so-called XY table that moves in directions orthogonal to each other in the horizontal direction. For this reason, the XY table is driven in response to the movement command from the control unit 5, and the chuck mechanism 22 can be positioned with high accuracy on a horizontal plane. As the XY table, a combination of a motor and a ball screw mechanism, a combination of two linear motors orthogonal to each other in a horizontal direction, and the like can be used.

  The rotation mechanism 24 has a motor 241. The rotation shaft (242 in FIG. 5) of the motor 241 extends vertically upward, and the horizontal positioning mechanism 23 is connected to the upper end thereof. Therefore, when a rotation command is given from the control unit 5, the motor 241 operates and the horizontal positioning mechanism 23, the chuck mechanism 22, and the chuck mechanism 22 hold the motor 241 around the rotation axis (AX <b> 3 in FIG. 5). The rotated work W is integrally rotated.

  Here, in the present embodiment, the horizontal positioning mechanism 23 is provided between the chuck mechanism 22 and the rotation mechanism 24, but the technical significance is that the center axis of the chuck mechanism 22, the work W held by the chuck mechanism 22. The point is that the relative positional relationship between the symmetrical axis (reference numeral AX4 in FIG. 5) of the gear Wb and the rotation axis of the motor 241 can be adjusted by the horizontal positioning mechanism 23. In other words, by aligning the center axis of the chuck mechanism 22 with the rotation axis of the motor 241, the work W gripped by the chuck mechanism 22 can be rotated around the shaft Wa. However, when the symmetric axis of the gear Wb is displaced from the shaft portion Wa, the motor 241 is misaligned, and the gear Wb rotates eccentrically. Therefore, by providing the horizontal positioning mechanism 23 and driving it so as to correct the shift amount and the shift direction, the symmetric axis of the gear Wb and the rotation axis of the motor 241 can be matched. Thereby, it is possible to capture an image of the gear Wb by the imaging unit 3 with high accuracy, and it is possible to improve the inspection accuracy of the work W.

  The vertical positioning mechanism 25 includes a holding plate 251 for holding the motor 241, a base plate 252 disposed below the motor 241, four connecting pins 253 for connecting the holding plate 251 and the base plate 252, and a base plate 252. And an elevating part 254 for elevating and lowering in the direction. The elevating unit 254 vertically moves the base plate 252 in response to an elevating command from the control unit 5 to move the rotation mechanism 24, the horizontal positioning mechanism 23, and the chuck mechanism 22 integrally in the vertical direction. The height position of the work W can be optimized at the PA and the inspection position PI.

  As shown in FIG. 3, the holding tables 21A and 21B configured as described above are fixed on the support plate 261 with a predetermined distance therebetween. Further, the support plate 261 is supported by the turning drive unit 262 at an intermediate position between the holding tables 21A and 21B. The turning drive unit 262 is capable of turning the support plate 261 by 180 ° around a turning axis AX1 extending in the vertical direction in response to a turning command from the control unit 5, and the holding tables 21A and 21B are rotated as shown in FIG. It is possible to switch between a first position located at the pre-alignment position PA and the inspection position PI, and a second position at which the holding tables 21A and 21B are located at the inspection position PI and the pre-alignment position PA, respectively. For example, in parallel with performing the pre-alignment process on the work W held on the holding table 21A located at the pre-alignment position PA, the holding table 21A is switched from the first position to the second position by the turning drive unit 262. Shifts from the pre-alignment position PA to the inspection position PI, and the pre-aligned work W can be positioned at the inspection position PI. After the inspection of the work W is completed, the holding table 21A is shifted from the inspection position PI to the pre-alignment position PA by turning in the reverse direction, and the inspection-processed work W is positioned at the pre-alignment position PA. Can be. As described above, in the present embodiment, the position switching mechanism 26 that switches the position of the work W by the support plate 261 and the turning drive unit 262 is configured.

  The pre-alignment position PA is a position where the pre-alignment processing is performed as described above, and the alignment camera 27 is disposed above the holding table 21A (or 21B) positioned at the pre-alignment position PA. As shown in FIG. 3, the alignment camera 27 is disposed on the opposite side of the motor 241 with respect to the workpiece W, that is, above the workpiece W, and extends radially outward with respect to the symmetry axis AX4 of the workpiece W. Line sensor 271. For this reason, the upper surface of the work W can be imaged by the line sensor 271 while rotating the work W, and by rotating the work W at least once, a convex portion (tooth end) formed on the outer peripheral portion of the gear Wb is formed. ) And the concave portion (root) are obtained.

  Although not shown in FIG. 3, an alignment illumination unit 28 (FIG. 2) is provided for illuminating the work W held on the holding table 21A (or 21B) and performing an alignment process satisfactorily. Have been. Therefore, the work W can be imaged by the alignment camera 27 while the work W is rotated by the rotation mechanism 24 and the work W is illuminated by the alignment illumination unit 28. Then, the image data of the work W is sent to the control unit 5, and the misalignment is corrected so that the symmetry axis of the gear Wb and the rotation axis of the motor 241 are matched, that is, a pre-alignment process is performed.

  On the other hand, the inspection position PI is a position where the inspection processing is performed, and the imaging unit 3 is disposed above the holding table 21A (or 21B) positioned at the inspection position PI. At the inspection position PI, the imaging unit 3 can image the workpiece W while rotating the workpiece W in a state where the axis of symmetry of the gear Wb and the rotation axis of the motor 241 coincide. Then, the image data of the work W is sent to the control unit 5, and an inspection process for inspecting the gear Wb for the presence or absence of a flaw or a defect is executed.

  This imaging unit 3 has a plurality of inspection cameras 31 and a plurality of inspection illumination units 32, as shown in FIG. In the imaging unit 3, a plurality of inspection illumination units 32 are arranged so as to illuminate the work W held on the holding table 21A (or 21B) positioned at the inspection position PI from various directions. The rotating mechanism 24 rotates the workpiece W, and the inspection illumination unit 32 illuminates the workpiece W, so that the inspection camera 31 can image the workpiece W from various directions. The plurality of pieces of image data are sent to the control unit 5, and the control unit 5 performs inspection of the work W.

  The holding table 21A (or 21B) holding the work W thus inspected is shifted from the inspection position PI to the pre-alignment position PA by the position switching mechanism 26 as described above. Then, the unloaded unit 4 unloads the inspected work W from the holding table 21A (or 21B). The unloading unit 4 is basically the same as the loading unit 1. That is, the unloading unit 4 includes a work storage unit (not shown) for temporarily storing the inspected work W, a work detection sensor 41 (FIG. 2), and an unloader 42 (FIG. 2). In response to the operation command from the work 5, the inspected work W is transported from the holding table 21A (or 21B) to the work accommodating section.

  As shown in FIG. 2, the control unit 5 includes a well-known CPU (Central Processing Unit) for executing a logical operation, a ROM (Read Only Memory) for storing initial settings and the like, and temporarily stores various data during operation of the apparatus. It is composed of a RAM (Random Access Memory) and the like for temporarily storing. The control unit 5 functionally includes an arithmetic processing unit 51, a memory 52, a drive control unit 53, an external input / output unit 54, an image processing unit 55, and an illumination control unit 56.

  The drive control unit 53 controls the drive of a drive mechanism provided in each unit of the apparatus, for example, the loader 12, the chuck mechanism 22, and the like. The external input / output unit 54 inputs signals from various sensors provided in each unit of the device, and outputs signals to various actuators and the like provided in each unit of the device. The image processing unit 55 captures image data from the alignment camera 27 and the inspection camera 31 and performs image processing such as binarization. The illumination control unit 56 controls turning on and off of the alignment illumination unit 28 and the inspection illumination unit 32, and the like.

  The arithmetic processing unit 51 has an arithmetic function, and controls a drive control unit 53, an image processing unit 55, an illumination control unit 56, and the like in accordance with a program stored in the memory 52, so that a series of operations will be described below. Execute the processing of

  Reference numeral 6 in FIG. 2 denotes a display unit functioning as an interface with the operator, which is connected to the control unit 5 and has a function of displaying an operation state of the inspection apparatus 100, and is constituted by a touch panel and has an input from the operator. It also has a function as an input terminal that accepts the information. Further, the present invention is not limited to this configuration, and a display device for displaying an operation state and an input terminal such as a keyboard and a mouse may be employed.

  FIG. 4 is a flowchart showing the inspection operation of the work by the inspection device shown in FIG. FIG. 5 is a diagram schematically showing the inspection operation. In FIG. 5, dots are given to the holding table 21B and the work W held by the holding table 21B in order to clearly distinguish the operations of the holding tables 21A and 21B.

  In the inspection apparatus 100, the arithmetic processing unit 51 controls each unit of the apparatus according to an inspection program stored in the memory 52 of the control unit 5 in advance, and executes the following operation. Here, various operations performed on one work W while focusing on the work W will be described with reference to FIGS. 4 and 5. The control unit 5 determines that the workpiece W is not present on the holding table 21A located at the pre-alignment position PA as shown in the column (a) of FIG. When it is confirmed that the work W is accommodated in the first work accommodation section, loading of the work W onto the holding table 21A is started (step S1). In this loading step, the loader 12 grips the uninspected work W in the work accommodating portion and transports the work W from the loading unit 1 to the holding table 21A. In the present embodiment, in order to smoothly perform the loading process and the subsequent misalignment detection process, before the transfer of the work W to the holding table 21A, the horizontal positioning mechanism is used as shown in FIG. 23, the center axis AX2 of the chuck mechanism 22 and the rotation axis AX3 of the motor 241 are made to coincide with each other, and the three movable members 221 to 223 are separated from each other to prepare for receiving the work W.

  When the work W is conveyed to the holding table 21A by the loader 12, the chuck mechanism 22 moves the three movable members 221 to 223 closer to each other as described above to sandwich a part of the shaft Wa of the work W. The work W is gripped. More specifically, during the loading operation, the movable members 221 to 223 move close to each other, and the upper end surfaces of the movable members 221 to 223 and the projecting members 225 engage with the step portions of the shaft Wa, and The workpiece W is held while the center axis AX2 is aligned with the axis of the shaft Wa (see the column (b) in FIG. 5). Thus, the loading step is completed. At this point, the rotation axis AX3 of the motor 241 is coincident with the center axis AX2 of the chuck mechanism 22 and the axis of the shaft Wa. However, in the workpiece W manufactured by forging or casting, for example, the symmetric axis AX4 of the gear Wb is displaced from the axis of the shaft Wa as shown in FIG. May have occurred.

  Therefore, in the present embodiment, while the untested work W is illuminated by the alignment lighting unit 28 (FIG. 2), the gear Wb is imaged by the alignment camera 27 while rotating the untested work W by the motor 241 of the holding table 21A. The image data is stored in the memory 52 (step S2).

  After the completion of the imaging, the rotation drive unit 262 switches the first position to the second position. That is, the turning drive unit 262 turns the support plate 261 by 180 ° about the turning axis AX1, and as a result, the holding table 21A holding the untested work W is moved to the pre-alignment position PA as shown in FIG. To the inspection position PI, and at the same time, the work W is moved to a height position where the image pickup unit 3 can image the work W by the lifting / lowering unit 254 (step S3).

  Further, in the present embodiment, in parallel with the movement, the image data of the work W is read from the memory 52, and the work W is misaligned with respect to the rotating mechanism 24 (motor 241) (in the present embodiment, the misalignment amount Δ and the misalignment direction). Is detected (step S4), and subsequently, the misalignment of the holding table 21A is corrected (step S5). The detection process of the misalignment (step S4) will be described later in detail. For the misalignment correction, the chuck mechanism 22 is moved by the horizontal positioning mechanism 23 so as to eliminate the misalignment detected in step S4. Thereby, as shown in the column (c) of FIG. 5, the symmetric axis of the gear Wb and the rotation axis of the motor 241 coincide with each other when the holding table 21A arrives at the inspection position PI or before and after the arrival, and immediately after the workpiece imaging step. (Step S6) can be started.

  In this step S6, the rotation mechanism 24 of the holding table 21A positioned at the inspection position PI operates to start the work rotation. At this time, the work W held on the holding table 21A is in a so-called centering state that has been subjected to the above-described misalignment correction, and rotates around the symmetry axis AX4. In addition, the plurality of inspection illumination units 32 are turned on in response to the rotation, and illuminate the rotating workpiece W from a plurality of directions. Here, the inspection lighting unit 32 is turned on after the work is rotated, but the lighting timing is not limited to this, and the lighting of the inspection lighting unit 32 is started at the same time as the start of rotation or before the start of rotation. Is also good.

  While the work W is being rotated and illuminated in this way, the plurality of inspection cameras 31 image the work W from various directions, and image data of images of the work W from multiple directions (hereinafter, referred to as “work images”). Is transmitted to the control unit 5. On the other hand, the control unit 5 stores the image data in the memory 52, and inspects the work W based on the image data at the following timing.

  After such image acquisition, the work rotation is stopped in the holding table 21A, and the inspection illumination unit 32 is turned off in the imaging unit 3. Further, the turning drive unit 262 turns the support plate 261 by 180 ° inversion about the turning axis AX1, whereby the holding table 21A moves from the inspection position PI to the pre-alignment position PA while holding the inspected work W, and moves up and down. The work W is moved to the original height position by the part 254 (step S7). In parallel with the movement of the work W, the control unit 5 reads out the image data from the memory 52, determines whether the gear Wb has a flaw or a defect based on the work image, and holds the gear Wb in the holding table 21A. A work inspection is performed on the work W thus performed (step S8).

  The work W returned to the pre-alignment position PA is gripped by the unloader 42, and is then transferred from the holding table 21A to the unloader 42 by releasing the gripping by the movable members 221 to 223. Subsequently, the unloader 42 conveys the work W to the unloading unit 4 and conveys it to the work storage unit (not shown) (Step S9). The above-described series of steps (Steps S1 to S9) are alternately repeated by the holding tables 21A and 21B.

  Next, a description will be given of a misalignment detection step (Step S4) executed as one aspect of the misalignment detection method according to the present invention in the present embodiment. However, in order to have the contents of the present invention accurately understood, First, a comparative example of detecting the misalignment by rotating the work W one round in the above-described inspection apparatus 100 will be described with reference to FIGS. 6 and 7.

  FIG. 6 is a flowchart showing a comparative example with respect to the misalignment detection method according to the present invention, and FIG. 7 is a diagram showing an example of various waveforms obtained by the misalignment detection method shown in FIG. The misalignment detection step is executed by the arithmetic processing unit 51 operating as follows in accordance with the misalignment detection program stored in the memory 52 of the control unit 5 in advance.

  In this comparative example, the center axis AX2 of the chuck mechanism 22 and the rotation axis AX3 of the motor 241 are made to coincide with each other by the horizontal positioning mechanism 23 prior to executing the misalignment detection step, and the rotation is held by the chuck mechanism 22. The shaft Wa of the work W matches the rotation axis AX3 of the motor 241. However, in the workpiece W formed by forging or the like, the symmetric axis AX4 of the gear Wb may be displaced from the axis of the shaft Wa, and the motor 241 may be misaligned. Therefore, in the present comparative example, the misalignment of the workpiece W loaded on the holding table 21A is detected by a procedure described in detail below. Note that the misalignment of the work W held by the holding table 21B is detected in the same manner, and a description of the misalignment detection of the work W will be omitted.

  When the loading of the uninspected work W onto the holding table 21A is completed, the gear Wb is imaged from above by the alignment camera 27 while the work W is rotated while the work W is illuminated by the alignment illumination unit 28. Thereby, the work image I1 for one rotation of the gear Wb is obtained, and the image data of the work image I1 is sent to the image processing unit 55 (step S401). The image processing unit 55 creates a continuous image I3 for three rounds of the work by connecting three work images I1 (step S402). Further, a noise component is removed by a smoothing filter process to create a smoothed image Is (step S403). Further, the smoothed image Is is converted into a binary image by a binarization process such as Otsu's method, and the binary image data of a region corresponding to the gear Wb of the work W (hereinafter, referred to as an “edge region”) is converted into edge image data. (Step S404).

  Next, the arithmetic processing unit 51 executes the following series of processes (steps S405 to S412) to determine the center deviation of the work W from the edge image data. That is, by converting the edge image data indicating the edge region into the run-length data, an edge waveform F indicating the shape of the edge region for three rounds is derived as shown in, for example, a column (a) of FIG. S405). The horizontal axis X in each of the graphs shown in the columns (a) and (b) of FIG. 7 indicates the pixel position of the smoothed image Is (corresponding to the rotation angle of the work W and the position of the gear Wb in the circumferential direction), The vertical axis Y indicates the edge position in the longitudinal direction of the line sensor 271 (the radial direction of the gear Wb). The method of deriving the edge waveform F is not limited to the above-described conversion of the edge image data into the run-length data, and the edge waveform F may be derived by another method. This is the same in the embodiment described later.

  Further, a moving average process is performed on the edge waveform F to derive a moving average waveform Fs shown in the column (b) of FIG. Here, in the gear Wb of the work W (corresponding to an example of the “outer peripheral portion of the work” of the present invention), the convex portion and the concave portion are repeatedly provided, and there is a periodic phase. Assuming that the number of pixels of the phase (the range of the rotation angle) is L1, the filter size of the moving average filter is set to (L1 / 2). Each phase in the moving average waveform Fs thus obtained has one maximum value and one minimum value, one of which corresponds to the tooth tip position of the gear Wb, and the other corresponds to the tooth bottom position. .

  In this comparative example, the local maximum value of each phase included in the moving average waveform Fs is extracted, and a function Fn indicating a change in the tooth tip position (or the tooth bottom position) accompanying the rotation of the work W is derived from those discrete points. (Step S407). The graph shown in the column (c) of FIG. 7 is an example of the function Fn, and the sharp left end of the graph is a problem in data processing when deriving the function. Does not indicate the tip position (bottom position). Therefore, in this comparative example, the average value AV of the function Fn is derived except for the steep data at the left end (step S408). Here, as is apparent from the column (c) of FIG. 7, when the misalignment has not occurred, the tooth tip position (the tooth bottom position) has a line of the average value AV (regardless of the rotation angle of the work W). (Broken line in the figure). On the other hand, when the misalignment occurs, the work W fluctuates in the rotation cycle of the work W as shown by the solid line in the figure, and the maximum value of the fluctuation amount from the average value AV corresponds to the deviation amount Δ (see FIG. 5). , The rotation angle indicating the maximum value corresponds to the deviation direction.

Therefore, in this comparative example, a function Fz indicating the relative variation from the average value AV for one round of the work W is extracted as shown in a section (d) of FIG. 7 (step S409), and based on the function Fz. A deviation direction is derived (step S410), and a deviation amount Δ is derived (step S411). More specifically, based on the pixel position Xmax at which the function Fz shows the maximum and the number L of pixels corresponding to one round of the work W, the shift direction is expressed by the following equation: shift direction = Xmax / L × 360 °
It is demanded by. Of course, the pixel position Xmin at which the function Fz shows the minimum may be used instead of the pixel position Xmax. Further, the amplitude A is calculated from the maximum value Ymax and the minimum value Ymin of the function Fz as follows: A = (Ymax−Ymin) / 2
And the value obtained by multiplying the amplitude A by the resolution of the line sensor 271 is defined as the deviation amount Δ.

  As described above, in this comparative example, edge image data (image data of the outer peripheral portion of the gear Wb) is obtained from the smoothed image Is created based on the work image I1 obtained by imaging while rotating the work W. At the same time, the edge image data is converted into run-length data to obtain an edge waveform F. The edge waveform F accurately indicates the shape of the outer peripheral portion of the gear Wb, and the misalignment of the workpiece W with respect to the rotation axis AX3 of the motor 241 can be calculated with high accuracy based on the edge waveform F. Then, by driving the horizontal positioning mechanism 23 so as to eliminate the misalignment (= the misalignment direction + the misalignment amount Δ) derived in this manner, the misalignment can be satisfactorily corrected. Furthermore, since the workpiece W is imaged while rotating the workpiece W after correcting the misalignment, the workpiece W is inspected, so that the workpiece inspection can be performed with high accuracy.

  Further, in this comparative example, a continuous image I3 is created by connecting three work images I1 for one round of the gear Wb, and the misalignment is detected based on the continuous image I3. The effect is obtained. For example, the continuous image I3 may be acquired while rotating the work W at the pre-alignment position PA three times, but the rotation amount of the work W required for detecting the misalignment is large, and the detection time of the misalignment becomes long. On the other hand, in this comparative example, the rotation amount of the work W can be set to one rotation, and the time required for detecting the misalignment can be reduced. However, also in this comparative example, it is necessary to rotate the work W once, and it is desired to reduce the time.

  Therefore, in the present embodiment, the step of detecting the misalignment (step S4) shown in FIGS. 8 and 9 is executed. FIG. 8 is a flowchart showing a process of detecting a misalignment which is a first embodiment of the misalignment detecting method according to the present invention. FIG. 9 shows an example of various waveforms obtained by the misalignment detecting method shown in FIG. It is a figure which shows the derivation | leading-out method of a center gap typically. The misalignment detection step is executed by the arithmetic processing unit 51 operating as follows according to the misalignment detection program stored in the memory 52 of the control unit 5 in advance. In the following, similarly to the description of the comparative example, the detection of the misalignment of the work W loaded on the holding table 21A will be described, and the description of the detection of the misalignment of the work W held on the holding table 21B will be omitted. I do. This applies to the second embodiment and the third embodiment described later.

  In the present embodiment, the moving average waveform Fs0 (dotted line in the column (c) of FIG. 9) for the work W in the centered state is obtained for various works and stored in the memory 52. Then, the moving average waveform Fs0 corresponding to the type of the work W loaded on the holding table 21A is read from the memory 52 (Step S421). When the detection of the misalignment is repeatedly performed for the same type of work W, the moving average waveform Fs0 may be obtained only for the first time (step S421), and after that, step S421 may be omitted. This also applies to the second and third embodiments described later.

  When the loading of the uninspected work W onto the holding table 21A is completed, the center axis AX2 of the chuck mechanism 22, the rotation axis AX3 of the motor 241, and the axis of the shaft Wa coincide with each other, as in the comparative example. I have. Then, while the work W is being rotated while the work W is being illuminated by the alignment illumination unit 28, the gear Wb is imaged from above by the alignment camera 27. In the present embodiment, in order to shorten the detection time, the work W is rotated by the work rotation angle θ to acquire the work partial image I11 for the angle φ of the gear Wb (step S422). In this embodiment, the rotation angle θ is set to 90 ° in order to obtain a workpiece partial image I11 for φ = 90 °. Here, the reason for setting φ = 90 ° is as follows.

  As described in detail in the comparative example, when the edge waveform F11 is obtained from the image I1 for one round of the outer circumference of the workpiece W, four characteristic portions (that is, four characteristic portions) are obtained for one round of the edge waveform F11 as shown in FIG. , The inflection point b1, the extremum a1, the inflection point b2, and the extremum a2) appear in order at 90 ° intervals. Each feature corresponds to a center deviation, and the center waveform is compared by comparing an edge waveform (or a moving average waveform) corresponding to the feature with a previously centered edge waveform (or a moving average waveform). The deviation can be derived. This means that the workpiece partial image I11 for 90 ° includes information sufficient to detect the misalignment, and in the present embodiment, the misalignment is detected using this. In the present embodiment, a moving average waveform having the same quality as the edge waveform is used to eliminate the influence of noise and improve detection accuracy.

  Thus, the work partial image I11 of φ (= 90 °) of the gear Wb is obtained, and the image data of the work partial image I11 is sent to the image processing unit 55. The image processing unit 55 creates a smoothed image Is1 by removing noise components by a smoothing filter process (step S423). Further, the smoothed image Is1 is converted into a binary image by a binarization process such as Otsu's method, and the binary image data of the edge area of the work W is extracted as edge image data (step S424).

  Next, the arithmetic processing unit 51 executes the following series of processing (steps S425 to S431), thereby obtaining the center deviation of the work W based on the characteristic part. That is, by converting the edge image data indicating the edge area into the run-length data, an edge waveform F11 indicating the shape of the edge area of 90 ° is derived, for example, as shown in column (a) of FIG. S425). As described above, the characteristic portions can be divided into two types, extreme values a1 and a2, and inflection points b1 and b2, and the principle of detecting the misalignment differs for each type. Derivation of misalignment based on value "and" Derivation of misalignment based on inflection point "are illustrated separately. In each graph shown in FIG. 9, the horizontal axis X indicates the pixel position of the image (corresponding to the rotation angle of the work W and the position in the circumferential direction of the gear Wb), and the vertical axis Y indicates the longitudinal direction of the line sensor 271 (the gear Wb (In the radial direction).

  The moving average processing is performed on the edge waveform F11 to derive a moving average waveform Fs1 shown in the column (b) of FIG. 7 (step S426). Here, the convex portion and the concave portion are repeatedly provided on the outer peripheral portion of the gear Wb of the workpiece W, and since there is a periodic phase, the filter size of the moving average filter is set to (L1 / 2) as in the comparative example. You have set. Each phase in the moving average waveform Fs1 thus obtained has one maximum value and one minimum value, one of which corresponds to the tooth tip position (the position of the convex portion) of the gear Wb and the other one. This corresponds to the tooth bottom position (the position of the concave portion).

Next, as shown in columns (b) and (c) of FIG. 9, extreme points of each phase included in the moving average waveform Fs1 are extracted (step S427). Then, based on the relative relationship between three consecutive extreme points E1 to E3 indicating the maximum values, it is determined whether the characteristic portions included in the work partial image I11 are the extreme values a1 and a2 and the inflection points b1 and b2. Is selected. More specifically, as shown in FIG. 9, the Y coordinate values of the extreme points E1 to E3, that is, the edge positions Y1 to Y3 have the following relationship:
Y2> Y1 and Y2> Y3
Or
Y2 <Y1 and Y2 <Y3
In the case of (1), it is determined that the type is an extreme value type that uses "derivation of the center deviation based on the extreme value" as the method of deriving the center deviation. On the other hand, the edge positions Y1 to Y3 of the extreme points E1 to E3 have the following relationship:
Y1 <Y2 <Y3
Or
Y1>Y2> Y3
In the case of (2), it is determined that the inflection point type uses "derivation of center deviation based on inflection point" as a method of deriving the center deviation (step S428). When it is determined in step S428 that the type is the extreme value type, the process proceeds to step S429 to derive the center deviation based on the extreme value, while when it is determined that the type is the inflection point type, the process proceeds to step S430. The center deviation based on the inflection point is derived.

In step S429 (the derivation of the center deviation based on the extreme value), the X coordinate value (pixel position X2 in the line sensor 271) and the Y coordinate value (edge position Y2) of the extreme value point E2 are obtained. Then, the shift direction is expressed by the following equation: shift direction = X2 / L × 360 °
Here, L is obtained from the number of pixels corresponding to one round of the work W. In addition, the shift amount is calculated by the following equation: shift amount Δ = Y2-Yav
However, Yav is obtained from the average value of the local maximum values of each phase included in the moving average waveform Fs0 of the workpiece W in the centered state. Note that a minimum value may be used instead of the maximum value.

On the other hand, in step S430 (center misalignment based on the inflection point), a line (straight line or curved line) connecting the maximum values in the moving average waveform Fs0 of the work W in the centered state as shown in the column (c) of FIG. 3.) LN0 and a line (straight line or curve) LN1 connecting the maximum values in the moving average waveform Fs1 of the work W to be inspected are obtained. In FIG. 9, a straight line connecting the extreme points E1 and E2 is defined as a line LN1. Then, after calculating the coordinates (Xc, Yc) of the intersection CP of the two lines LN0 and LN1, the inclination T of the line LN1 at the intersection CP is calculated by the following equation: T = (Y1-Y2) / (X1-X2)
Obtain according to. Further, the deviation amount Δ is calculated by the following equation: Δ = T / ω
Where ω = 2π / L
Is derived according to Here, the reason why the deviation amount Δ can be obtained by the above equation will be described. Assuming that a curve formed by connecting adjacent extreme points E1 and E2 as described above is a sine wave, a function Fz representing a center deviation is given by: Fz = A · sin (ωX + k)
Where A is the amplitude (corresponding to the shift amount), k is the shift direction,
Can be defined as Then, the differential value Fz ′ of the function Fz is given by: Fz ′ = A · ω · cos (ωX + k)
Becomes Since the slope of this waveform is cos (ωX + k) = 1 at the inflection point,
Fz '= T = A · ω
Becomes Therefore, the shift amount Δ can be obtained by the above equation. Further, since the number of pixels for one round of the work W is L, the displacement direction k is k = X2 / L × 2π + π / 2.
Becomes

  As described above, in the first embodiment, the work W is rotated by 90 ° to capture a work partial image corresponding to the angle φ (= 90 °) of the outer peripheral portion of the work W, and the work W is rotated based on the work partial image. An edge waveform F that partially shows the shape of the outer peripheral portion is obtained, and the center deviation of the workpiece W with respect to the rotation mechanism 24 is calculated based on the edge waveform F. Therefore, the misalignment of the work W can be detected in a short time with high accuracy.

  By the way, in the first embodiment, a work partial image of 90 ° is acquired, but the same operation and effect can be obtained even if a work partial image of an angle exceeding 90 ° (but less than 360 °) is acquired. Is obtained. However, since the angle θ for rotating the work W increases as the angle φ increases, it is desirable to set the angle φ to a value close to 90 ° in order to reduce the time required for detecting misalignment. Further, although the angle is the same, the angle θ can be halved by increasing the imaging area as described below, and the misalignment detection time can be further reduced (second and third embodiments). .

  FIG. 10 is a view showing a configuration of a work holding unit equipped with a second embodiment of the misalignment detecting device according to the present invention. FIG. 11 is a plan view schematically showing the positional relationship between the line sensor and the work in the work holding unit shown in FIG. The inspection apparatus 100 according to the second embodiment differs from the first embodiment in the arrangement of the line sensor 271 and the processing content in the misalignment detection step, and other configurations are basically the same as those in the first embodiment. Is the same as Therefore, the following description will focus on the differences, and the same components will be denoted by the same reference numerals and description thereof will be omitted.

In the second embodiment, the line sensor 271 is arranged offset by a predetermined distance d1 with respect to the rotation axis AX3 of the motor 241. The offset distance d1 is given by the following equation: offset distance d1 = R · sin (φ / 4)
Here, R is the radius of the gear Wb of the work W (the distance from the symmetry axis AX4 to the outer peripheral surface of the gear),
In the second embodiment in which the angle φ is set to 90 °, as in the first embodiment, the line sensor 271 has an offset distance d1 = R · sin (90 ° / 4)
The motor 241 is disposed apart from the rotation axis AX3. For this reason, the arrangement direction of the light receiving elements (not shown) of the line sensor 271 is orthogonal to the radial direction of the rotation axis AX3, and the two imaging regions AR1 provided linearly while being separated by a certain distance in the orthogonal direction. , AR2, the outer peripheral portion of the work W can be locally imaged. Therefore, a function Fz1 indicating the state of the center shift of the work W obtained by the same method as in the first embodiment based on the work image imaged in the first imaging area AR1 while the work W in the state of the center shift makes one round. Is a graph shown in the upper part of FIG. Similarly, a function Fz2 indicating the state of misalignment of the workpiece W based on the workpiece image captured in the second imaging area AR2 is a graph illustrated in the lower part of FIG.

  As is clear from both graphs, in the inspection apparatus 100 configured as described above, even when the rotation angle θ of the workpiece W is set to half the angle φ, the angle (φ / φ) is determined by the first imaging region AR1. 2) A work local image (hereinafter, referred to as a “first work local image”) is imaged at the same time as the work local image corresponding to an angle (φ / 2) by the second imaging region AR2 (hereinafter, “a second work local image”). Is imaged. Moreover, since the line sensor 271 is offset by the offset distance d1 with respect to the rotation axis AX3 of the motor 241, a work partial image (= first work local image + A second work local image) is obtained, and a characteristic portion is always included in an edge waveform or a moving average waveform obtained from one of the first work local image and the second work local image (see FIG. 14 described later). ). For example, in the example shown in FIG. 12, the extreme value a2 appears in the second work local image captured in the second imaging region AR2 in the range (angle of φ / 2) sandwiched by the broken lines. Therefore, the center deviation can be obtained based on three consecutive extreme points included in the edge waveform and the moving average waveform created based on the second work local image.

  As described above, in the second embodiment, it is possible to detect the misalignment of the work W while setting the rotation angle θ of the work W to half the angle φ. Specifically, after detecting the misalignment according to the flowchart shown in FIG. 13, the inspection device 100 performs steps S3 to S7 to inspect the work W as in the first embodiment.

  FIG. 13 is a flowchart showing a process of detecting a misalignment which is a second embodiment of the misalignment detecting method according to the present invention, and FIG. 14 is a diagram showing a combination of a feature part appearance mode and a misalignment derivation method. . This misalignment detection step is executed by the arithmetic processing unit 51 operating as follows in accordance with the misalignment detection program stored in the memory 52 of the control unit 5 in advance.

  In the second embodiment, as in the first embodiment, a moving average waveform Fs0 for the centered work W is obtained for each type of work and stored in the memory 52. Then, the moving average waveform Fs0 is read from the memory 52 corresponding to the type of the work W loaded on the holding table 21A (Step S441).

  When the loading of the uninspected work W onto the holding table 21A is completed, the center axis AX2 of the chuck mechanism 22 and the rotation axis AX3 of the motor 241 coincide with each other, as in the comparative example. Then, while the work W is being rotated while the work W is being illuminated by the alignment illumination unit 28, the gear Wb is imaged from above by the alignment camera 27. In the present embodiment, in order to further reduce the detection time, the work rotation angle θ is set to an angle (φ / 2), and the rotation angle θ of the gear Wb (φ / 2) during rotation for each of the imaging regions AR1 and AR2. Is obtained (step S442). In this embodiment, the rotation angle θ is set to 45 ° in order to obtain a work partial image for φ = 90 °, and the first work local image for an angle (φ / 2) is set by the first imaging region AR1. And a second work local image corresponding to an angle (φ / 2) by the second imaging area AR2, and the image data thereof is sent to the image processing unit 55. The image processing unit 55 creates a smoothed image by removing noise components by a smoothing filter process for each work local image (step S443). Further, for each smoothed image, the smoothed image is converted into a binary image by a binarization process such as Otsu's method, and the binary image data of the edge region of the work W is extracted as edge image data (step S444). .

  Next, the arithmetic processing unit 51 executes the following series of processing (steps S445 to S451) to determine the center deviation of the work W based on the characteristic part. That is, by converting the edge image data into run-length data for each edge region, an edge waveform indicating the shape of the 45-degree edge region is derived (step S445). Further, a moving average process is performed for each edge waveform to derive a moving average waveform (step S446).

  Next, an extreme point of each phase is extracted for each moving average waveform (step S447). Then, it is determined whether or not a characteristic portion is included in the first work local image based on a relative relationship between three consecutive extreme points indicating the maximum value among the extreme points corresponding to the first work local image. I do. When a characteristic part is included, it is determined whether the characteristic part is the extreme value a1, a2 or the inflection point b1, b2. The determination is also made for the second work local image. Here, the method described in the first embodiment can be used for determining the characteristic portion. On the other hand, the determination that the characteristic portion does not exist is made based on the extreme points E1 to E3 and the average Yav of the maximum value of each phase included in the moving average waveform Fs0 of the workpiece W in the centered state. be able to. That is, when it is confirmed that the extreme points E1 to E3 are monotonically increasing or monotonically decreasing, and that the Y coordinate values of the extreme points E1 to E3 are all larger or smaller than the average value Yav, the characteristic portion is determined. It can be determined that it does not exist.

  As described above, in the present embodiment, one of the first work local image and the second work local image does not have a characteristic part, but the other has a characteristic part, and the characteristic part appearance pattern is shown in FIG. It is a pattern. Therefore, in the present embodiment, the center deviation deriving method is selected according to the appearance pattern of the characteristic portion. In other words, when the extreme value a1 (or a2) appears in one of the first work local image and the second work local image, the above-described “derivation of the center shift based on the extreme value” is described as the center shift derivation method. On the other hand, when the inflection point b1 (or b2) appears, the above-described “derivation of the center deviation based on the inflection point” is selected as the center deviation derivation method, and the process proceeds to step S449 to obtain the extreme value. On the other hand, when it is determined that the type is the inflection point type, the process proceeds to step S450 to derive the center deviation based on the inflection point. Note that these deriving methods are the same as those in the first embodiment, and thus redundant description will be omitted.

  As described above, in the second embodiment, a workpiece partial image is captured in the same manner as in the first embodiment, and an edge waveform partially indicating the shape of the outer peripheral portion of the workpiece W is obtained based on the workpiece partial image. The misalignment of the workpiece W with respect to the rotation mechanism 24 is calculated based on the edge waveform. For this reason, the misalignment of the work W can be detected with high accuracy. In addition, since the rotation angle θ of the work W is set to half of that in the first embodiment, the time required for detecting the misalignment of the work W can be further reduced.

In the second embodiment, the offset distance of the line sensor 271 is set to the distance d1, but as shown in FIG. 15, the offset distance is set to the offset distance d2 = R · cos (φ / 4)
(Third embodiment).

  In the third embodiment, a work obtained by a method similar to that of the first embodiment based on a first work local image captured in the first imaging area AR1 while the work W in a misaligned state makes one round. The function Fz1 indicating the W misalignment state is a graph shown in the upper part of FIG. Similarly, a function Fz2 indicating the state of misalignment of the work W obtained based on the second work local image captured in the second imaging area AR2 is a graph shown in the lower part of FIG.

  As is clear from both graphs, in the inspection apparatus 100 configured as described above, even when the rotation angle θ of the work W is set to half the angle φ, the first work local image is captured. At the same time, a second work local image is captured. In addition, since the line sensor 271 is offset by the offset distance d2 with respect to the rotation axis AX3 of the motor 241, a work partial image corresponding to a total angle φ (90 ° in the present embodiment) is obtained, and the first work An edge waveform or a moving average waveform obtained from one of the local image and the second work local image always includes at least one characteristic portion. For example, in the example shown in FIG. 16, the inflection point b1 appears in the second work local image captured in the second imaging area AR2 in the angle range (φ / 2) sandwiched by the broken lines. Therefore, the center deviation can be obtained based on three consecutive extreme points included in the edge waveform or the moving average waveform created based on the second work local image.

  As described above, also in the third embodiment, it is possible to detect the misalignment of the work W with high accuracy in a shorter time than in the first embodiment. As described above, the second embodiment and the third embodiment execute the same configuration and the same operation except for the offset distance. Further, in the second and third embodiments, the angle φ is set to 90 °, but the angle φ is optional within a range of 90 ° or more and less than 360 °. However, the offset distances d1 and d2 differ depending on the value of the angle φ. For example, when the angle φ is 90 °, the offset distance d1 is shorter than the offset distance d2, and the second embodiment is advantageous in terms of technical significance of reducing distortion of the work local image. When the angle φ increases and the offset distance d2 becomes shorter than the offset distance d1, the third embodiment is more advantageous. Therefore, when setting the apparatus, it is desirable to adopt the shorter embodiment in consideration of the offset distances d1 and d2.

  The rotating mechanism 24 in the above embodiment corresponds to an example of the “rotating unit” of the present invention, and the alignment camera 27 corresponds to an example of the “imaging unit” of the present invention. In addition, the arithmetic processing unit 51 functions as the “edge waveform deriving unit” and the “center deviation calculating unit” of the present invention. The coordinate values of the extreme points E1 to E3 correspond to an example of the “position information” of the present invention. The rotation mechanism 24, the alignment camera 27, the arithmetic processing unit 51, and the image processing unit 55 constitute the "center misalignment detection device" of the present invention. The holding tables 21A and 21B correspond to an example of the "work holding unit" of the present invention. Further, the moving average waveform corresponds to an example of the “edge waveform” of the present invention.

  The present invention is not limited to the above-described embodiment, and various changes other than those described above can be made without departing from the gist of the present invention. For example, in the above embodiment, the extreme points E1 to E3 indicating the maximum value are used for determining the characteristic portion and deriving the center deviation, but the extreme points E1 to E3 indicating the minimum value may be used.

  In the above-described embodiment, the workpiece W having the gear Wb is set as the detection target. However, by using the present invention, the outer periphery where the convex portions and the concave portions are provided in a shape rotationally symmetric about the axis of symmetry and are periodically repeated. The misalignment can be detected with high accuracy for the entire work having the portion.

  In the above embodiment, the misalignment detection device according to the present invention is applied to the inspection device 100 that detects the misalignment by alternately positioning the two holding tables 21A and 21B at the pre-alignment position PA. The present invention can also be applied to an inspection device having one or three or more holding tables. Further, in the above-described embodiment, the present invention is applied to the inspection apparatus 100 in which the pre-alignment position PA is separated from the inspection position PI. The present invention can also be applied to an inspection apparatus that performs an inspection process after performing detection and misalignment correction. In the inspection device configured as described above, a part of the inspection camera 31 may also function as the alignment camera 27.

  Further, in the above-described embodiment, one of the four characteristic portions is specified with a small amount of information in order to detect the misalignment. However, a sine wave ( (See column (d) of FIG. 7, FIGS. 12 and 16). That is, the function Fz can be derived without rotating the work W by one rotation.

  INDUSTRIAL APPLICABILITY The present invention can be applied to all misalignment detection techniques for detecting misalignment of a work having a rotation portion symmetrical about an axis of symmetry around a rotation portion of a work in which a convex portion and a concave portion are periodically and repeatedly provided.

5. Control unit 21A, 21B ... Holding table (work holding unit)
24 rotating mechanism (rotating part)
27 ... Alignment camera (imaging unit)
51: arithmetic processing unit (edge waveform deriving unit, center deviation calculating unit)
55 image processing unit 100 inspection device 241 motor (rotating unit)
271: Line sensor AR1: First imaging area AR2: Second imaging area AX3: Rotation axis (of motor 241) AX4: Symmetry axis d1, d2: Offset distance F, F11 Edge waveform I1: Work image W: Work Δ ... The amount of deviation

Claims (9)

A center misalignment detecting device that detects a misalignment of the work with respect to a rotating part, the work having an outer peripheral portion provided with a protrusion and a recess periodically and rotationally symmetrical about a symmetry axis. ,
An imaging unit configured to capture a workpiece partial image corresponding to an angle φ ( = 90 ° ) of an outer peripheral portion of the workpiece rotated by the rotating unit;
An edge waveform deriving unit that determines an edge waveform indicating the shape of the outer peripheral portion of the work included in the work partial image,
From the edge waveform derived by the edge waveform deriving unit, position information indicating the position of the convex portion or position information indicating the position of the concave portion is acquired, and the core of the workpiece with respect to the rotating unit is obtained based on the position information. A misalignment calculation unit for calculating misalignment,
A misalignment detection device comprising: The misalignment detection device according to claim 1,
The imaging unit has an imaging area for capturing the workpiece partial image while the workpiece is rotated by the same angle θ as the angle φ by the rotating unit,
The center deviation detecting device, wherein the edge waveform deriving unit obtains the edge waveform based on an image captured by the imaging region. A center misalignment detecting device that detects a misalignment of the work with respect to a rotating part, the work having an outer peripheral portion provided with a protrusion and a recess periodically and rotationally symmetrical about a symmetry axis. ,
An imaging unit configured to capture a work partial image corresponding to an angle φ (90 ° ≦ φ <360 °) of an outer peripheral portion of the work rotated by the rotating unit;
An edge waveform deriving unit that determines an edge waveform indicating the shape of the outer peripheral portion of the work included in the work partial image,
From the edge waveform derived by the edge waveform deriving unit, position information indicating the position of the convex portion or position information indicating the position of the concave portion is acquired, and the core of the workpiece with respect to the rotating unit is obtained based on the position information. A misalignment calculation unit for calculating misalignment,
With
The imaging unit,
A first imaging region and a second imaging region for imaging the outer peripheral portion of the work at different positions,
While the work is rotated by the rotation unit by an angle θ (where φ / 2 ≦ θ <φ), the first imaging region captures a first work local image corresponding to the angle θ and the second imaging. misalignment detecting apparatus characterized by obtaining the workpiece partial image by area imaging the second workpiece partial picture of the angle θ min. The misalignment detection device according to claim 3,
The misalignment detecting device in which the angle θ is φ / 2. The misalignment detection device according to claim 4,
The imaging unit is a line sensor in which the first imaging region and the second imaging region are provided linearly and are offset from the rotation axis of the rotation unit by a distance d in a radial direction,
d = R · sin (φ / 4) (Equation 1)
d = R · cos (φ / 4) (Expression 2)
Here, R is the distance from the symmetry axis of the work to the outer peripheral surface,
A misalignment detection device that satisfies one of the two. The misalignment detection device according to claim 5,
When the distance d obtained by the above equation 1 is shorter than the distance d obtained by the above equation 2, the line sensor is arranged such that the above equation 1 is satisfied, while the distance d obtained by the above equation 2 is satisfied. Is smaller than the distance d obtained by the expression 1, the misalignment detection device arranged so that the expression 2 is satisfied. A misalignment detection device according to any one of claims 1 to 6,
The misalignment detecting device calculates the misalignment of the work with respect to the rotating unit based on three or more pieces of continuous position information. A first step of holding, by a work holding unit, a work having an outer peripheral portion in which a convex portion and a concave portion are provided in a shape that is rotationally symmetric about a symmetric axis and that is periodically and repeatedly provided;
A second step of rotating the work around the rotation axis of the rotating unit by rotating the work holding unit holding the work by 90 ° by a rotating unit;
A third step of acquiring a workpiece partial image corresponding to an angle φ ( = 90 ° ) of an outer peripheral portion of the workpiece during rotation of the workpiece by the rotating unit;
A fourth step of obtaining an edge waveform indicating the shape of the outer peripheral portion of the work included in the work partial image;
A fifth step of calculating a center shift of the work with respect to the rotating section based on the edge waveform. A first step of holding, by a work holding unit, a work having an outer peripheral portion in which a convex portion and a concave portion are provided in a shape that is rotationally symmetric about a symmetry axis and that are periodically and repeatedly provided;
  A second step of rotating the work around the rotation axis of the rotating unit by rotating the work holding unit holding the work by an angle θ by a rotating unit;
  A third step of acquiring, by the imaging unit, a workpiece partial image corresponding to an angle φ of the outer peripheral portion of the workpiece during rotation of the workpiece by the rotating unit;
  A fourth step of obtaining an edge waveform indicating the shape of the outer peripheral portion of the work included in the work partial image;
  A fifth step of calculating a misalignment of the work with respect to the rotating section based on the edge waveform,
Said angle φ is greater than or equal to 90 ° and less than 360 °,
  The angle θ is not less than φ / 2 and less than φ,
  The imaging unit has a first imaging region and a second imaging region for imaging the outer peripheral portion of the work at different positions,
  In the third step, while the work is rotated by the rotation unit by the angle θ, the first imaging region captures a first work local image corresponding to the angle θ, and the second imaging region sets the angle θ. Obtains the work partial image by capturing the second work local image
A method for detecting misalignment.

Images