JP2022052199A - Component mounting system and component mounting method as well as component mounting device - Google Patents

Abstract

To provide a component mounting system and a component mounting method, as well as a component mounting device, capable of improving mounting accuracy of a component without reducing mounting efficiency of the component.SOLUTION: A component mounting method includes an imaging step (ST13), a positional deviation amount acquiring step (ST17), a learning data creating step (ST1), a learning step (ST2), a positional deviation amount estimating step (ST3), and a component mounting step (ST18 and ST14). The imaging step images a substrate by an imaging unit which integrally moves with a mounting head when a component is mounted on a mounting point of the substrate. The positional deviation amount acquiring step acquires a positional deviation amount from the mounting point of the mounted component. The learning data creating step creates learning data on the basis of images of one substrate and a positional deviation amount of a component mounted on the one substrate. The learning step generates a learning model on the basis of the learning data. The positional deviation amount estimating step estimates the positional deviation amount when the component is mounted on the substrate from the images of the substrate using the generated learning model. The component mounting step mounts the component on the mounting point of the substrate on the basis of the estimated positional deviation amount.SELECTED DRAWING: Figure 8

Description

The present invention relates to a component mounting system, a component mounting method, and a component mounting device for manufacturing a mounting board by mounting components on a board.

Conventionally, a component mounting system including a component mounting device for mounting a component on a board and an inspection device for inspecting a mounting state such as a displacement amount of the component on the mounting board on which the component is mounted is known (for example, Patent Document 1). ). In Patent Document 1, changes in images of a plurality of mounting boards imaged by an inspection device are monitored, and it is determined whether or not there is a correlation between the changes in the images and the quality deterioration of the mounting boards in which the changes in the images are observed. It is described that the data used in the component mounting device (mounting usage data) is corrected when the relationship is recognized.

Japanese Unexamined Patent Publication No. 2018-148003

However, in the prior art including Patent Document 1, although the information acquired in the inspection device can be fed back to the component mounting device to improve the component mounting accuracy, the feedback information is based on the past work. It is not possible to correct even the misalignment caused by the distortion inherent in the board to be worked on. In addition, it is possible to take an image of the mounting point of the component with a camera built in the component mounting device and correct the positional deviation peculiar to the board before mounting the component, but the work of moving the camera above the mounting point to take an image. Since time is added, there is a problem that the mounting efficiency is lowered.

Therefore, an object of the present invention is to provide a component mounting system, a component mounting method, and a component mounting device capable of improving the component mounting accuracy without lowering the component mounting efficiency.

The component mounting system of the present invention is a component mounting system for mounting components on a board, and has a mounting head for holding the components, a component mounting unit for mounting the components at a mounting point on the board, and the mounting head. An image pickup unit that images the upper surface of the board when the mounting head moves integrally and mounts a component at a mounting point of the board, a mounting control unit that controls the component mounting unit, and mounting on the board. An inspection unit that inspects the amount of misalignment of the component from the mounting point, an image of the upper surface of one substrate imaged by the imaging unit, and the component mounted on the one substrate inspected by the inspection unit. The image pickup unit uses the training data creation unit that creates training data based on the misalignment amount, the learning unit that generates a training model based on the training data, and the generated learning model. The mounting control unit includes a position shift amount estimation unit that estimates an estimated position shift amount when mounting a component on the board from an image of the upper surface of the board, and the mounting control unit determines the component based on the estimated estimated position shift amount. The mounting unit is controlled to mount components at the mounting points of the board.

The component mounting method of the present invention is a component mounting method for mounting a component on a board, and when the mounting head mounts the component at a mounting point on the board by an image pickup unit that moves integrally with the mounting head, the component mounting method is described. An imaging step of capturing an image of the upper surface of a substrate, a position deviation amount detecting step of detecting a displacement amount of the component mounted on the substrate from the mounting point, an image of the upper surface of one substrate imaged, and the above. A learning data creation process for creating training data based on the displacement amount of the component mounted on one board, a learning process for generating a learning model based on the training data, and the generated learning model. Based on the position shift amount estimation process that estimates the estimated position shift amount when mounting components on the board from the image of the upper surface of the board that has been imaged, and the estimated position shift amount, the board is mounted. It includes a component mounting process for mounting components at points.

The component mounting device of the present invention is a component mounting device for mounting components on a board, and has a mounting head for holding the components, a component mounting unit for mounting the components at a mounting point on the board, and the mounting head. An image pickup unit that images the upper surface of the board when the mounting head moves integrally and mounts a component at a mounting point of the board, a mounting control unit that controls the component mounting unit, and mounting on the board. The position shift amount acquisition unit that acquires the position shift amount from the mounting point of the component, the image of the upper surface of the one substrate imaged by the image pickup unit, and the one substrate acquired by the position shift amount acquisition unit. A learning data creation unit that creates training data based on the displacement amount of the component mounted on the above, a learning unit that generates a learning model based on the training data, and the generated learning model are used. The mounting control unit includes an estimated position deviation amount estimation unit that estimates an estimated position deviation amount when a component is mounted on the substrate from an image of the upper surface of the substrate captured by the image pickup unit, and the mounting control unit has an estimated estimated position. The component mounting unit is controlled based on the amount of deviation, and the component is mounted at the mounting point of the board.

According to the present invention, it is possible to improve the mounting accuracy of the component without lowering the mounting efficiency of the component.

Configuration explanatory view of the component mounting system of the embodiment of the present invention Configuration explanatory view of the component mounting apparatus of one embodiment of the present invention Functional explanatory view of the component mounting apparatus according to the embodiment of the present invention. A block diagram showing a configuration of a control system of a component mounting device according to an embodiment of the present invention. The figure which shows the example of the substrate which the component is mounted by the component mounting apparatus of one Embodiment of this invention. (A) (b) (c) Explanatory drawing of an example of the captured image captured by the component mounting apparatus according to the embodiment of the present invention. Explanatory drawing of mounting point height estimation method of one Embodiment of this invention Flow chart of the component mounting method according to the embodiment of the present invention

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. The configurations, shapes, and the like described below are examples for explanation, and can be appropriately changed according to the specifications of the component mounting system, the component mounting device, the mounting inspection device, and the post-reflow inspection device. In the following, the corresponding elements are designated by the same reference numerals in all the drawings, and duplicate description will be omitted. In FIG. 1 and a part described later, as biaxial directions orthogonal to each other in the horizontal plane, the X direction of the substrate transport direction (horizontal direction in FIG. 1) and the Y direction orthogonal to the substrate transport direction (vertical direction in FIG. 5). Is shown. In FIG. 1 and a part described later, the Z direction (vertical direction in FIG. 1) is shown as a height direction orthogonal to the horizontal plane. The Z direction is a vertical direction or an orthogonal direction when the component mounting device is installed on a horizontal plane.

First, the configuration of the component mounting system 1 will be described with reference to FIG. The component mounting system 1 has a function of mounting components on a board to produce a mounting board. In the present embodiment, one component mounting line 4 is connected to the management computer 3 via the communication network 2. The component mounting line 4 included in the component mounting system 1 is not limited to one, and may be two or more.

The work in the component mounting line 4 is managed by the management computer 3. The management computer 3 has a function of creating data and various parameters necessary for operating the production equipment included in the component mounting line 4 and transmitting them to each production equipment. Further, data such as image data and inspection results are transmitted from each production facility to the management computer 3.

In FIG. 1, the component mounting line 4 includes a board supply device M1, a solder printing device M2, a component mounting device M3, a component mounting device M4, a mounting inspection device M5, a reflow device M6, a post-reflow inspection device M7, and a board recovery device M8. It has a connected structure. The board supplied from the board supply device M1 is carried into the solder printing device M2. The solder printing apparatus M2 executes a solder printing operation of screen-printing cream solder for joining parts on a substrate. The board after solder printing is sequentially delivered to the component mounting devices M3 and M4. The component mounting devices M3 and M4 execute component mounting operations for mounting components on the board after solder printing.

The board after mounting the components is carried into the mounting inspection device M5. The mounting inspection device M5 includes a mounting inspection camera 5, and the mounting inspection camera 5 captures an image of a component mounted on a substrate, recognizes the imaging result, and determines the amount of misalignment of the mounted component from the mounting point. inspect. The inspection result (inspection result before reflow) is transmitted to the management computer 3 and the component mounting devices M3 and M4. The board after the inspection is carried into the reflow device M6. The reflow device M6 heats the substrate according to a predetermined heating profile, melts and solidifies the cream solder for joining the components, and solders the components to the substrate.

The reflowed substrate is carried into the post-reflow inspection device M7. The post-reflow inspection device M7 includes a post-reflow inspection camera 6, and the post-reflow inspection camera 6 takes an image of a component soldered to a substrate, recognizes the image pickup result, and processes the soldered component from a mounting point. Inspect the amount of misalignment. The inspection result (inspection result after reflow) is transmitted to the management computer 3 and the component mounting devices M3 and M4. The board after the inspection is recovered by the board recovery device M8. As described above, the mounting inspection device M5 is an inspection unit that inspects the parts mounted on the board before reflow. Further, the post-reflow inspection device M7 is an inspection unit that inspects the parts mounted on the board after reflow.

Next, the configurations of the component mounting devices M3 and M4 will be described with reference to FIG. The component mounting devices M3 and M4 include a substrate transfer mechanism 11 having a transfer conveyor 10 that supports both sides of the substrate B in the Y direction from below. The substrate transfer mechanism 11 drives the transfer conveyor 10 to transfer the substrate B to be worked in the X direction. A substrate holding portion 12 for holding the substrate B at the working position W is provided at an intermediate point of the substrate transport mechanism 11. The substrate holding portion 12 includes a substrate clamping mechanism 14 (see FIG. 4) including a lower receiving member 13 that supports the lower surface of the substrate B and a clamping member (not shown) that fixes both side portions of the substrate B. The substrate holding portion 12 fixes and holds the substrate B conveyed to the working position W by the substrate conveying mechanism 11 in the vertical direction (Z direction) and the horizontal direction (X direction, Y direction) by the substrate clamping mechanism 14.

A mounting head 15 is arranged above the board transfer mechanism 11. The mounting head 15 includes a nozzle 16 that holds the component P at the lower end by vacuum suction, and a nozzle elevating mechanism 17 that raises and lowers the nozzle 16 in the vertical direction. The mounting head 15 reciprocates between the component supply unit (not shown) by the mounting head moving mechanism 18 and the upper part of the board B held at the working position W by the board holding unit 12. Then, the mounting head 15 lowers the nozzle 16 holding the component P supplied by the component supply unit, and mounts the component P at the mounting point Kn (see FIG. 3) of the substrate B.

As described above, the mounting head 15 and the mounting head moving mechanism 18 have a mounting head 15 for holding the component P, and are a component mounting unit 19 for mounting the component P at the mounting point Kn of the board B. The mounting head 15 may be configured to include a plurality of nozzles 16 and a nozzle elevating mechanism 17. Further, the nozzle 16 may be configured to grip and hold the component P by the chuck.

In FIG. 3, a land E is formed around the mounting point Kn of the substrate B. The land E is an electrode bonded to the terminal of the component P. Further, a paste-like cream solder C is printed on the land E by the solder printing apparatus M2. The nozzle elevating mechanism 17 lowers the nozzle 16 to a height at which the lower surface of the component P held by the nozzle 16 crushes the upper surface of the cream solder C on the land E by a predetermined pushing amount (arrow a). After that, the vacuum suction by the nozzle 16 is stopped, and the nozzle elevating mechanism 17 raises the nozzle 16, so that the component P is mounted (transferred) on the substrate B.

In FIGS. 2 and 3, the mounting head 15 is equipped with a head camera 20 that captures a recognition mark (not shown) provided on the upper surface Ba of the substrate B held at the working position W by the substrate holding portion 12. There is. The head camera 20 moves the recognition mark to a position where it can be imaged by the mounting head moving mechanism 18. The position of the mounting point Kn is corrected based on the position of the recognition mark captured by the head camera 20. In this way, the head camera 20 is an imaging unit that moves integrally with the mounting head 15 and images the upper surface Ba of the substrate B.

Next, the configuration of the control system of the component mounting devices M3 and M4 will be described with reference to FIG. The component mounting devices M3 and M4 include a control device 30, a board transfer mechanism 11, a board holding section 12, a component mounting section 19, a head camera 20, and a touch panel 21. The control device 30 includes a storage unit 31, a mounting control unit 32, an image pickup processing unit 33, a position shift amount acquisition unit 34, a learning data creation unit 35, a learning unit 36, and a position shift amount estimation unit 37. In addition to various data, the touch panel 21 has a function of displaying an operation screen or the like on a liquid crystal panel or the like, and a function of inputting operation commands and data.

The storage unit 31 is a storage device, and stores mounting data 38, captured image data 39, position shift amount data 40, learning data 41, learning model 42, estimated position shift amount data 43, and the like. In the mounting data 38, various information such as the component type and size of the component P mounted on the substrate B and the coordinates of the mounting point Kn on the substrate B are stored for each substrate type of the mounting substrate to be produced.

In FIG. 4, the image pickup processing unit 33 lowers the nozzle 16 after the nozzle 16 holding the component P moves horizontally (XY directions) above the mounting point Kn to perform alignment during the component mounting operation. Before this, the head camera 20 is controlled to image the upper surface Ba of the substrate B. That is, the head camera 20 (imaging unit) images the upper surface Ba of the substrate B when the nozzle 16 mounts the component P at the mounting point Kn. Since the imaging time by the head camera 20 is shorter than the moving time of the mounting head 15 (nozzle 16), the imaging of the upper surface Ba of the substrate B by the head camera 20 after the nozzle 16 moves above the mounting point Kn is performed. Very little additional time.

FIG. 3 shows a state in which the nozzle 16 holding the component P5 is horizontally aligned and stopped above the mounting point K5. FIG. 5 shows the state of the upper surface Ba of the board B near the mounting point K5 immediately before mounting the component P5 on the board B. The image pickup processing unit 33 controls the head camera 20 in a state where the nozzle 16 is stopped above the mounting point K5 to image the upper surface Ba of the substrate B. In this example, the imaging range of the head camera 20 includes components P1 to P3 mounted on the substrate B. The image pickup processing unit 33 associates the captured image of the upper surface Ba of the substrate B with the information for specifying the substrate B, the information for specifying the mounting point Kn where the nozzle 16 is stopped, the image pickup date and time, and the like, and the image pickup image data 39. Is stored in the storage unit 31.

6 (a) to 6 (c) show an example of the captured image 20a captured by the head camera 20 in a state where the nozzle 16 holding the component P5 is stopped above the mounting point K5. The positions of the components P1 to P3 included in the captured image 20a corresponding to the mounting point K5 in the horizontal direction (X direction, Y direction) include the amount of horizontal displacement of the nozzle 16 with respect to the mounting point K5, distortion of the substrate B, and the like. Varies depending on. The relative positional relationship between the nozzle 16 or the component P5 held by the nozzle 16 and the mounting point K5 can be estimated from the positions of the components P1 to P3 in the captured image 20a corresponding to the mounting point K5.

In FIG. 4, the misalignment amount acquisition unit 34 acquires the pre-reflow inspection result from the mounting inspection device M5 and the post-reflow inspection result from the post-reflow inspection device M7, and stores them in the storage unit 31 as the misalignment amount data 40. .. The pre-reflow inspection result and the post-reflow inspection result include information for specifying the inspected substrate B, information for specifying the mounting point Kn of the inspected component P, and the amount of misalignment from the mounting point Kn of the component P (hereinafter, "inspection"). Information associated with "amount of misalignment") is included. In this way, the misalignment amount acquisition unit 34 acquires the inspection misalignment amount from the mounting point Kn of the component P mounted on the substrate B from the mounting inspection device M5 and the post-reflow inspection device M7 (inspection unit).

In FIG. 4, the learning data creation unit 35 includes an image pickup image 20a corresponding to a mounting point Kn on the upper surface Ba of the substrate B imaged by the head camera 20 (imaging unit) included in the image pickup image data 39, and a position shift amount acquisition unit. The training data 41 is created based on the inspection position deviation amount of the component P mounted on the mounting point Kn of the board B acquired by 34. That is, the learning data 41 is created for each mounting point Kn. The created learning data 41 is stored in the storage unit 31. The learning unit 36 uses the learning data 41 of the mounting point Kn as teacher data, and generates a learning model 42 of the mounting point Kn, which will be described later, by a learning algorithm using machine learning or the like.

As the learning algorithm, a neural network (including deep learning using a multi-layered neural network), genetic programming, a decision tree, a Bayesian network, a support vector machine (SVM), or the like can be used. The generated learning model 42 is stored in the storage unit 31.

The misalignment amount estimation unit 37 uses the generated learning model 42 of the mounting point Kn to capture the substrate from the image captured image 20a corresponding to the mounting point Kn of the upper surface Ba of the substrate B imaged by the head camera 20 (imaging unit). The estimated amount of misalignment when the component P is mounted on B is estimated (calculated). That is, the estimated misalignment amount is estimated to be detected by the inspection unit when the nozzle 16 is lowered and the component P is mounted on the substrate B without correcting the horizontal position of the nozzle 16. Is. The estimated estimated misalignment amount is stored in the storage unit 31 as the estimated misalignment amount data 43. The calculated estimated position deviation amount is used for the horizontal position correction of the nozzle 16 in the mounting operation of the component P by the mounting control unit 32.

Here, with reference to FIG. 7, a method of estimating the amount of misalignment in which the estimated amount of misalignment is estimated by the component mounting devices M3 and M4 will be described. First, the learning data creating unit 35 uses the captured image data 39 stored in the storage unit 31 and the misalignment amount data 40 to capture an image 20a (imaged substrate) corresponding to the same mounting point Kn on the same substrate B. The learning data 41 is created based on the inspection position deviation amount (the position deviation amount from the mounting point Kn of the component P mounted on the substrate B inspected by the inspection unit) (ST1: learning). Data creation process).

That is, the learning data creation unit 35 specifies the captured image 20a corresponding to the mounting point Kn, the inspection position deviation amount of the mounting point Kn, and the board B and the mounting point Kn for the plurality of (for example, 50 sheets) boards B. Teacher data associated with the information is created and stored as learning data 41. In the example of FIG. 7, teacher data in which the captured image 20a corresponding to the mounting point Kn of the substrate B1 and the inspection position deviation amount of the mounting point Kn are associated with each other are created as the learning data 41 of the mounting point Kn. Further, the teacher data in which the captured image 20a corresponding to the mounting point Kn of the substrate B2 and the inspection position deviation amount of the mounting point Kn are associated with each other are created as the learning data 41 of the mounting point Kn.

In FIG. 7, the learning unit 36 then generates a learning model 42 of the mounting point Kn based on the learning data 41 of the mounting point Kn (ST2: learning step). The created learning model 42 is stored in the storage unit 31. Next, the misalignment amount estimation unit 37 uses the generated learning model 42 of the mounting point Kn to mount the substrate B99 from the captured image 20a corresponding to the mounting point Kn of the substrate B99 imaged after the learning model 42 is created. The estimated misalignment amount when the component P is mounted at the point Kn is estimated (calculated) (ST3: misalignment amount estimation step). The estimated estimated misalignment amount is stored in the storage unit 31 as the estimated misalignment amount data 43.

That is, the misalignment amount estimation unit 37 uses the learning model 42 of the mounting point Kn to stop the nozzle 16 above the mounting point Kn of the board B99, and from the captured image 20a corresponding to the mounting point Kn imaged to the board B99. The estimated amount of misalignment that is estimated to occur when the component P is mounted at the mounting point Kn of the above is calculated. In other words, the learning model 42 uses a calculation formula for estimating (calculating) the estimated position shift amount of the mounting point Kn from the captured image 20a corresponding to the mounting point Kn imaged by stopping the nozzle 16 above the mounting point Kn. It is a calculation method.

In FIG. 4, the mounting control unit 32 controls the component mounting unit 19 based on the estimated estimated misalignment amount, and the horizontal position of the nozzle 16 that holds the component P so as to correct the estimated misalignment amount. Is corrected (positional deviation correction), and the component P is mounted at the mounting point Kn of the board B. Thereby, the mounting accuracy of the component P can be improved.

As described above, the component mounting devices M3 and M4 have a mounting head 15, and move integrally with the component mounting unit 19 for mounting the component P at the mounting point Kn of the board B and the mounting head 15 for mounting. When the head 15 mounts the component P at the mounting point Kn of the board B, the image pickup unit (head camera 20) that images the upper surface Ba of the board B, the mounting control unit 32 that controls the component mounting unit 19, and the board B. It is provided with a position deviation amount acquisition unit 34 for acquiring an inspection position deviation amount from the mounting point Kn of the component P mounted on the above.

Further, the component mounting devices M3 and M4 obtain learning data 41 based on the captured image 20a of one substrate B imaged by the imaging unit and the inspection position deviation amount of one substrate B acquired by the position displacement amount acquisition unit 34. Using the learning data creation unit 35 to be created, the learning unit 36 that generates the learning model 42 based on the learning data 41, and the generated learning model 42, the substrate B is imaged from the image captured image 20a of the substrate B. A position shift amount estimation unit 37 for estimating an estimated position shift amount when the component P is mounted on B is provided. Then, the mounting control unit 32 mounts the component P at the mounting point Kn of the substrate B based on the estimated estimated position deviation amount. As a result, the mounting accuracy of the component P can be improved without lowering the mounting efficiency of the component P.

The component mounting system 1 includes a position shift amount acquisition unit 34, a learning data creation unit 35, a learning unit 36, and an image data acquisition unit in which the management computer 3 acquires image data 39 from the component mounting devices M3 and M4. It may be a configuration. In this case, in the management computer 3, the inspection result acquired from the mounting inspection device M5 or the post-reflow inspection device M7, which is the inspection unit, and the image captured by the imaging unit (head camera 20) acquired from the component mounting devices M3 and M4. A learning model 42 is generated based on 20a. The generated learning model 42 is transmitted to the component mounting devices M3 and M4.

Then, in the component mounting devices M3 and M4, the misalignment amount estimation unit 37 estimates the estimated misalignment amount using the learning model 42 received from the captured image 20a captured by the head camera 20 (imaging unit) and mounts the product. The control unit 32 controls the component mounting unit 19 based on the estimated position deviation amount, and mounts the component P at the mounting point Kn of the substrate B so as to correct the estimated position deviation amount.

Next, a component mounting method for mounting the component P on the substrate B by the component mounting devices M3 and M4 will be described according to the flow of FIG. First, the mounting control unit 32 controls the board transfer mechanism 11 to carry in the board B (ST10: board carry-in step). Next, the mounting control unit 32 controls the board holding unit 12 to hold the board B at the working position W (ST11: board holding step).

Next, the mounting control unit 32 controls the component mounting unit 19 to move the mounting head 15 in the horizontal direction, and horizontally moves the nozzle 16 holding the component P above the mounting point Kn (ST12: nozzle horizontal movement step). ). At this time, the head camera 20 (imaging unit) moves integrally with the mounting head 15 and moves to a position corresponding to the mounting point Kn. Next, the head camera 20 images a position corresponding to the mounting point Kn on the upper surface Ba of the substrate B (ST13: imaging step). That is, when the nozzle 16 mounts the component P at the mounting point Kn of the board B by the head camera 20 that moves integrally with the mounting head 15, the upper surface Ba of the board B is imaged. The captured image 20a corresponding to the captured mounting point Kn is stored in the storage unit 31 as captured image data 39.

In FIG. 8, the mounting control unit 32 then controls the component mounting unit 19 to lower the nozzle 16 to mount the component P at the mounting point Kn of the substrate B (ST14: nozzle lowering step). Hereinafter, the nozzle horizontal movement step (ST12), the imaging step (ST13), and the nozzle lowering step (ST14) to the next mounting point Kn are repeatedly executed. When the mounting of the component P on the board B held by the board holding unit 12 is completed, the mounting control unit 32 controls the board transfer mechanism 11 to carry out the board B (ST15: board carrying out step).

The board B carried out from the component mounting devices M3 and M4 is carried into the mounting inspection device M5 (inspection unit). The mounting inspection device M5 detects the amount of misalignment from the mounting point Kn of the component P mounted on the substrate B (position misalignment detection step before reflow). The substrate B after inspection is carried into the reflow device M6, and the substrate B after reflow is carried into the inspection device M7 (inspection unit) after reflow. The post-reflow inspection device M7 detects the amount of misalignment from the mounting point Kn of the component P mounted on the substrate B (post-reflow misalignment detection step). The substrate B after the inspection is recovered by the substrate recovery device M8.

In FIG. 8, when the number of substrates B on which the captured image 20a corresponding to the mounting point Kn is captured is less than a predetermined number in the imaging step (ST13) (No in ST16), the substrate loading step shown on the left side in FIG. 8 (No). Returning to ST10), the component mounting work for the next board B is executed. When the captured images 20a are captured by a predetermined number of substrates B (Yes in ST16), the misalignment amount acquisition unit 34 acquires the misalignment amount from the mounting inspection device M5 or the post-reflow inspection device M7, and the misalignment amount acquisition unit 34 obtains the misalignment amount. It is stored in the storage unit 31 as the amount data 40 (ST17: position shift amount acquisition step).

Next, the learning data creation step (ST1) is executed, and the learning data 41 is generated based on the captured image 20a of the upper surface Ba of the one substrate B and the amount of misalignment of the component P mounted on the one substrate B. Will be created. Next, the learning step (ST2) is executed, and the learning model 42 is generated based on the learning data 41. Next, the process proceeds to the substrate loading step (ST10) shown on the right side in FIG. 8, and the substrate B to be next is loaded. Next, a substrate holding step (ST11), a nozzle horizontal movement step (ST12), and an imaging step (ST13) are executed. As a result, the nozzle 16 that holds the component P moves above the mounting point Kn, and the captured image 20a corresponding to the mounting point Kn of the board B held by the board holding portion 12 is acquired.

In FIG. 8, the misalignment amount estimation step (ST3) is then executed. That is, using the generated learning model 42, the estimated position deviation amount when the component P is mounted from the captured image 20a corresponding to the mounting point Kn of the upper surface Ba of the imaged substrate B to the mounting point Kn of the substrate B. Is estimated. Next, the mounting control unit 32 controls the component mounting unit 19 based on the estimated estimated misalignment amount to correct the horizontal misalignment of the nozzle 16 holding the component P (ST18: misalignment correction step). Next, the nozzle lowering step (ST14) is executed, and the component P held by the nozzle 16 is mounted at the mounting point Kn of the substrate B.

That is, the misalignment correction step (ST18) and the nozzle descent step (ST14) are component mounting steps for mounting the component P at the mounting point Kn of the substrate B based on the estimated estimated misalignment amount. In this way, by estimating and correcting the amount of misalignment based on the image captured image 20a corresponding to the mounting point Kn imaged before mounting the component P at the mounting point Kn, the mounting efficiency of the component P is lowered. The mounting accuracy of the component P can be improved without any problem.

In FIG. 8, hereinafter, a nozzle horizontal movement step (ST12), an imaging step (ST13), a misalignment amount estimation step (ST3), a misalignment correction step (ST18), and a nozzle lowering step (ST14) to the next mounting point Kn are performed. However, it is executed repeatedly. When the mounting of the component P on the substrate B held by the substrate holding portion 12 is completed, the substrate unloading step (ST15) is executed. Then, until the production of a predetermined number of mounting boards is completed (No in ST19), the process returns to the board loading step (ST10) shown on the right side in FIG. 8, and the component mounting work on the next board B is executed.

In the above, the example of correcting the horizontal (X direction, Y direction) position deviation of the nozzle 16 in the position deviation correction step (ST18) has been described, but the corrected position deviation is limited to the horizontal direction. Instead, the θ direction (direction of rotation with the axis in the Z direction as the rotation axis) or the Z direction may also be corrected. That is, the inspection unit (mounting inspection device M5, post-reflow inspection device M7) also detects the amount of misalignment in the θ direction (Z direction), and the misalignment estimation unit 37 corresponds to the mounting point Kn using the learning model 42. The estimated displacement amount in the θ direction (Z direction) at the mounting point Kn is also estimated from the captured image 20a. Then, in the position shift correction step (ST18), the position shift in the θ direction (Z direction) of the nozzle 16 holding the component P may be corrected.

The component mounting system, component mounting method, and component mounting device of the present invention have the effect of improving the component mounting accuracy without lowering the component mounting efficiency, and are useful in the field of mounting components on a substrate. Is.

1 Component mounting system 15 Mounting head 19 Component mounting section 20 Head camera (imaging section)
20a Captured image B Board Ba Top surface K5, Kn Mounting point M3, M4 Component mounting device M5 Mounting inspection device (inspection unit)
M7 Post-reflow inspection equipment (inspection department)
P1 to P5, P parts

Claims (5)

It is a component mounting system that mounts components on a board.
A component mounting unit that has a mounting head for holding components and mounts the components at the mounting points of the board.
An image pickup unit that moves integrally with the mounting head and images the upper surface of the board when the mounting head mounts a component at a mounting point of the board.
A mounting control unit that controls the component mounting unit,
An inspection unit that inspects the amount of misalignment from the mounting point of the component mounted on the board, and
A learning data creation unit that creates learning data based on an image of the upper surface of one substrate imaged by the imaging unit and the amount of misalignment of the component mounted on the one substrate inspected by the inspection unit.
A learning unit that generates a learning model based on the learning data,
Using the generated learning model, the image pickup unit is provided with a position shift amount estimation unit that estimates an estimated position shift amount when mounting a component on the substrate from an image of the upper surface of the substrate captured by the image pickup unit.
The mounting control unit is a component mounting system that controls the component mounting unit based on the estimated estimated position deviation amount to mount the component at a mounting point of the board. The component mounting system according to claim 1, wherein the inspection unit inspects the component mounted on the substrate before reflow. The component mounting system according to claim 1, wherein the inspection unit inspects the component mounted on the substrate after reflow. It is a component mounting method for mounting components on a board.
An imaging process in which the upper surface of the substrate is imaged when the mounting head mounts a component at a mounting point of the substrate by an imaging unit that moves integrally with the mounting head.
A misalignment detection step for detecting the misalignment of the component mounted on the substrate from the mounting point, and a misalignment detection step.
A learning data creation process for creating learning data based on an image of the upper surface of one substrate imaged and the amount of misalignment of the component mounted on the one substrate.
A learning process that generates a learning model based on the learning data,
Using the generated learning model, a position shift amount estimation process for estimating an estimated position shift amount when mounting a component on the board from an image of the upper surface of the board captured, and a position shift amount estimation step.
A component mounting method including a component mounting step of mounting a component at a mounting point of the board based on the estimated estimated misalignment amount. A component mounting device that mounts components on a board.
A component mounting unit that has a mounting head for holding components and mounts the components at the mounting points of the board.
An image pickup unit that moves integrally with the mounting head and images the upper surface of the board when the mounting head mounts a component at a mounting point of the board.
A mounting control unit that controls the component mounting unit,
A position shift amount acquisition unit that acquires a position shift amount from the mounting point of the component mounted on the board, and a position shift amount acquisition unit.
Learning data creation that creates learning data based on the image of the upper surface of one substrate captured by the imaging unit and the displacement amount of the component mounted on the one substrate acquired by the displacement amount acquisition unit. Department and
A learning unit that generates a learning model based on the learning data,
Using the generated learning model, the image pickup unit is provided with a position shift amount estimation unit that estimates an estimated position shift amount when mounting a component on the substrate from an image of the upper surface of the substrate captured by the image pickup unit.
The mounting control unit is a component mounting device that controls the component mounting unit based on the estimated estimated position deviation amount to mount the component at a mounting point of the board.

Images