Substrate work device

Abstract

This substrate work device comprises: a work unit that performs predetermined work on a substrate; a first camera that captures an image of the substrate and acquires an inspection image before the work is performed; a storage unit that stores a master image of the substrate; and an abnormality detection unit that performs abnormality recognition on the substrate by detecting difference between the master image and the inspection image. Before detecting the difference, the abnormality detection unit derives a first feature element for the master image and a second feature element for the inspection image on the basis of a common position of interest in the master image and the inspection image and performs image correction on at least one of the master image and the inspection image so that the first feature element and the second feature element overlap.

Description

Substrate working device

[0001] The present invention relates to a substrate working device provided with a working unit that performs predetermined work on a substrate on which components are mounted.

[0002] For example, in a substrate mounting device that mounts components on a substrate, before mounting a specific component, an inspection may be performed to check whether there is a foreign object at the mounting position of the substrate. Patent Document 1 discloses an inspection device that determines the presence or absence of a foreign object on an inspection target substrate based on the difference between a master image of a good substrate and an inspection image of the substrate to be inspected. The difference between the two images corresponds to the change from the master image to the inspection image. If a foreign object that is not present in the master image, such as a dropped component, is reflected in the inspection image, the presence or absence of the dropped component becomes the change in the image.

[0003] However, change elements other than the presence of a foreign object may be detected as the difference in the image. For example, the fixed state of the substrate may vary between the acquisition of the master image and the acquisition of the inspection image due to the impact during component mounting or the operations of substrate conveyance and loading / unloading. In this case, image misalignment may occur between the two images due to the posture deviation of the substrate, and this may be detected as the difference. Also, depending on the production line settings, the master image and the inspection image may be acquired by different cameras. In this case, image misalignment may occur between the two images due to the individual differences of the cameras, and this may be detected as the difference.

[0004] Japanese Unexamined Patent Application Publication No. 2012 - 112669

[0005] An object of the present invention is to provide a substrate working device that can accurately recognize abnormalities in a substrate when performing predetermined work on the substrate.

[0006] A substrate work apparatus according to one aspect of the present invention comprises: a work unit that performs predetermined work on a substrate on which components are mounted; a first camera that captures an image of the substrate and acquires an inspection image before the work unit performs the work on the substrate; a storage unit that stores a master image showing the initial state of the substrate; an anomaly detection unit that detects the difference between the master image and the inspection image and recognizes an anomaly in the substrate on which the work unit is performing the work; and a control unit that controls the operation of the work unit based on the result of the anomaly recognition. The anomaly detection unit, before detecting the difference, derives a first feature element for the master image and a second feature element for the inspection image based on a common point of interest in the master image and the inspection image, and corrects at least one of the master image or the inspection image so that the first feature element and the second feature element overlap.

[0007] Figure 1 is a simplified block diagram showing the configuration of a mounted circuit board production line incorporating a mounter as an example of a circuit board work apparatus according to the present invention. Figure 2 is a plan view showing the schematic configuration of the mounter. Figure 3 is a side view showing the schematic configuration of the head unit portion of the mounter. Figure 4 is a block diagram showing the electrical configuration of the mounter. Figure 5 is an explanatory diagram of foreign object inspection using a difference image between a master image and an inspection image. Figure 6 is a schematic diagram for explaining inspection examples 1 to 3 of foreign object inspection. Figures 7(A) and (B) show a first embodiment of image correction that corrects the image misalignment between the master image and the inspection image. Figure 8 shows a modified example of the first embodiment. Figure 9 is a flowchart showing the component mounting process on a circuit board, including image correction according to the first embodiment. Figure 10 shows a position of interest in image correction according to the second embodiment. Figure 11 shows a second embodiment of image correction based on the difference between the master image and the inspection image. Figure 12 shows a modified example of the second embodiment. Figure 13 is a flowchart showing the component mounting process on a circuit board, including image correction according to the second embodiment. Figure 14 is a block diagram showing an embodiment of brightness correction.

[0008] Embodiments of the present invention will be described in detail below with reference to the drawings. In the embodiments described below, the substrate handling apparatus according to the present invention is shown as an example in which components are mounted on a printed circuit board. That is, the predetermined operation on the board is a component mounting operation, and the work unit is shown as a mounting head that picks up and mounts the components. The components are, for example, chip components such as chip resistors and chip capacitors, IC package type electronic components such as BGA, QFP or SOP, power supply related components such as capacitors and transformers, connectors, heat sinks, etc. Note that the substrate handling apparatus of the present invention is not limited to a mounter, but can be applied to various devices equipped with a work unit that performs predetermined operations on a board. For example, the work unit may be a print head that performs screen printing on the board, or a coating head that applies adhesives or chemicals to the board.

[0009] [Printed Circuit Board Production Line] Figure 1 is a simplified block diagram showing the configuration of a printed circuit board production line 1, which incorporates a mounter as an example of a circuit board handling apparatus according to the present invention. In Figure 1, each component of the printed circuit board production line 1 is shown as a simplified block. The printed circuit board production line 1 includes a printer 11, a print inspection machine 12, a mounter 13, a circuit board inspection machine 14, a reflow oven 15, and a visual inspection machine 16, which are arranged in tandem along a conveyor 10 that transports printed circuit boards in the direction indicated by the arrows in the figure. At the upstream end of the printed circuit board production line 1, a loader 17 is located to transport circuit boards to the printer 11, and at the downstream end, an unloader 18 is located to remove the produced circuit boards from the visual inspection machine 16.

[0010] The printing press 11 applies solder to the pad portion of the printed circuit board. For example, the printing press 11 places a mask with openings for solder application onto the printed circuit board and applies solder paste from above the mask. The printing inspection machine 12 takes an image of the soldered printed circuit board to determine the two-dimensional and three-dimensional shape of the solder and inspects whether the position, amount, and height of the solder applied to the printed circuit board are appropriate.

[0011] The mounter 13 is equipped with a head for mounting components and mounts the required electronic components and other components onto the printed circuit board. In the example in Figure 1, a triple arrangement of mounters 13 is shown. That is, the first mounter 13A, the second mounter 13B, and the third mounter 13C are arranged in tandem from the upstream side. The triple mounters 13A, 13B, and 13C can mount different components or the same components on a single printed circuit board, or each can mount components on separate printed circuit boards. In the following embodiments, examples are shown in which the circuit board processing apparatus according to the present invention is one or more of the triple mounters 13A, 13B, and 13C. Note that the circuit board processing apparatus may also be the printing press 11 described above.

[0012] The circuit board inspection machine 14 images the printed circuit board after it has passed through the mounter 13 and inspects for misalignment of electronic components mounted on the printed circuit board, component lifting, mounting omissions, soldering defects, etc. The reflow oven 15 heats the printed circuit board with the mounted electronic components to melt the solder and fix the electronic components to the printed circuit board. The visual inspection machine 16 images the printed circuit board after the heating treatment in the reflow oven 15 and inspects for misalignment of electronic components, component lifting, mounting omissions, soldering defects, etc., in the same way as the circuit board inspection machine 14.

[0013] [Mounter Structure] Figure 2 is a plan view showing the schematic configuration of the mounter 13, and Figure 3 is a side view showing the schematic configuration of the head unit 4 portion of the mounter 13. In Figures 2 and 3, X, Y, and Z directions are indicated. The X direction is the transport direction of the substrate P. Note that the X direction may also be referred to as the left-right direction, the Y direction as the front-back direction, and the Z direction as the up-down direction.

[0014] The mounter 13 includes a device base 130, a substrate transport unit 2, a component supply unit 3, a head unit 4, a component recognition camera 7, and a control device 8, all of which are arranged on the device base 130. The head unit 4 is equipped with a substrate recognition camera 5, which can be used as a first or second camera, and a scan camera unit 6.

[0015] The device base 130 has a rectangular shape in plan view and is a flat top surface, to which the substrate transport unit 2 and component supply unit 3 are assembled. The substrate transport unit 2 transports the substrate P on which components are mounted. The substrate transport unit 2 has a pair of conveyors 21 and 22 on the device base 130 that transport the substrate P in the left-right direction. The conveyors 21 and 22 bring the substrate P into the mounter 13 from the right side and transport it to the left to a predetermined working position, in this case the position of the substrate P shown in Figure 2, where it is temporarily stopped. At this working position, components are mounted on the substrate P. After the mounting work, the conveyors 21 and 22 transport the substrate P to the left and transport it out of the mounter 13. The conveyors 21 and 22 constitute a part of the transport conveyor 10 shown in Figure 1.

[0016] The component supply unit 3 supplies components to be mounted on the circuit board P. The component supply units 3 are located on both the front and rear sides of the circuit board transport unit 2. Each component supply unit 3 is equipped with a plurality of tape feeders 31 arranged in the left-right direction. Each tape feeder 31 holds a reel around which tape containing and holding electronic components such as ICs, transistors, resistors, and capacitors at predetermined intervals is wound. The tape feeder 31 intermittently feeds the tape from the reel, supplying components to the component supply position at the tip of the feeder. Figure 2 shows an example of a tape feeder 31, but trays holding large electronic components or other components may also be attached to the component supply unit 3.

[0017] The head unit 4 takes components from the component supply unit 3 and mounts them onto the circuit board P. In other words, the head unit 4 is a work unit that performs the task of mounting components onto the circuit board P as a predetermined task on the circuit board P. The head unit 4 is positioned above the device base 130 so as to be movable in the X and Y directions. The head unit 4 takes components from the tape feeder 31 at the component supply position and mounts the components to predetermined positions on the circuit board P at the work position. A support beam 23 extending in the X direction is erected above the device base 130. The head unit 4 is movably supported on an X-axis fixed rail 24 fixed to the support beam 23.

[0018] The support beam 23 is supported by a Y-axis fixed rail 25, the ends of which extend in the Y direction, and is movable in the Y direction along this Y-axis fixed rail 25. An X-axis servo motor 26 and a ball screw shaft 27 are positioned relative to the X-axis fixed rail 24. A Y-axis servo motor 28 and a ball screw shaft 29 are positioned relative to the Y-axis fixed rail 25. The head unit 4 moves in the X direction by the rotational drive of the ball screw shaft 27 by the X-axis servo motor 26, and moves in the Y direction by the rotational drive of the ball screw shaft 29 by the Y-axis servo motor 28.

[0019] The head unit 4 is equipped with multiple heads 4H for holding and transporting parts. In this embodiment, an example is shown in which a total of six heads 4H are arranged in a line in the X direction. Each head 4H includes a shaft 41 extending in the Z direction and a suction nozzle 42 attached to the lower end of the shaft 41. The shaft 41 can move up and down relative to the head unit 4 and rotate around the nozzle central axis (R axis). The suction nozzle 42 picks up and holds the parts and mounts them on the surface of the substrate P.

[0020] The circuit board recognition cameras 5 are fixedly mounted on both the left and right sides of the head unit 4. The circuit board recognition cameras 5 capture images of various marks attached to the surface of the circuit board P that has been transported to the work position of the mounter 13 by the conveyors 21 and 22. In Figure 2, as an example of the marks, a pair of fiducial marks Fid attached to the diagonals of a rectangular circuit board P are shown. The fiducial marks Fid are marks used to detect the amount of positional deviation of the transported circuit board P from the origin coordinates of the work position. The position of the fiducial marks Fid is identified in the image data obtained by the circuit board recognition cameras 5, and the amount of positional deviation from the origin coordinates is determined. This amount of positional deviation is referenced during component mounting, and components are mounted on the circuit board P by the head 4H so as not to cause positional deviation. Note that either the left or right circuit board recognition camera 5 may be omitted.

[0021] In this embodiment, the substrate recognition camera 5 is also used to acquire a master image and an inspection image for abnormality recognition of the substrate P. The master image is an image showing the initial state of the substrate P. In this embodiment, the master image is an image taken when the substrate P is transported to the working position of the head unit 4 by the substrate transport unit 2. The inspection image is an image of the work area of ​​the substrate P taken before the head unit 4, acting as the work unit, performs a specific operation on the substrate P. In this embodiment, the inspection image is an image of the mounting area of ​​a specific component taken by the head unit 4 immediately before mounting the specific component onto the substrate P. Specific examples of abnormality recognition of the substrate P using the master image and the inspection image will be described in detail later.

[0022] The scan camera unit 6 is mounted near the lower end of the head unit 4 so as to be movable in the X direction relative to the head unit 4. The scan camera unit 6 is a unit for image recognition of the state in which a part is held by the suction nozzle 42. The scan camera unit 6 includes a scan camera 62 that images the part held by the suction nozzle 42. The scan camera unit 6 moves in the X direction along a ball screw shaft 61 that extends in the X direction and is attached to the head unit 4.

[0023] The part recognition camera 7 is integrated into the device base 130. The part recognition camera 7 is a camera whose imaging field of view is above the device base 130, and it images the part held by the suction nozzle 42 from the bottom side for part recognition. Based on the captured image, the suction state of the part by the suction nozzle 42 is inspected.

[0024] [Electrical Configuration of the Mounter] Next, the control configuration of the mounter 13 will be described. Figure 4 is a block diagram showing the electrical configuration of the mounter 13. The mounter 13 is equipped with a control device 8 located within the device base 130. The control device 8 controls the operation of each part of the mounter 13. Figure 4 shows the Z-axis servo motor 43 and the R-axis servo motor 44, which were omitted from Figures 2 and 3. The Z-axis servo motor 43 and the R-axis servo motor 44 are motors incorporated into the head unit 4. The Z-axis servo motor 43 is a drive source that raises and lowers the head 4H along the Z-axis when the head 4H picks up or mounts a component. The R-axis servo motor 44 is a drive source that rotates the head 4H around the R-axis.

[0025] The control device 8 consists of a processor and the like that operates by loading a predetermined program, and functionally comprises an imaging control unit 81, an image processing unit 82, an axis control unit 83, an anomaly detection unit 84, a main control unit 85, and a storage unit 86. An input unit 87, consisting of an operation panel and a keyboard, is connected to the control device 8. The input unit 87 accepts input of various information such as setting information and control instructions for the mounter 13.

[0026] The imaging control unit 81 controls the imaging operations of the substrate recognition camera 5, the scan camera 62, and the component recognition camera 7, as well as various other cameras equipped on the mounter 13. For example, the imaging control unit 81 provides control signals to these cameras to specify the timing for performing imaging operations and control signals to specify the exposure time. In this embodiment, the imaging control unit 81 causes the substrate recognition camera 5 to perform the imaging of the master image and the inspection image described above.

[0027] The image processing unit 82 applies image processing techniques such as edge detection and pattern recognition with feature extraction to image data acquired by the substrate recognition camera 5 and other devices to extract various information from the image. Specifically, the image processing unit 82 performs processing to identify the coordinates of points of interest, such as the fiducial mark Fid and solder pattern SP on the substrate P, based on the image data acquired by the substrate recognition camera 5. In addition, the image processing unit 82 performs processing to identify the shape, position, etc., of the component held by the suction nozzle 42, based on the image data acquired by the scan camera 62 and the component recognition camera 7.

[0028] The axis control unit 83 controls the movement of the head unit 4 in the XY direction by controlling the X-axis servo motor 26 and the Y-axis servo motor 28. The axis control unit 83 also controls the lifting, lowering, and rotational movements of the head 4H by controlling the Z-axis servo motor 43 and the R-axis servo motor 44 of the head unit 4. Furthermore, the axis control unit 83 controls the loading, stopping, and unloading of the substrate P to the mounter 13 by controlling the conveyors 21 and 22.

[0029] The anomaly detection unit 84 recognizes an anomaly in the circuit board P on which the head unit 4 performs mounting work by detecting the difference between the master image and the inspection image. In addition, before detecting the difference, the anomaly detection unit 84 performs image correction to correct any misalignment between the master image and the inspection image and to adjust the scale. Functionally, the anomaly detection unit 84 includes an image correction unit 841, a brightness calibration unit 842, an image comparison unit 843, and a determination unit 844.

[0030] Before detecting the difference, the image correction unit 841 performs a process to derive a first feature element for the master image and a second feature element for the inspection image based on a common point of interest in both the master image and the inspection image. Subsequently, the image correction unit 841 corrects at least one of the master image or the inspection image so that the first feature element and the second feature element overlap.

[0031] The location of interest is a feature portion on the surface of the substrate P that can be imaged by the substrate recognition camera 5 and is identifiable by shape or color. In the first embodiment described below, an example is shown where the location of interest is a fiducial mark Fid (a mark for substrate recognition) attached to the substrate P. In the second embodiment, an example is shown where the location of interest is a specific solder pattern SP applied to the substrate P. The first and second feature elements are substantial or virtual feature elements that can be derived based on one or more of the locations of interest. In the first embodiment, an example is shown where the feature element is a straight line drawn between two fiducial marks Fid that are spaced apart from each other. In the second embodiment, an example is shown where the feature element is a straight line drawn between two solder patterns SP that are spaced apart from each other.

[0032] The luminance calibration unit 842 performs luminance calibration to correct luminance discrepancies between cameras when the master image and the inspection image are captured by different cameras. When the master image and the inspection image are captured by the same camera, the luminance at the same position in both images will be the same if the lighting conditions are the same. However, when the two images are captured by different cameras, the luminance may differ due to individual differences in the cameras. The luminance calibration unit 842 performs a process to calibrate the luminance of both images so that the luminance difference due to individual differences in the cameras is not detected as a difference between the master image and the inspection image. The luminance calibration will be described in detail based on the example shown in Figure 14.

[0033] The image comparison unit 843 compares the master image and the inspection image and derives a difference image of the two images. The difference image is obtained by determining the brightness difference at the same position of pixels in the master image and the inspection image, and then imaging the portion where the brightness difference exists. For example, if some foreign object that is not present in the master image is captured in the inspection image, a brightness difference will be detected between the master image and the inspection image at the location where the foreign object is captured. Therefore, the foreign object can be detected based on the difference image derived by the image comparison unit 843.

[0034] The determination unit 844 evaluates the difference image derived by the image comparison unit 843 and determines whether or not a foreign object is visible in the inspection image. For example, the determination unit 844 calculates the area (luminance difference area) of the portion that appears in the difference image as a difference in luminance between the master image and the inspection image. Furthermore, the determination unit 844 checks whether or not the luminance difference area exceeds a predetermined threshold, and if it exceeds the threshold, it determines that a foreign object is present at the mounting position of the substrate P.

[0035] The main control unit 85 comprehensively controls various operations on the mounter 13. For example, the main control unit 85 provides control signals to the imaging control unit 81, the image processing unit 82, and the axis control unit 83, etc., to perform operations such as capturing images, performing image processing on image data, and driving the head unit 4 and the head 4H.

[0036] Furthermore, the main control unit 85 includes an inspection control unit 851 that causes the mounter 13 to perform operations for inspecting for foreign objects on the substrate P. The inspection control unit 851 provides control signals to the imaging control unit 81 and the axis control unit 83, causing the substrate recognition camera 5 to capture a master image and an inspection image. The inspection control unit 851 also causes the abnormality detection unit 84 to perform predetermined processing for foreign object inspection. Furthermore, the inspection control unit 851 controls the operation of the head unit 4 based on the abnormality recognition result by the abnormality detection unit 84. Specifically, if the determination unit 844 determines that a foreign object is present, the inspection control unit 851 stops the component mounting work by the head unit 4 and issues an error alert.

[0037] The memory unit 86 stores various information about the substrate P and electronic components, various setting values ​​and parameters related to the mounter 13, control data, operation programs, etc. The memory unit 86 also stores various information about the substrate P and the components to be mounted. In this embodiment, identification information of components that should be inspected for foreign objects immediately before mounting is also stored in the memory unit 86 in advance. Furthermore, the memory unit 86 also stores master image data for foreign object inspection and information about the above-mentioned points of interest.

[0038] [Foreign Object Inspection and Problems] A specific example of foreign object inspection using a difference image between a master image and an inspection image will be explained with reference to Figure 5. In the process of mounting electronic components onto a substrate P, an inspection is sometimes performed to check whether foreign objects are present at the mounting location of a specific component, such as a BGA (Ball Grid Array), before mounting that component. Figure 5 shows an example of foreign object inspection using a difference image DI between a master image MI and an inspection image EI for the mounting area PA of a BGA.

[0039] A BGA is a thin, rectangular component that packages an IC chip and has a mounting surface with a grid-like arrangement of solder balls. On the mounting area PA of the substrate P, solder patterns SP are printed in a grid arrangement, which are joined to each of the solder ball groups. The solder patterns SP have a circular shape when viewed from above. If foreign matter is present in the mounting area PA, the mounting surface will lift away from the mounting area PA. As a result, some or all of the solder balls will not be able to contact the solder patterns SP, resulting in a mounting defect. Therefore, inspection for foreign matter is required before mounting.

[0040] The master image MI is captured by the substrate recognition camera 5 when the substrate P is transported by conveyors 21 and 22 to the work position of the mounter 13 and fixed in place. The acquired master image MI data is stored in the storage unit 86. The master image MI is an image of a normal mounting area PA that is free of foreign matter. That is, only the grid-arranged solder pattern SP is visible in the mounting area PA in the master image MI.

[0041] The inspection image EI is captured by the substrate recognition camera 5 immediately before the head unit 4 mounts the BGA to the mounting area PA. Between the time the substrate P is brought in and immediately before the BGA is mounted, foreign matter may enter the mounting area PA for some reason. This foreign matter may be other electronic components or dust and debris from the factory. The inspection image EI shown in Figure 5 illustrates a mounting area PA with multiple chip components C scattered around. Note that the inspection image EI does not necessarily have to be captured immediately before the mounting of a specific component such as the BGA; it can be captured at a time close to the start of mounting. Also, other work may be performed between the capture of the inspection image EI and the mounting of the specific component.

[0042] The difference image DI is an image that makes the brightness difference between the master image MI and the inspection image EI apparent. If the brightness difference between the master image MI and the inspection image EI is zero across the entire image, nothing will appear in the difference image DI. In the example in Figure 5, the inspection image EI contains a chip component C that is not present in the master image MI. In the inspection image EI, the brightness of the pixels where the chip component C is reflected differs from the brightness of the corresponding pixels in the master image MI. Therefore, a brightness difference region corresponding to the shape of the chip component C appears in the difference image DI. The determination unit 844 determines whether or not the BGA can be mounted in the current mounting area PA based on the ratio of the total area of ​​the brightness difference region to the total area of ​​the difference image DI.

[0043] Figure 6 is a schematic diagram illustrating inspection examples 1 to 3 for foreign object inspection. Inspection examples 1 to 3 are variations in the acquisition method of the master image MI and inspection image EI. Inspection example 1 is an example in which the master image MI and inspection image EI are captured by one substrate recognition camera 5 of one mounter 13. That is, when the substrate P is brought in, the substrate recognition camera 5 captures the mounting area of ​​a specific component such as a BGA to acquire the master image MI, and immediately before mounting the specific component, the same substrate recognition camera 5 captures the same mounting area to acquire the inspection image EI. In inspection example 1, the substrate recognition camera 5 captures the inspection image EI and also acts as a first camera that captures the master image MI before capturing the inspection image EI.

[0044] Inspection Example 2 is an example in which the master image MI and the inspection image EI are separately captured by two substrate recognition cameras 5A and 5B installed on one mounter 13. Specifically, when the substrate P is loaded, the master image MI is captured by one substrate recognition camera 5A (the second camera). Then, immediately before mounting the specific component, the inspection image EI is captured by the other substrate recognition camera 5B (the first camera).

[0045] Inspection Example 3 is an example in which the master image MI and the inspection image EI are separately captured by substrate recognition cameras 5 installed on different mounters 13A and 13C. Specifically, when the substrate P is loaded onto the first mounter 13A, the master image MI is captured by the substrate recognition camera 5 (the second camera) of the first mounter 13A. Then, when the substrate P is loaded onto the third mounter 13C by the conveyor 10 and immediately before mounting the specific component, the inspection image EI is captured by the substrate recognition camera 5 (the first camera) of the third mounter 13C. The inspection image EI may also be captured by the substrate recognition camera 5 of the second mounter 13B.

[0046] In the foreign object inspection using the master image MI and the inspection image EI described above, there may be cases where change factors other than the presence of foreign objects appear in the difference image DI, and as a result, foreign objects cannot be accurately detected. For example, the fixed state of the substrate P at the working position may vary between the acquisition of the master image MI and the acquisition of the inspection image EI due to the impact during component mounting or the operations of substrate conveyance / loading and unloading. In this case, image misalignment may occur between the two images due to the posture deviation of the substrate P, and this may appear in the difference image DI. Also, when the master image MI and the inspection image EI are acquired by different cameras, image misalignment such as scale difference may occur between the two images due to the individual differences of the cameras, and this may appear in the difference image DI. The image misalignment is often a rotational misalignment in which the inspection image EI rotates mainly with respect to the master image MI.

[0047] In inspection example 1 of FIG. 6, the fixing state of the substrate P may vary due to the impact of component mounting executed between the imaging of the master image MI and the imaging of the inspection image EI. In inspection example 2, the problem is the variation in the fixing state of the substrate P due to the impact of component mounting similar to that in inspection example 1, as well as the influence of the individual differences and fixed angle differences of the two substrate recognition cameras 5A and 5B. In inspection example 3, in addition to the variation in the fixing state of the substrate P due to the impact of component mounting and the individual differences and fixed angle differences between the two substrate recognition cameras 5 and 5 of the two mounters 13A and 13C, the variation in the fixing state of the substrate P accompanying the conveyance of the substrate P from the first mounter 13A to the third mounter 13C also becomes a problem. Hereinafter, an embodiment for correcting such an image shift between the master image MI and the inspection image EI before deriving the difference image DI will be described.

[0048] [First Embodiment] FIGS. 7(A) and 7(B) are diagrams showing a first embodiment of image correction for correcting the image shift between the master image MI and the inspection image EI. In FIG. 7(A), an example in which the inspection image EI is rotated counterclockwise with respect to the master image MI is shown. When an impact during component mounting is applied to the substrate P, the substrate P rotates at the fixed position, and as a result, a rotational shift may occur in which the inspection image EI is in a rotated state with respect to the master image MI. In such a case, the abnormality detection unit 84 (image correction unit 841) of the control device 8 executes the following image correction.

[0049] The master image MI and the inspection image EI are imaged so that at least two fiducial marks Fid separated from each other are reflected as the attention position AP. In the example of FIG. 7(A), the master image MI and the inspection image EI include a pair of fiducial marks Fid1 and Fid2 located at the diagonal of the rectangular image. The image processing unit 82 of the control device 8 performs image processing for recognizing the fiducial marks Fid1 and Fid2 when the master image MI and the inspection image EI are respectively acquired. Further, the image processing unit 82 specifies the coordinates of the fiducial marks Fid1 and Fid2 based on the result of the recognition.

[0050] The image correction unit 841 of the anomaly detection unit 84 determines a first straight line L1 drawn between two fiducial marks Fid1 and Fid2 in the master image MI as a first feature element. The image correction unit 841 also determines a second straight line L2 drawn between two fiducial marks Fid1 and Fid2 in the inspection image EI as a second feature element. Then, the image correction unit 841 performs image correction by rotating at least one of the master image MI or the inspection image EI so that the first straight line L1 and the second straight line L2 overlap.

[0051] Specifically, the image correction unit 841 determines the slope and length of the first line L1 and the second line L2, respectively. Figure 7(B) shows the slopes of lines L1 and L2. The first line L1 is represented by the linear function y = ax + c and is a line with a slope a. The angle that the first line L1 makes with the x-axis is called the first angle α. The second line L2 is represented by the linear function y = bx + c and is a line with a slope b. The angle that the second line L2 makes with the x-axis is called the second angle β.

[0052] The angle formed by the second line L2 with respect to the first line L1 is the rotational displacement angle R of the inspection image EI relative to the master image MI. In other words, the first line L1 and the second line L2 are lines connecting a pair of fiducial marks Fid1 and Fid2, which are common points of interest AP in the master image MI and the inspection image EI. When there is no rotational displacement between the master image MI and the inspection image EI, the first line L1 and the second line L2 completely overlap. That is, the slopes of both lines L1 and L2 are a = b. On the other hand, when the rotational displacement occurs, the slope is a ≠ b, and the first line L1 and the second line L2 intersect with a rotational displacement angle R. The rotational displacement angle R is expressed by the following equation: R = tan －１ R [rad] Also, since the rotational displacement angle R is the difference between the first angle α and the second angle β, it can be expressed as tanR = tan(α - β) = (a - b) / (1 + ab).

[0053] In other words, by performing a process to calculate the rotational misalignment angle R between the first line L1 and the second line L2, which are derived from the master image MI and the inspection image EI respectively, it is possible to determine whether or not the inspection image EI is rotated relative to the master image MI. That is, if the rotational misalignment occurs, the rotational misalignment angle R is detected without the first line L1 and the second line L2 overlapping, so the occurrence of the rotational misalignment can be easily detected. Furthermore, image correction to correct the rotational misalignment can also be easily performed. The image correction only requires rotating at least one of the master image MI or the inspection image EI by the rotational misalignment angle R so that the first line L1 and the second line L2 overlap. Therefore, the image correction process can be simplified.

[0054] The image correction may be performed on at least one of the master image MI and the examination image EI, or on both. In the latter case, the correction can be performed by rotating both the master image MI and the examination image EI so that the first angle α and the second angle β shown in Figure 7(B) are both zero degrees, that is, so that the first line L1 and the second line L2 are aligned with the x-axis.

[0055] In addition to the rotational misalignment correction described above, the image correction unit 841 can perform a correction that adjusts the scale ratio between the master image MI and the inspection image EI based on the length ratio of the first straight line L1 and the second straight line L2. As shown in inspection examples 2 and 3 in Figure 6, if the master image MI and the inspection image EI are captured by different cameras, the scale ratios of the two images MI and EI may not match. If the scale ratios of the master image MI and the inspection image EI match, the lengths of the first straight line L1 and the second straight line L2 will be the same. However, if the scale ratios do not match, the lengths of the first straight line L1 and the second straight line L2 will be different.

[0056] In other words, by determining the length ratio of the first line L1 and the second line L2, it is possible to confirm whether the scale ratio of the master image MI and the inspection image EI match. If the scale ratios do not match, the lengths of the first line L1 and the second line L2 will differ, making it easy to detect the difference in the scale ratios of both images MI and EI. Image correction to correct the difference in scale ratio only requires reducing or enlarging at least one of the master image MI or the inspection image EI so that the first line L1 and the second line L2 become the same length. Therefore, the image correction process can be simplified.

[0057] Figure 8 shows an image correction method according to a modified example of the first embodiment. In the example above, the focus points were a pair of fiducial marks Fid1 and Fid2 located diagonally opposite each other in the rectangular image. In the example in Figure 8, the focus points are a pair of fiducial marks Fid1 and Fid3 along the long side of the rectangular image, or a pair of fiducial marks Fid1 and Fid4 along the short side.

[0058] The image correction unit 841 determines a first straight line L11 and a second straight line L12 drawn between a pair of fiducial marks Fid1 and Fid3 along the long sides of the master image MI and the inspection image EI. Alternatively, the image correction unit 841 determines a first straight line L21 and a second straight line L22 drawn between a pair of fiducial marks Fid1 and Fid4 along the short sides of the master image MI and the inspection image EI. Furthermore, the image correction unit 841 determines the inclination and length of the first straight lines L11, L21 and the second straight lines L12, L22.

[0059] If a rotational misalignment angle R is detected between the first straight line L11 and the second straight line L12, or between the first straight line L21 and the second straight line L22, the image correction unit 841 performs image correction by rotating at least one of the master image MI or the inspection image EI so that the first straight line L11 and the second straight line L12, or the first straight line L21 and the second straight line L22, overlap. If there is a difference in the length between the first straight line L11 and the second straight line L12, or between the first straight line L21 and the second straight line L22, the image correction unit 841 performs image correction by reducing or enlarging at least one of the master image MI or the inspection image EI so that the first straight line L11 and the second straight line L12, or the first straight line L21 and the second straight line L22, become the same length.

[0060] Figure 9 is a flowchart showing the component mounting process on a substrate P, including image correction according to the first embodiment. When the process starts, the axis control unit 83 of the control device 8 (Figure 4) controls the conveyors 21 and 22 to transport the substrate P to the mounter 13 (step S1). The transported substrate P is fixed to a predetermined working position on the mounter 13 for component mounting.

[0061] The axis control unit 83 controls the X-axis servo motor 26 and the Y-axis servo motor 28 to move the head unit 4 towards the position on the substrate P where the fiducial marks Fid1 and Fid2 are located (step S2). The axis control unit 83 stops the head unit 4 at a position where the substrate recognition camera 5 mounted on the head unit 4 can capture images of the fiducial marks Fid1 and Fid2 set to the point of interest AP. The imaging control unit 81 causes the substrate recognition camera 5 to capture images of the fiducial marks Fid1 and Fid2 (step S3). If all of the fiducial marks Fid1 and Fid2 set to the point of interest AP do not fit within a single field of view, the imaging control unit 81 causes the substrate recognition camera 5 to take multiple images.

[0062] The image processing unit 82 performs image processing on the image acquired in step S3 to recognize fiducial marks Fid1 and Fid2. Furthermore, the image processing unit 82 identifies the coordinates of fiducial marks Fid1 and Fid2 based on the recognition results (step S4). Subsequently, the image correction unit 841 of the anomaly detection unit 84 creates a first straight line L1 connecting the pair of fiducial marks Fid1 and Fid2, and calculates the slope and length of the first straight line L1 (step S5).

[0063] Subsequently, the imaging control unit 81 causes the substrate recognition camera 5 to image the surface of the substrate P in order to acquire a master image MI (step S6). The image acquired here is an image of the region of the substrate P that does not include the fiducial marks Fid1 and Fid2, but includes the mounting region PA of a specific component. In this embodiment, the master image MI exemplified in Figure 7(A) is a composite image of the images of the fiducial marks Fid1 and Fid2 acquired in step S3 and the image of the region including the mounting region PA acquired in step S6. Note that if both the fiducial marks Fid1 and Fid2 and the mounting region PA can be captured in a single image, the image acquisition in step S6 may be omitted.

[0064] Once the master image MI is acquired, the axis control unit 83 causes the head unit 4 to start the component mounting operation on the substrate P (step S7). The inspection control unit 851 of the main control unit 85 determines, based on component data etc. stored in the memory unit 86, whether or not the component to be mounted next by the head unit 4 is a component subject to foreign object inspection (step S8). If it is not a component subject to foreign object inspection (NO in step S8), the next component is mounted. If component mounting is to be performed in the downstream mounter 13, the substrate P is transported. For example, the substrate P is transported from the first mounter 13A to the third mounter 13C, and then the next component is mounted.

[0065] On the other hand, if the part is subject to foreign object inspection (YES in step S8), the inspection control unit 851 causes each part to perform the following processes in order to acquire the inspection image E1 and correct the image. The axis control unit 83 moves the head unit 4 to the position where the fiducial marks Fid1 and Fid2 are attached to the substrate P (step S9). After moving, the imaging control unit 81 causes the substrate recognition camera 5 to capture images of the fiducial marks Fid1 and Fid2 (step S10).

[0066] The image processing unit 82 obtains the coordinates of fiducial marks Fid1 and Fid2 based on the image acquired in step S10 (step S11). The image correction unit 841 creates a second straight line L2 connecting the pair of fiducial marks Fid1 and Fid2, and calculates the slope and length of the second straight line L2 (step S12). Subsequently, the imaging control unit 81 causes the substrate recognition camera 5 to image the surface of the substrate P in order to acquire an inspection image EI (step S13).

[0067] The image correction unit 841 calculates a correction value based on the first straight line L1 calculated in step S5 and the second straight line L2 calculated in step S12 (step S14). Specifically, the image correction unit 841 calculates the rotational displacement angle R between the first straight line L1 and the second straight line L2, and the length ratio between the first straight line L1 and the second straight line L2 as the correction value.

[0068] Next, the image correction unit 841 performs image correction (step S15). If a rotational misalignment angle R is detected, image correction is performed to rotate at least one of the master image MI or the inspection image EI by the amount of the rotational misalignment angle R so that the first line L1 and the second line L2 overlap. If the lengths of the first line L1 and the second line L2 do not match, image correction is performed to reduce or enlarge the image of at least one of the master image MI or the inspection image EI so that they become the same length.

[0069] After image correction, the image comparison unit 843 derives a difference image DI of the master image MI and the inspection image EI based on the brightness difference between the two images (step S16). The determination unit 844 calculates the area of ​​the portion in the difference image DI that appears as a brightness difference between the master image MI and the inspection image EI. The determination unit 844 checks whether the calculated brightness difference area exceeds a predetermined threshold. That is, the determination unit 844 determines whether there is a difference of a certain area or more between the master image MI and the inspection image EI (step S17).

[0070] If there is no difference of a certain area or more between the master image MI and the inspection image EI (NO in step S17), it means that there are no foreign objects in the mounting area PA. In this case, the inspection control unit 851 causes the inspection target component to be mounted onto the substrate P (step S18). After the mounting of the component is completed, the process returns to step S8 and is repeated. On the other hand, if there is a difference of a certain area or more (YES in step S17), it is highly likely that there are foreign objects in the mounting area PA. In this case, the inspection control unit 851 stops the mounting of the inspection target component by the head unit 4 and issues an error alert (step S19).

[0071] [Second Embodiment] In the second embodiment, an example is shown in which a specific solder pattern is used as the focus point AP instead of the fiducial mark Fid. Figure 10 is a diagram showing the focus point AP in image correction according to the second embodiment. Figure 10 shows the mounting area PAA acquired in one image capture by the substrate recognition camera 5. Solder patterns SP for joining with BGA components are arranged in a grid pattern in the mounting area PAA. Of these solder patterns SP, specific solder patterns SP1 and SP2 are set as the focus point AP. In Figure 10, an example is shown in which solder patterns SP1 and SP2 located diagonally across the rectangular field of view are set as the focus point AP.

[0072] Figure 11 shows an example of image correction that corrects image misalignment between a master image MI and an inspection image EI using solder patterns SP1 and SP2. In these images MI and EI, the solder patterns SP appear as circles. Similar to the first embodiment, the image correction unit 841 determines a first straight line L1 drawn between a pair of solder patterns SP1 and SP2 in the master image MI as a first feature element. The image correction unit 841 also determines a second straight line L2 drawn between a pair of solder patterns SP1 and SP2 in the inspection image EI as a second feature element. Based on the slopes of the obtained first straight line L1 and second straight line L2, the rotational misalignment angle R of both straight lines L1 and L2 is determined. Then, the image correction unit 841 performs image correction by rotating at least one of the master image MI or the inspection image EI by the rotational misalignment angle R so that the first straight line L1 and the second straight line L2 overlap.

[0073] Furthermore, the image correction unit 841 performs correction to adjust the scale ratio between the master image MI and the inspection image EI based on the length ratio of the first straight line L1 and the second straight line L2. If the lengths of the first straight line L1 and the second straight line L2 are different, the image correction unit 841 performs image correction to reduce or enlarge at least one of the master image MI or the inspection image EI so that both straight lines L1 and L2 become the same length.

[0074] Figure 12 shows an image correction method according to a modified example of the second embodiment. In the example above, a pair of solder patterns SP1 and SP2 located diagonally opposite each other in a rectangular image were designated as the focus points. In the example of Figure 12, a pair of solder patterns SP1 and SP3 along the long side of the rectangular image, or a pair of solder patterns SP1 and SP4 along the short side, are designated as the focus points.

[0075] The image correction unit 841 determines a first straight line L11 and a second straight line L12 drawn between a pair of solder patterns SP1 and SP3 along the long sides of the master image MI and the inspection image EI. Alternatively, the image correction unit 841 determines a first straight line L21 and a second straight line L22 drawn between a pair of solder patterns SP1 and SP4 along the short sides of the master image MI and the inspection image EI. Furthermore, the image correction unit 841 determines the slope and length of the first straight lines L11, L21 and the second straight lines L12, L22.

[0076] The image correction unit 841 determines the rotational displacement angle R between the first line L11 and the second line L12, or between the first line L21 and the second line L22. Then, the image correction unit 841 performs image correction by rotating at least one of the master image MI or the inspection image EI by the rotational displacement angle R so that the first line L11 and the second line L12, or the first line L21 and the second line L22, overlap. In addition, the image correction unit 841 performs image correction by reducing or enlarging at least one of the master image MI or the inspection image EI so that the first line L11 and the second line L12, or the first line L21 and the second line L22, are the same length.

[0077] Figure 13 is a flowchart showing the component mounting process on the substrate P, including image correction according to the second embodiment. In the following description, operations similar to those in the first embodiment will be omitted. The image correction is assumed to use the diagonally positioned solder patterns SP1 and SP2 shown in Figure 11. When the process starts, the substrate P is loaded into the mounter 13 (step S21), and the substrate P is fixed to a predetermined working position on the mounter 13. The head unit 4 is moved to a position where the mounted substrate recognition camera 5 can capture images of the specific solder patterns SP1 and SP2 designated as the point of interest AP on the substrate P (step S22).

[0078] The imaging control unit 81 causes the substrate recognition camera 5 to image the mounting area PAA, which includes specific solder patterns SP1 and SP2 (step S23). The difference from the first embodiment is that the image acquired in step S23 includes both the mounting area PAA and the solder patterns SP1 and SP2 at the point of interest AP. In other words, the image in step S23 is an image for image correction and is also a master image MI for foreign object inspection.

[0079] The image processing unit 82 performs image processing to recognize the solder patterns SP1 and SP2 in the image acquired in step S23, and also performs processing to identify the coordinates of the solder patterns SP1 and SP2 (step S24). Subsequently, the image correction unit 841 of the anomaly detection unit 84 creates a first straight line L1 connecting the pair of solder patterns SP1 and SP2, and calculates the slope and length of the first straight line L1 (step S25).

[0080] Next, the axis control unit 83 causes the head unit 4 to start the component mounting operation on the substrate P (step S26). The inspection control unit 851 of the main control unit 85 determines whether the next component to be mounted by the head unit 4 is a component subject to foreign object inspection (step S27). If it is not a component subject to foreign object inspection (NO in step S27), the next component is mounted.

[0081] On the other hand, if the part is subject to foreign object inspection (YES in step S27), the inspection control unit 851 moves the head unit 4 to the position where the solder patterns SP1 and SP2 are located on the substrate P in order to acquire the inspection image E1 and correct the image (step S28). After moving, the imaging control unit 81 causes the substrate recognition camera 5 to image the mounting area PAA including the solder patterns SP1 and SP2 (step S29). Similar to step S23, the image captured in step S29 includes the solder patterns SP1 and SP2 and the mounting area PAA, and is therefore an image for image correction, and is also an inspection image EI for foreign object inspection.

[0082] The image processing unit 82 obtains the coordinates of solder patterns SP1 and SP2 based on the image acquired in step S29 (step S30). The image correction unit 841 creates a second straight line L2 connecting the pair of solder patterns SP1 and SP2, and calculates the slope and length of the second straight line L2 (step S31).

[0083] The image correction unit 841 calculates a correction value (step S32) based on the first straight line L1 calculated in step S25 and the second straight line L2 calculated in step S31. Specifically, the image correction unit 841 calculates the rotational displacement angle R between the first straight line L1 and the second straight line L2, and the length ratio between the first straight line L1 and the second straight line L2 as the correction value.

[0084] Next, the image correction unit 841 performs image correction (step S33). If a rotational misalignment angle R is detected, image correction is performed to rotate at least one of the master image MI or the inspection image EI by the amount of the rotational misalignment angle R so that the first line L1 and the second line L2 overlap. If the lengths of the first line L1 and the second line L2 do not match, image correction is performed to reduce or enlarge the image of at least one of the master image MI or the inspection image EI so that they become the same length.

[0085] After image correction, the image comparison unit 843 derives a difference image DI between the master image MI and the inspection image EI based on the brightness difference between the two images (step S34). Based on the difference image DI, the determination unit 844 checks whether the brightness difference area between the master image MI and the inspection image EI exceeds a predetermined threshold. That is, the determination unit 844 determines whether there is a difference of a certain area or more between the master image MI and the inspection image EI (step S35).

[0086] If there is no difference of a certain area or more between the master image MI and the inspection image EI (NO in step S35), the inspection control unit 851 causes the inspection target component to be mounted onto the substrate P (step S36). After the mounting of the component is completed, the process returns to step S27 and is repeated. On the other hand, if there is a difference of a certain area or more (YES in step S35), the inspection control unit 851 stops the mounting of the inspection target component by the head unit 4 and issues an error alert (step S37).

[0087] In step S24 described above, when identifying the coordinates of solder patterns SP1 and SP2, a step may be added to determine whether these solder patterns SP1 and SP2 can be used as points of interest AP. Solder patterns SP may not be suitable for use as points of interest AP due to factors such as dirt, overlapping foreign matter, or printing defects. Therefore, the usability of specific solder patterns SP1 and SP2 can be determined, and if they are deemed unusable, other solder patterns SP may be used as points of interest AP. For example, if solder pattern SP1 is unusable, pairs of solder patterns SP3 and SP4, pairs of solder patterns SP2 and SP4, or pairs of solder patterns SP2 and SP3 can be used as points of interest AP.

[0088] According to the second embodiment, the image acquired for image correction and the master image MI or inspection image EI for foreign object inspection can be the same image. Therefore, the number of images taken by the substrate recognition camera 5 can be reduced, and the cycle time can be improved. However, since the shape accuracy of the solder pattern SP is inferior to that of the fiducial mark Fid, the accuracy of image correction is superior in the first embodiment. Therefore, the decision of whether to adopt the first or second embodiment may be made by considering whether to prioritize shortening the cycle time or improving the accuracy of foreign object inspection. Depending on the components to be mounted, it may also be possible to select whether to use the fiducial mark Fid or the solder pattern SP for image correction.

[0089] [Embodiment with added brightness calibration] Figure 14 is a block diagram showing an embodiment in which brightness correction is added to the first or second embodiment described above. In inspection example 3 shown in Figure 6, the master image MI and the inspection image EI are captured separately by the substrate recognition cameras 5 equipped on the first mounter 13A and the third mounter 13C, respectively. In Figure 6, the scale difference due to individual differences between the substrate recognition cameras 5, 5 of the two mounters 13A and 13C was mentioned, but it is desirable to also consider the brightness difference between different cameras.

[0090] When performing brightness calibration, when the substrate recognition camera 5 of the first mounter 13A captures the master image MI, the calibration plate RP shown in Figure 2 is also captured to obtain "Brightness Calibration Image 1". The calibration plate RP is located on the device base 130 and is installed within the range from which the head unit 4 can move. A brightness calibration pattern is drawn on the calibration plate RP. Similarly, when the substrate recognition camera 5 of the third mounter 13C captures the inspection image EI, the calibration plate RP is also captured to obtain "Brightness Calibration Image 2".

[0091] The luminance calibration unit 842 of the control device 8 performs luminance calibration correction based on "luminance calibration image 1" and "luminance calibration image 2" before the image comparison unit 843 detects the difference image DI. When acquiring a master image MI and an inspection image EI with different cameras, images with different luminances may be acquired between the cameras. For example, even when imaging is performed under the same lighting conditions, the luminance of the images acquired by the substrate recognition camera 5 of the first mounter 13A and the substrate recognition camera 5 of the third mounter 13C may differ. Also, as the lighting environment changes over time, the luminance of the images acquired by the two substrate recognition cameras 5 may differ. If the difference in luminance between the two substrate recognition cameras 5 is detected as the difference between the master image MI and the inspection image EI, the difference image DI will become inaccurate.

[0092] The luminance calibration unit 842 determines a luminance correction value so that "luminance calibration image 1" and "luminance calibration image 2" have the same luminance. This luminance correction value is referenced when the image comparison unit 843 compares the master image MI and the inspection image EI, so that the difference in luminance between the two substrate recognition cameras 5 is not detected as a difference in the difference image DI. This ensures the accuracy of the difference image DI.

[0093] [Other Embodiments] The present invention is not limited to the embodiments described above, and other embodiments may be adopted. For example, the image correction unit 841 of the anomaly detection unit 84 performs image correction such as image rotation and scale adjustment on either the master image MI or the inspection image EI, or both. In the former case, that is, when one of the master image MI or the inspection image EI is a correction image to be corrected and the other is an image that does not require correction, the image that does not require correction may be corrected so that its resolution matches that of the correction image.

[0094] Image corrections such as rotation, enlargement, and reduction can be performed using affine transformations. Generally, images subjected to affine transformations have reduced resolution. When either the master image MI or the examination image EI is image-corrected, the difference in resolution between the two images may be detected as a difference in the difference image DI. For example, when image correction is performed on the examination image EI, the resolution of the master image MI is corrected to match the resolution of the examination image EI after affine transformation. This suppresses the intervention of differences based on the resolution difference between the master image MI and the examination image EI.

[0095] [Inventions included in the above embodiments] The embodiments described above include the following inventions.

[0096] A substrate work apparatus according to one aspect of the present invention comprises: a work unit that performs predetermined work on a substrate on which components are mounted; a first camera that captures an image of the substrate and acquires an inspection image before the work unit performs the work on the substrate; a storage unit that stores a master image showing the initial state of the substrate; an anomaly detection unit that detects the difference between the master image and the inspection image and recognizes an anomaly in the substrate on which the work unit is performing the work; and a control unit that controls the operation of the work unit based on the result of the anomaly recognition. The anomaly detection unit, before detecting the difference, derives a first feature element for the master image and a second feature element for the inspection image based on a common point of interest in the master image and the inspection image, and corrects at least one of the master image or the inspection image so that the first feature element and the second feature element overlap.

[0097] In this embodiment, the first and second feature elements derived from a common point of interest in both the master image and the inspection image are corrected so that they overlap. Therefore, if there is a discrepancy in the substrate orientation between the master image and the inspection image, for example, if the substrate in one image is rotated relative to the substrate in the other, such image discrepancies can be corrected by image correction. Furthermore, even if the image scales of the master image and the inspection image are different, this can also be corrected by the image correction. Consequently, abnormalities in the substrate can be accurately recognized based on the difference between the master image and the inspection image.

[0098] In the above-described substrate processing apparatus, the first camera may image the substrate in order to acquire the master image before capturing the inspection image.

[0099] In this embodiment, the first camera is also used for acquiring the master image. Therefore, the device configuration can be simplified. Furthermore, since both the master image and the inspection image are acquired with the same camera, image correction to compensate for differences between the two images caused by individual differences in the cameras is unnecessary.

[0100] The above-described substrate processing apparatus may further include a second camera that captures images of the substrate in order to acquire the master image.

[0101] According to this embodiment, since master images and inspection images are acquired by different cameras, the degree of freedom in arranging the work units can be increased, such as by arranging multiple work units in tandem.

[0102] In the above-described substrate processing apparatus, it is desirable to further include a transport unit for transporting the substrate to the work position of the work unit, the master image is captured when the substrate is transported to the work position of the work unit, the inspection image is captured immediately before the work unit performs the work on the substrate, and the control unit preferably stops the work performed by the work unit when the abnormality detection unit recognizes an abnormality.

[0103] In this embodiment, the image acquired when the circuit board is brought into the work position becomes the master image, and the image acquired immediately before performing a predetermined operation on the circuit board becomes the inspection image. Therefore, it is possible to detect the difference between the master image and the inspection image by eliminating the influence of factors causing image misalignment that occurred between the time the circuit board was brought in and immediately before the operation, such as misalignment of the circuit board's orientation due to the mounting of other components. In addition, since the operation is stopped when an abnormality is detected, unnecessary work and component loss can be avoided.

[0104] In the above-described substrate processing apparatus, it is desirable that the abnormality detection unit determines a first straight line drawn between two spaced-apart points of interest in the master image as the first feature element, and a second straight line drawn between two spaced-apart points of interest in the inspection image as the second feature element, and performs image correction by rotating at least one of the master image and the inspection image so that the first straight line and the second straight line overlap.

[0105] According to this embodiment, if the substrate of one image (master image or inspection image) is rotated differently from the substrate of the other image, the first and second lines will not overlap, making it easy to detect the rotational misalignment. Furthermore, image correction can be simplified by simply rotating the image so that the first and second lines overlap.

[0106] In the above-described substrate processing apparatus, when the inspection image is captured by the first camera and the master image is captured by the second camera, it is desirable that the abnormality detection unit perform a correction to adjust the scale ratio between the inspection image and the master image based on the length ratio of the first straight line and the second straight line.

[0107] According to this embodiment, if the scale ratio of the master image and the inspection image do not match, the lengths of the first line and the second line will differ, making it easy to detect the difference in the scale ratio of the two images. Furthermore, image correction only requires reducing or enlarging at least one of the images so that the first line and the second line are the same length, thus simplifying the image correction process.

[0108] In the above-described circuit board handling apparatus, the location of interest may be a circuit board recognition mark attached to the circuit board. Alternatively, the location of interest may be a specific solder pattern applied to the circuit board.

[0109] A circuit board on which components are mounted generally has marks or solder patterns for circuit board recognition. According to the above embodiment, these general-purpose parts are used as points of interest, so it is not necessary to create special points of interest on the circuit board for image correction.

[0110] Furthermore, the abnormality detection unit determines whether the specific solder pattern can be used, and if it determines that it cannot be used, it may use another solder pattern as the focus position. This allows for flexibility in setting the focus position.

[0111] In the above-described substrate processing apparatus, the first camera and the second camera may each capture calibration images capable of detecting brightness, and the anomaly detection unit may, before detecting the difference, perform brightness calibration correction based on the calibration images acquired by the first camera and the second camera, respectively.

[0112] When acquiring master images and inspection images with different cameras, images with different brightness levels may be acquired between the cameras. For example, even when imaging is performed under the same lighting conditions, the brightness of the acquired images may differ between the two cameras. Furthermore, changes in the lighting environment over time may also cause differences in the brightness of the acquired images between the two cameras. According to the above embodiment, the intervention of differences based on such differences in brightness between cameras can be suppressed.

[0113] In the above-described substrate processing apparatus, the abnormality detection unit may perform image correction on both the master image and the inspection image so that the first feature element and the second feature element overlap.

[0114] According to this embodiment, it is possible to more accurately correct for pose discrepancies and scale differences between the master image and the inspection image.

[0115] In the above-described substrate processing apparatus, the abnormality detection unit may, when one of the master image or the inspection image is a correction image to be corrected and the other is an image that does not require correction, correct the image that does not require correction so that its resolution matches that of the correction image.

[0116] Generally, corrected images often have a lower resolution than the original image. When either the master image or the inspection image is corrected, the difference in resolution between the two images can be detected as a difference. According to the above embodiment, the intervention of differences based on the resolution difference between images can be suppressed.

[0117] As described above, the present invention provides a substrate work apparatus that can accurately recognize abnormalities in a substrate P when performing a predetermined operation on the substrate P, such as component mounting.

Claims

1. A substrate work apparatus comprising: a work unit that performs predetermined work on a substrate on which components are mounted; a first camera that captures an image of the substrate and acquires an inspection image before the work unit performs the work on the substrate; a storage unit that stores a master image showing the initial state of the substrate; an anomaly detection unit that detects the difference between the master image and the inspection image and recognizes an anomaly in the substrate on which the work unit is performing the work; and a control unit that controls the operation of the work unit based on the result of the anomaly recognition, wherein, before detecting the difference, the anomaly detection unit derives a first feature element for the master image and a second feature element for the inspection image based on a common point of interest in the master image and the inspection image, and corrects the image of at least one of the master image or the inspection image so that the first feature element and the second feature element overlap.

2. A substrate processing apparatus according to claim 1, wherein the first camera images the substrate in order to acquire the master image before capturing the inspection image.

3. A substrate processing apparatus according to claim 1, further comprising a second camera for imaging the substrate in order to acquire the master image.

4. A substrate processing apparatus according to claim 2 or 3, further comprising a transport unit for transporting the substrate to the work position of the work unit, wherein the master image is captured when the substrate is transported to the work position of the work unit, the inspection image is captured immediately before the work unit performs the work on the substrate, and the control unit stops the work performed by the work unit when the abnormality detection unit recognizes an abnormality.

5. A substrate processing apparatus according to any one of claims 1 to 3, wherein the abnormality detection unit determines, as the first characteristic element, a first straight line drawn between two spaced-apart points of interest in the master image, and as the second characteristic element, a second straight line drawn between two spaced-apart points of interest in the inspection image, and performs image correction by rotating at least one of the master image and the inspection image so that the first straight line and the second straight line overlap.

6. A substrate processing apparatus according to claim 5, wherein when the inspection image is captured by the first camera and the master image is captured by the second camera, the abnormality detection unit performs a correction to adjust the scale ratio between the inspection image and the master image based on the length ratio of the first straight line and the second straight line.

7. A substrate work apparatus according to any one of claims 1 to 3, wherein the position of interest is a substrate recognition mark attached to the substrate.

8. A substrate work apparatus according to any one of claims 1 to 3, wherein the position of interest is a specific solder pattern applied to the substrate.

9. A circuit board work apparatus according to claim 8, wherein the abnormality detection unit determines whether the specific solder pattern can be used, and when it determines that it cannot be used, uses another solder pattern as the position of interest.

10. A substrate processing apparatus according to claim 3, wherein the first camera and the second camera each capture a calibration image capable of detecting brightness, and the abnormality detection unit performs brightness calibration correction based on the calibration images acquired by the first camera and the second camera, respectively, before detecting the difference.

11. A substrate processing apparatus according to claim 1, wherein the abnormality detection unit corrects both the master image and the inspection image so that the first feature element and the second feature element overlap.

12. A substrate processing apparatus according to claim 1, wherein the abnormality detection unit corrects the non-correction image so that the resolution of the non-correction image matches the resolution of the correction image when one of the master image or the inspection image is a correction image to be corrected and the other is a non-correction image that does not require correction.

Images