JP2005249615A - Different material inclusion-inspecting method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a different material inclusion-inspecting method responsive to a change in inspection environment. <P>SOLUTION: In the different material inclusion-inspecting method, a continuously supplied workpiece 1 is siccessively imaged by a camera 10 to acquire image data and the inclusion of a different workpiece is inspected on the basis of the image data. The inspecting method is equipped with a step (S130) for registering the image data S of the first workpiece 1 (master workpiece) as reference image data N, a step (S150) for comparing the image data P of each workpiece 1 supplied in succession to the master workpiece with the reference image data N to determine whether the workpiece 1 is assumed to be the same and a step (S170) for setting the image data P of the workpiece 1 determined to be the same to renew the reference image data N. Since the image data (S or P) of the workpiece 1 determined to be the same is set to the reference image data N to be compared with the image data P of the work 1, the inspection method is responsive to a change in inspection environment. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a method for inspecting mixing of different kinds of workpieces, in which continuously supplied workpieces are sequentially picked up by an image pickup device to obtain image data, and based on the image data, inspection of mixing of different kinds of workpieces is performed. is there.

  2. Description of the Related Art Conventionally, there is known a mixed product inspection method for sequentially capturing workpieces that are continuously supplied with an imaging device, acquiring image data, and inspecting mixing of different types of workpieces based on the image data. . In this case, the master work is imaged in advance and reference image data (for example, the number of reference pixels) is registered, and the registered reference image data is read out every time the same kind of the master work is produced, and the work image data is read out. And the reference image data are compared to determine whether they can be regarded as the same product type.

  However, if the master work is imaged in advance and the reference image data is registered, the inspection environment changes between the reference image data registration by the master work and the inspection of the current work (for example, illumination illuminance change, line layout change). If there is a difference in the paint state of the workpiece (for example, black and matte black), the illuminance (reflected light from the workpiece) detected by the imaging device will be different even for the same workpiece. There is a risk that a procedure such as changing the threshold value or re-registering the reference image data may be required. In particular, when producing many types (for example, several hundred types) on one line, it is necessary to readjust threshold values and re-register reference image data for all types, which requires a lot of man-hours. .

  The present invention has been made in view of the above-described problems, and an object of the present invention is to provide an inspection method for mixing different types that is resistant to changes in the inspection environment.

  In order to achieve the above object, the inspection method for mixed product of different types according to claim 1 obtains image data by sequentially imaging workpieces that are continuously supplied with an imaging device, and based on the image data, This is a method for inspecting mixing of different kinds of workpieces. Then, a registration step of registering the image data of the master work supplied first as reference image data, and comparing the image data of each work supplied following the master work with the reference image data A determination step for determining whether or not the images can be regarded, and an update step for updating the reference image data with the image data of the workpiece determined to be the same type determined by the determination step as new reference image data. .

  As described above, according to the inspection method for mixing different kinds of the present invention, the image data is registered as the reference image data as it is only for the first supplied master work (non-defective product determination). The image data of the current workpiece is compared with the image data of the workpiece determined as the same type as the reference image data.

  Therefore, since the latest image data determined with the same product type is always compared as the reference image data, even if the illumination illuminance varies depending on the production timing or when the illumination changes gradually during production, It is possible to execute the mixed product inspection without taking steps such as adjusting the threshold value, changing the threshold value, and re-registering the reference image data. That is, the inspection method for mixing different types of the present invention is an inspection method that is more resistant to changes in the inspection environment than the case where master image data is fixedly registered in advance as reference image data.

  Since the master work becomes the reference image data as it is without being compared, it is preferable to take an image after confirming that the master work is a non-defective product.

  In addition, since the inspection method for mixing different kinds of the present invention is strong against changes in the inspection environment such as illumination illuminance changes, as described in claim 2, image data changes particularly in the inspection environment (detected by the imaging device). This is more effective for a configuration obtained by identifying a predetermined area included in an image based on luminance affected by illuminance (change in reflected light from the workpiece). In addition, when acquiring image data by processing other than that (for example, color), the present invention is applicable because it is affected by changes in the inspection environment such as changes in illumination intensity.

  More specifically, as described in claim 3, the number of pixels in a predetermined area extracted by binarization processing may be used as image data. In addition to the black and white binarization processing, gray processing and color processing using RGB data of each pixel can be applied as the luminance processing. In addition to the number of pixels, position (coordinates), average luminance (density), and the like may be applied as image data.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
(First embodiment)
FIG. 1 is a block diagram showing a schematic configuration of an image processing apparatus for executing a different product mixture inspection. In FIG. 1, reference numeral 1 denotes a work on a production line 2 (for example, a conveyor belt) as a subject, and in this embodiment, a sensor is housed in a circuit board on which a sensor chip is mounted in a case. And the workpiece | work 1 is continuously conveyed in the direction of a white arrow as shown in FIG.

  As shown in FIG. 1, the image processing apparatus 100 includes a camera 10 as an imaging apparatus, a control unit 20 that performs predetermined processing on an image obtained by imaging, and a notification unit 30 that outputs a determination result by the control unit 20 to the outside. It consists of.

  The camera 10 is arranged on the production line 2 and sequentially images the workpiece 1 flowing continuously. In the present embodiment, a CCD camera including a CCD element is applied as an imaging element. Then, an image signal obtained by imaging is output to the control unit 20.

  The control unit 20 is composed of a microcomputer, and includes a central processing unit 21 (hereinafter referred to as a CPU) that executes various processes, a memory 22 such as a ROM, an EEPROM, and a RAM, an I / O, and a bus that connects these configurations. It is comprised by a line (not shown).

  The CPU 21 performs predetermined processing on the image signal from the camera 10 to generate image data, stores the image data in the memory 22, and compares the reference image data stored in the memory 22 with the image data. Then, the determination result is output to the notification unit 30 and the line control unit 110. Note that the CPU 21 in the present embodiment binarizes the captured image (signal) into white or black according to a predetermined threshold (brightness), and extracts the number of pixels (image data) in a predetermined area (for example, black area). Count. Further, by comparing the number of pixels with the reference pixel number (reference image data), it is determined whether the imaged work 1 is a non-defective product (same product type) or a defective product (different product type). In the present embodiment, the master work image data is not stored as the reference image data in the memory 22 in advance (fixed registration), but is registered at the time of production of the product type. Details thereof will be described later.

  The notification unit 30 outputs a determination result to the outside based on a determination signal from the CPU 21. For example, a display device such as a CRT or a sound output device that outputs an alarm sound or the like can be applied. In the present embodiment, a CRT that displays a determination result together with a captured image of the work 1 and a sound output device that outputs an alarm sound in the case of a defective product (different product) are applied.

  The line control unit 110 controls the driving state of the production line 2 based on a determination signal from the CPU 21. In the present embodiment, when a different kind determination signal is received, the work 1 that has been determined to have a different kind is delivered to a defective discharge conveyor (not shown) branched from the production line 2.

  The image processing apparatus 100 configured as described above is applied to the inspection for mixing different types of products. Next, the different product mixing inspection method, which is a feature of the present embodiment, will be described using the flow shown in FIG.

  In a state where the production of the workpiece 1 of the predetermined type has started and the workpiece 1 is being transported on the production line 2, the CPU 21 of the control unit 20 receives the detection signal of the workpiece 1 (S100), and the detection signal is It is determined whether or not the workpiece is the first workpiece 1 (S110). In the present embodiment, a non-contact type detection device (not shown) is disposed immediately before the imaging area of the camera 10 in the production line 2, and the workpiece 1 is detected by the detection device. Then, after the production of the previous product is finished (a state in which all the workpieces are paid out from the production line 2), the workpiece 1 corresponding to the detection signal of the workpiece 1 received first by the CPU 21 is determined as the master workpiece. In addition to the detection device, the workpiece 1 may be detected by a detection signal from the camera 10. For example, when the workpiece 1 is supplied to the imaging area of the camera 10 by the robot, the robot 1 It may be a position signal from.

  If it is determined in S110 that the work 1 detected in S100 is the first work, the CPU 21 captures an image using the work 1 as a master work and compares each pixel with a threshold value stored in advance in the memory 22. Then, binarization processing is performed. Then, the number of pixels of the extracted predetermined area (area determined to be black) is acquired as image data S of the master work (S120). Then, the image data S is stored in the memory 22 as the reference image data N (S130).

  In S110, when it is determined that the workpiece 1 detected in S100 is not a master workpiece, the CPU 21 captures an image captured by the camera 10 and compares each pixel with a threshold value stored in advance in the memory 22. Perform binarization. Then, the number of pixels in the extracted predetermined area is acquired as image data P (S140), and this image data P is compared with the reference image data N (S150). Note that a predetermined tolerance for comparison in S150 is stored in the memory 22 in advance, and in S150, the CPU 21 compares the reference image data N with the tolerance and compares it with the image data P.

  Here, when the image data P of the work 1 is compared with the reference image data N and can be regarded as the same product type (the image data P is included within a predetermined range of the reference image data N (including tolerance)), the CPU 21 determines that the work 1 Are determined as non-defective products (same product type) (S160). Specifically, the CPU 21 outputs the image signal of the work 1 and the non-defective signal to the CRT of the notification unit 30 and displays that the CRT of the notification unit 30 is the image of the work 1 and the non-defective product. Further, the CPU 21 outputs a non-defective product signal to the line control unit 110, and the line control unit 110 conveys the workpiece 1 as a non-defective product (same product type) to the next process (after the different product mixing inspection process).

  Further, the CPU 21 newly updates the image data P of the work 1 as the reference image data N and stores it in the memory 22 (S170). That is, when the imaged workpiece 1 is determined as a non-defective product (same product type), the CPU 21 sets the image data P of the workpiece 1 as the reference image data N, and executes the next workpiece 1 determination (S150).

  On the other hand, when the image data P of the work 1 and the reference image data N are compared and cannot be regarded as the same product type (when the image data P is not included in the predetermined range of the reference image data N (including tolerance)), the CPU 21 The workpiece 1 is determined as a defective product (different product) (S180). Specifically, the CPU 21 outputs the image signal and defective product signal of the work 1 to the CRT and sound output device of the notification unit 30. Therefore, the CRT of the notification unit 30 displays the image of the work 1 and the fact that it is a defective product, and the audio output device generates an alarm sound as a defective product. Further, the CPU 21 also outputs a defective product signal to the line control unit 110, and the line control unit 110 conveys the workpiece 1 as a defective product (different product) to a defective discharge conveyor branched from the production line 2. That is, it is excluded from non-defective products. In addition, it is confirmed whether the workpiece | work 1 conveyed by this defect discharge conveyor is a different kind by a separate worker.

  In this case, the CPU 21 does not update the reference image data N, and uses the reference image data N so far to determine the next workpiece 1 (S150).

  Note that while the workpiece 1 is sequentially supplied to the camera 10, the above-described S100 to S180 are repeatedly executed.

  Conventionally, when continuously supplying workpieces 1 are sequentially captured by the camera 10 to acquire image data, and inspection of mixing of different types of workpieces 1 is performed based on the image data, the master workpiece is imaged in advance. Then, reference image data (for example, the number of reference pixels) is registered, and each time the same product type as the master work is produced, the fixedly registered reference image data is read, and the image data of work 1 is compared with the reference image data. However, it was determined whether they could be regarded as the same variety.

  However, if the reference image data is fixedly registered in advance, the inspection environment changes between the registration of the reference image data and the inspection of the workpiece 1 (for example, illumination intensity change, line layout change, workpiece 1 coating state). If there is a difference (for example, black and matte black), the illuminance (reflected light from the work 1) detected by the camera 10 is different even for the same work 1, so that the illumination is adjusted and the threshold is changed each time. Therefore, it has become necessary to take measures such as re-registration of the reference image data. In particular, when producing a wide variety (for example, several hundred types) of workpieces 1 on a single production line 1, it is necessary to readjust threshold values and re-register reference image data for all types. It was necessary.

  However, according to the different kind mixing inspection method of this embodiment shown in S100 to 180, only the work 1 (master work) supplied first is used as the reference image data N as it is as the image data S obtained by the binarization process. Although registration (non-defective product determination) is performed, for the subsequent workpiece 1, the image data (S or P) of the workpiece 1 in which the non-defective product (same product type) is determined immediately before the workpiece 1 is set as the reference image data N, and the image of the workpiece 1 Data P is compared.

  Accordingly, since the latest image data (S or P) that is always determined to be non-defective (same product type) is used as the reference image data N, the image data P of the work 1 is compared and determined, so only when the inspection environment changes depending on the production timing. In addition, even when the inspection environment changes during production (for example, the illuminance gradually changes), mixing different types of products without taking measures such as adjusting the illumination, changing the threshold, or re-registering the reference image data N An inspection can be performed. That is, the different product mixture inspection method of the present invention is an inspection method that is more resistant to changes in the inspection environment than the case where the master work image data S is fixedly registered in advance as the reference image data N.

  In addition, since the different kind mixing inspection method of the present embodiment is resistant to changes in the inspection environment such as the illumination illuminance change, as shown in the present embodiment, the change in the inspection environment (the illuminance (detected by the camera 10 ( It is more effective for a configuration in which a predetermined region included in an image is identified (binarization processing) by luminance affected by a change) of reflected light from the work 1 and image data (S or P) is acquired. .

  In the case of the first work 1 (master work), the image data S becomes the reference image data N as it is without being compared, so it is confirmed that it is a non-defective product (the same kind as the kind to be produced). In addition, it is preferable to take an image with the camera 10.

  Although the preferred embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments, and can be implemented with various modifications.

  In the present embodiment, the configuration in which the workpiece 1 is transported on the production line 2 and sequentially supplied to the camera 10 of the image processing apparatus 100 is shown. However, the configuration is not limited to the above as long as the workpiece 1 is sequentially supplied to the camera 10.

  In this embodiment, the CPU 21 of the control unit 20 binarizes each pixel based on a predetermined threshold level (luminance) from an image captured by the camera 10 and identifies a predetermined area. showed that. However, other than the luminance, for example, a configuration in which the predetermined area is identified by a color (RGB value) may be used. If it is the structure which images the workpiece | work 1 with the camera 10 and carries out the predetermined | prescribed process and makes the image data (S or P), since it will receive to the influence of the change of an inspection environment, such as the illumination intensity change of illumination, Applicable.

  In the present embodiment, the CPU 21 binarizes the image of the work 1 to white or black, and sets the number of pixels in a predetermined area (for example, an area determined to be black) extracted by the binarization process as image data. An example of (S or P) is shown. However, as image data (S or P), position (coordinates), average luminance (density), or the like may be applied in addition to the number of pixels. In addition to the black and white binarization processing, gray processing or color processing using RGB data of each pixel may be applied as the luminance processing.

  In this embodiment, when determining a defective product (different product) in S180, the CPU 21 outputs a defective product signal to the line control unit 110, conveys the work 1 from the production line 2 to the defective discharge conveyor, An example of automatically excluding from good products was shown. However, in addition to this, when determining a defective product in S180, the CPU 21 may output a signal for stopping the production line 2 to the line control unit 110, for example. In this case, after the line is stopped, the worker confirms the actual work 1 and manually inputs the data of the pass / fail result (whether the product is the same product type or a different product type), whereby the image data P is newly added to the reference image data. A configuration may be adopted in which whether it is updated as N or the reference image data N is maintained as it is is determined, and the production line 2 resumes conveyance.

It is a block diagram which shows schematic structure of the image processing apparatus for performing the different kind mixing inspection of this invention. It is a flow for demonstrating the different kind mixing inspection method.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Work 10 ... Camera (imaging device)
20 ... Control unit 21 ... Central processing unit (CPU)
22 ... Memory 30 ... Notification unit 100 ... Image processing device 110 ... Line control unit

Claims (3)

A continuously supplied workpiece is sequentially imaged by an imaging device to acquire image data, and based on the image data, a different kind mixing inspection method for inspecting mixing of different kinds of works,
A registration step of registering image data of the master work supplied first as reference image data;
A determination step of determining whether the image data of each workpiece supplied following the master workpiece is compared with the reference image data, and can be regarded as the same product type,
An inspection method for mixing different types of products, comprising: an update step of updating the reference image data by using, as new reference image data, the image data of the workpiece determined by the determining step as the same type.   The method of claim 1, wherein the image data is obtained by identifying a predetermined area included in the image based on brightness of the image captured by the imaging device.   The method of claim 2, wherein the image data is the number of pixels in a predetermined area extracted by binarization processing.

Images