JP2013024852A - Molding image processing inspection device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image inspection device easily and inexpensively performing an inspection of a plurality of parts and an inspection of a foreign object and a chip with a dimension of substantially 0.1 mm.SOLUTION: A hand part 3 of a robot 1 grips a workpiece 4. The hand part 3 has such a structure as having a grip element 14 for reducing repetitive displacement of a gripping state. Further, for reducing the displacement, the workpiece 4 gripped by the robot 1 is held over a laser sensor 11 or a camera 5, and the position of the workpiece 4 is compared with data previously obtained by holding a nondefective item of the workpiece 4 over the laser sensor 11 or the camera 5. On the basis of the data, the robot 1 corrects the position of the workpiece 4 by a coordinate conversion program. The workpiece 4 is held over the camera 5, emits a trigger signal, and captures an image. The image is checked by comparing it with an image of the nondefective item of the workpiece that has been registered previously, and, when it is determined to be the nondefective item, releases the workpiece 4 to a position of a nondefective chute 9 and, when it is determined to be a defective item, releases it at a position of a defective chute 10.

Description

  The present invention relates to a plastic molded product image processing inspection apparatus.

Current inspection equipment 1
The camera is set up on the conveyor, and when the workpiece moving on the conveyor comes to a certain position, it is picked up by a trigger signal to the camera. Device that performs discrimination

Current inspection equipment 2
Position the work with a positioning cylinder, etc. when the work placed on the conveyor and moving on the conveyor comes to a certain position, and take a picture with the trigger signal to the camera after there is no position shift. , A device for determining the quality by comparing the image with a good image stored in advance

Current inspection equipment 3
When the camera is placed on the conveyor and the workpiece moving on the conveyor reaches a certain position, a trigger signal is sent to the camera, the workpiece is rotated, continuous pictures are taken, and the image is stored in advance. A device that determines pass / fail by comparing with a non-defective image.

Difficulty of setting,
In order to determine the quality of a workpiece, it is necessary to take multiple images at various angles. For this purpose, the position of the installation of multiple cameras must be devised. In addition, each lighting device suitable for detection of defective products by each camera is required, and advanced knowledge is required for selecting the type of lighting device. The algorithm for selecting the camera lens and detecting the defective part must be set. As for the algorithm, the image inspection equipment manufacturer is researching software suitable for the type of defective part. If a part of the algorithm type is placed, it cannot be counted as position correction, threshold setting, expansion, contraction. If you put a part of the type of lighting, you can count dome lighting, bar lighting, LED lighting, red, blue and white.
Which kind of algorithm to use in this setting requires advanced knowledge and experience.
In addition, combination conditions with lighting are added, making it more difficult to set.
Since it is impossible to detect defective products due to the camera position, the influence of ambient light, and the change of the lighting position at the site, it is necessary to re-set the settings and there is no hard work at the site.

Photo angle It is necessary to take pictures at various angles depending on the type and location of the defect in order to determine the quality of the workpiece. From the top, from the side, from the diagonal, from the bottom, it is necessary to devise the position to install multiple cameras.

Pass / fail judgment accuracy When some of the types of defects are placed, they are foreign matter, uneven color, blurred printing, missing characters, missing workpieces, welds, bubbles, and the like. As a high-accuracy image inspection means, when it is desired to detect a foreign object size of 0.1 mm or more as a defective product, accurate positioning is performed so that the positional relationship between the camera, illumination, and workpiece is 0.1 mm or less. A device is required. However, when it is desired to inspect from various angles, a large number of positioning devices are required, resulting in a complicated problem.
There is an image position correction function in the camera image processing software even if the position is shifted by 0.1 mm or more, and it is common to find the features of the image that has already been stored and automatically correct the position shift for comparison. However, the accuracy of image inspection is lower than the method of positioning accurately.
The cause is that even if the image is corrected by software by the camera position correction, the positional relationship between the work, illumination, and camera is shifted, so the shadowing, the light reflection, and the distortion of the work shape change. The accuracy drops.

  In order to solve the above problems, the present invention has a hand part gripper 14 with high repeated positioning accuracy at the tip of the hand part 3 of the articulated robot 1, and grips the workpiece 4 with the hand part gripper 14. The workpiece 4 gripped by the hand gripper 14 is measured with a position / distance measuring device such as a laser 11 or a camera 5 to measure the gripping position error of the workpiece 4, and the coordinates of the multi-joint robot 1 are determined based on the error data. By correcting the position using the conversion program and correcting the gripping position, the relative positions of the camera 5, the illumination 6, and the workpiece 4 are made the same as the positional relationship stored in advance, and the workpiece 4 is subjected to image inspection. When the camera 5 is held over, the image inspection accuracy can be improved by making the illumination angle hitting the workpiece 4 the same, and the accuracy of the workpiece can be improved more quickly and accurately than the inspection by the human eye. There is provided an image inspection apparatus characterized by capable of a 査.

  The robot 1 grips the workpiece 4 and picks up and picks up a plurality of surfaces of the workpiece from various angles freely with the teaching of an articulated robot, so that the workpiece can be detected more quickly and accurately than human eye inspection. Can be inspected.

  Since the robot 1 holds the workpiece 4 and holds it in front of the camera, the angle between the workpiece 4 and the camera 5 can be freely set and teaching can be performed while viewing the image, so that fine adjustment is facilitated.

  When the robot 1 grips the work 4 and a grab error occurs and the image is corrected and inspected by the position correction of the image processing software, the positional relationship between the camera 5 and the work 4 can be corrected. Since the positional relationship of is changed, the way of lighting changes and the inspection result becomes unstable and the inspection accuracy decreases. Therefore, when the hand grip 14 of the robot 1 is repeatedly shaped with high positioning accuracy, the positional relationship of the camera 5, the illumination 6, and the workpiece 4 is the same, so that the position correction of the image processing software is not relied upon. The accuracy is improved because the inspection can be performed.

  The illumination type and illuminance can be automatically switched to the most suitable selection by issuing a signal to the illumination device according to the defective part detection type of the work 4.

  By exchanging the hand grip 14 for gripping the workpiece 4, creating a robot program for each type of workpiece and teaching the robot position, it is possible to inspect various types of workpieces.

  If a gripping error occurs when the hand gripper 14 of the item 0011 is repeatedly shaped with high positioning accuracy, the laser sensor 11 incorporated in the inspection apparatus By measuring the linear and rotational misalignment and using the data to correct the position using the robot's coordinate conversion program to correct the gripping position, the position is held at the same position as the stored work, camera and illumination. Therefore, unlike image software position correction, it is possible to increase image inspection accuracy.

  The image is captured by the camera 5 of the image inspection apparatus instead of the laser sensor 11 and the position data of the work 4 is taken in, and the position is corrected using the coordinate conversion program of the robot and the gripping position is corrected using the data. It can be held over the same position as the work, camera and lighting. As described above, the position / distance measuring apparatus may be a laser sensor or an image inspection apparatus, but there is no problem with an apparatus and method more suitable for the workpiece.

  The high-speed parallel link robot of 6-axis multi-joint robot can be used, so it is possible to inspect at high speed, and the inspection of the molded product is completed within the molding cycle of the injection molding machine. I can go down.

BEST MODE FOR CARRYING OUT THE INVENTION

  An embodiment of the present invention is shown in FIGS. The combination of two cameras 5, one laser sensor 11, and one temporary placement table 13 is the best mode of the shortest time inspection for performing six-surface inspection of the workpiece 4.

  1, 2, and 3, the reciprocating unit 7 is installed in the frame unit 12, the receiving unit 8 is installed on the reciprocating unit 7, A jig for receiving the workpiece 4 is attached. The work 4 taken out from the injection molding machine by the take-out machine is opened to the receiving part 8.

  The cradle unit 8 is moved to the bottom of the robot 1 by the reciprocating unit 7, and the robot 1 sequentially grips and takes out the workpiece 4, and sequentially inspects the surface to be inspected of the gripped workpiece on the camera 5.

  In the inspection method, it is necessary to store the image in advance by sequentially holding the front or oblique surface to be inspected of a non-defective work as a master over the camera 5. Next, the work to be inspected is successively held over the camera 5 in the same manner, and quality determination is made based on the difference between the non-defective product storage work image and the inspection work image.

  When a non-defective signal is output by the good / bad determination signal, the robot 1 moves onto the non-defective shooter 9 to release the workpiece being gripped. When a defective product signal is output, the workpiece moved onto the defective product shooter 10 is released.

  When it is desired to inspect the entire surface of the work 4, since the work 4 is gripped by the hand grip 14 of the robot 1, the gripping location cannot be inspected by the camera 5. Once the workpiece 4 is released, another surface is gripped and the place where it has been gripped is held over the camera 5 to inspect the entire surface.

  4, 5, and 6 show an example in which six surfaces of the work 4 are sequentially inspected. The workpiece 4 is gripped by the hand gripping element 14. Since the pair of cameras 5 and the illumination 6 are opposed to each other, it is possible to inspect both the front and back surfaces of the workpiece 4 simultaneously.

  In the inspection of FIG. 6, since the camera 5 can see the places other than the place where it is gripped, it can be inspected.

  In this case, if the material of the gripping part 14 of the hand part is made of a transparent acrylic resin or the like, the workpiece 4 can be seen through the acrylic resin or the like, so that inspection becomes possible, and the cycle of inspection time is as much as it is not temporarily placed. Can be shortened.

  7A and 7B are state diagrams in which the work 4 is gripped by the hand gripping element 14. The hand portion gripping element 14 is processed so as to fit the uneven shape of the place where the workpiece is gripped. As a result, the repeated positioning accuracy when gripped becomes stable. 8A and 8B are diagrams showing a state in which the hand unit grip 14 is open.

  FIG. 9 is a plan view of one embodiment of the grip cylinder part 2 and the hand part 3. FIG. 10 is a front view of an embodiment of the grip cylinder part 2 and the hand part 3. FIG. 11 is a side view of an embodiment of the grip cylinder part 2 and the hand part 3. The hand portion gripper 14 shown in FIG. 11 has a structure that can be easily replaced. If the hand grip 14 is made of a transparent acrylic resin or the like, the acrylic resin can pass through and the work 4 can be seen, so that it can be inspected at the gripping place.

  The accuracy of image inspection increases as the gripping position of the gripped work is brought closer to the gripping position registered in advance by gripping a good master work in advance and holding it over the camera 5. However, when the position of the workpiece in the actually gripped state is slightly shifted, the angle of illumination changes with respect to the workpiece, so that the image changes and the inspection result becomes defective. Even if the image can be inspected by correcting the position of the camera and the image with the position correction software of the image inspection apparatus, the relative position between the camera and the workpiece is actually shifted, so that the influence of the illumination position cannot be eliminated.

FIGS. 12 to 16 are diagrams illustrating an embodiment for explaining a method of correcting the position of the work 4 held by the laser sensor 11. In FIG. 12, the workpiece 4 is moved by the robot 1 within the measurement range of the laser 11 and the laser value at this position is measured. Let this position be the Z1 point. Next, as shown in FIG. 13, the workpiece 4 is moved in the negative Z direction to measure the laser value. Let this position be the Z2 point. A distance between the Z1 point and the Z2 point is represented by A in FIG.
Let B be the difference in laser value between the Z1 point and the Z2 point. From this data, if the inclination of the workpiece 4 is θ, TANθ = B / A.

  14, after gripping a good master work in advance and correcting the tilt difference angle with the data measured by the laser in the same manner as in the item 0029 using the tool coordinates of the robot, and then converting it to user coordinates to correct the tilt. The work 4 is rotated 90 degrees so that it is horizontal. Then, the distance between the top surface of the workpiece 4 and the laser 11 is measured. This position is assumed to be X3 point.

  A good master work is grasped in advance and the Z-direction top surface distance difference C between the data measured by the laser and the X3 point in the same manner as in the item 0030 is corrected with the tool coordinates of the robot. Rotate 90 degrees to make it vertical. Then, the distance between the vertical surface of the workpiece 4 and the laser 11 is measured again at the Z1 point.

  A good master work is grasped in advance, and the X-direction distance difference D between the data measured by the laser and the Z1 point is corrected again with the tool coordinates of the robot in the same manner as in item 0031. Thus, the inclination θ is the correction of the robot user coordinates, and the Z direction and the X direction are corrections of the robot tool coordinates. The corrected position is the same as the position where a good master work is previously gripped, and is shown in FIG.

  In the subsequent operation, the position where the workpiece 4 is sequentially held over the camera 5 is determined by the user coordinate correction of the robot and the tool coordinate correction. Can be the same as the position.

  As a high-accuracy molded product image processing inspection apparatus based on an image, the position / distance of a laser, an image apparatus, or the like that can correct the relative position of the workpiece 4, the camera 5, and the illumination 6 by a robot, rather than only the inspection method based on the position correction of the camera. The measuring device is more suitable for high-precision molded product inspection using images.

It is a front view which shows one Example of this invention. It is a top view which shows one Example of this invention. It is a side view which shows one Example of this invention. It is a 1st, 3rd surface imaging position form figure of a workpiece | work. It is a 2 and 4 surface imaging position form figure of a workpiece | work. It is a 5th, 6th surface imaging position form figure of a work. It is a form top view which shows one Example of the hand part gripper closing of this invention. It is a form front view which shows one Example of the hand part gripper closing of this invention. It is a form top view which shows one Example of the hand part holding element opening of this invention. It is a form front view which shows one Example of the hand part holding element opening of this invention. It is a top view which shows one Example of the holding | grip cylinder part of this invention. It is a front view which shows one Example of the holding | grip cylinder part of this invention. It is a side view which shows one Example of the holding | grip cylinder part of this invention. Laser Z1 point measurement chart Laser Z2 point measurement chart Laser X3 point measurement chart Laser again Z1 point measurement diagram Work position diagram after coordinate correction

DESCRIPTION OF SYMBOLS 1 Articulated parallel link robot 2 Gripping cylinder part 3 Hand part 4 Work 5 Camera part 6 Illumination part 7 Reciprocating movement part 8 Placement part 9 Good product shooter 10 Defective product shooter 11 Laser sensor 12 Frame part 13 Temporary placement part 14 Hand part Gripper

Claims (2)

  Hold the gripper with high positioning accuracy repeatedly at the tip of the hand of the articulated robot, grip the workpiece with the gripper, and set the relative position between the camera, illumination, and the workpiece to the same gripping position as the positional relationship stored in advance in the image. Image processing aiming to inspect the workpiece faster and more accurately than the inspection by the human eye by picking up multiple images of the workpiece over the image inspection camera freely from various angles according to the teaching of the articulated robot Inspection device.   The image inspection apparatus according to claim 1, wherein the workpiece gripped by the gripper is measured by a position / distance measuring device such as a laser or an image device to measure a gripping position error with respect to a positional relationship in which the workpiece is stored in advance. By correcting the position using the coordinate conversion program of the articulated robot based on the error data and correcting the gripping position, the relative position between the camera, the illumination, and the workpiece is set to the same gripping position as the image stored in advance, An image processing inspection apparatus for increasing an image inspection accuracy when an illumination angle hitting the workpiece becomes the same when the workpiece is held over an image inspection camera.

Images