JP2011164061A - Transparent object detection system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a system simply and precisely detecting a transparent object. <P>SOLUTION: The transparent object detection system includes: a camera 12 picking up a first region including a transparent object 114 having characteristics of changing a polarization direction of transmitted light to pick up a vertical polarization image and a horizontal polarization image; a mount base 113 for mounting the transparent object 114; a polarization filter 112 provided at a position including a range of the second region picked up so as to include at least the transparent object 114 from among the first region across the mount base 113 from the camera 12; and an image processing device 13 detecting the transparent object on the basis of a distribution of vertical/lateral polarization degree of a vertical/lateral polarization degree image based on the vertical polarization image and the horizontal polarization image. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a transparent body detection system. More specifically, the present invention relates to a transparent body detection system that can easily and accurately detect a transparent body by image processing.

  Transparent plastic molded products (hereinafter referred to as transparent bodies) are used in various fields such as food packaging means, home appliances, and automobile parts. However, when a transparent object is an object to be imaged, the transparent object has low reflectance and high transmittance and cannot be contrasted with the background. Therefore, the transparent object is detected using a camera that uses only normal luminance information. Was difficult.

  For such a problem, as shown in FIG. 16, Patent Document 1 discloses a light projecting method in which light is polarized in a specific direction by a light projection side polarizing plate 95 and then projected onto a predetermined region including the transparent body 90. An apparatus 91, a polarizing device 92 having a light receiving side polarizing plate 96 that transmits only specific polarized light out of the light transmitted through the predetermined region, and an image obtained by receiving the light transmitted through the light receiving side polarizing plate 96, A television camera 93 that captures a two-dimensional image, and an image processing device 94 that detects a transparent body 90 using a brightness / darkness distribution in the image obtained from the television camera 93, and that is configured to emit light polarized in a specific direction. Then, light is projected onto a predetermined area including the transparent body 90, and the light transmitted through the predetermined area is transmitted through a light-receiving side polarizing plate 96 that transmits only specific polarized light, and then is transmitted to the CCD in the TV camera 93. Received light and obtained from this CCD Using the brightness distribution in the image, the transparent object detection method for detecting a transparent body 90 is disclosed.

(Problem 1: Arrangement of lighting device)
In the technique described in Patent Document 1, it is necessary to dispose the illumination device 97 and the light projecting side polarizing plate 96 below the transparent body 90 with respect to the television camera 93. However, in an actual line or the like, there is a process of washing the transparent body, and water drops may fall on the lower side. Therefore, it is difficult to arrange the lighting device 97 below the transparent body in consideration of safety. There is a case.

(Problem 2: Relocation from tray to detector)
In addition, the technique described in Patent Document 1 has a problem in that when the process is performed integrally with the tray or the mounting table and the tray, it is necessary to mount the process once on the transparent body detection device.

(Problem 3: Necessity of rotation mechanism and calibration)
Furthermore, the technique described in Patent Document 1 requires that the transmission polarization directions of the light-projecting side polarizing plate 95 and the light-receiving side polarizing plate 96 coincide with each other or be orthogonal to each other. It was necessary to rotate and measure based on each image. When this is performed according to a change in the type of the transparent body 90, etc., this leads to a time loss on the process line. In addition, providing such a rotation mechanism is a factor in increasing the size of the apparatus. Furthermore, such a rotating mechanism is prone to change over time (center position deviation, etc.) of the mechanism itself with vibration as in a factory, requiring maintenance such as failure due to wear of moving parts. It was.

  Therefore, the present invention relates to a transparent body detection system that enables simple and highly accurate detection of a transparent body, and more specifically, without detecting an illumination device on the lower side and regardless of the mounting state of the transparent body. It is another object of the present invention to provide a transparent body detection system that does not involve calibration work or a rotation mechanism.

  In order to achieve such an object, the transparent body detection system according to claim 1 captures a first region including a transparent body having a property of changing a polarization direction of transmitted light, and obtains a vertically polarized image and a horizontally polarized image. At a position including the range of the second area to be imaged including at least the transparent body among the first area, with the image capturing means for imaging, the mounting table for mounting the transparent body, and the mounting table from the image imaging means. An installed polarizing filter and an image processing device that detects a transparent body based on a distribution of vertical and horizontal polarization degrees of a vertical and horizontal polarization degree image based on a vertical polarization image and a horizontal polarization image.

  Therefore, since the vertical and horizontal polarization degree images can be taken in, it becomes possible to detect the polarization change of the transparent body without depending on the direction of the polarizing filter on the transparent base, and the calibration work becomes unnecessary. In addition, a rotating mechanism is not required (Problem 3). Moreover, the light which permeate | transmits a transparent body should just be the light (light biased in the specific polarization direction) which permeate | transmitted the polarizing filter, and it is not necessary to arrange | position an illuminating device directly under a polarizing filter (issue 1). Furthermore, even if a mounting table is installed on the tray, it is sufficient if there is a polarizing filter between the tray and the mounting table, and an operation for replacing the mounting table with the transparent body detection device is not necessary ( Problem 2).

  In addition, the transparent body detection system according to claim 2 is an image capturing unit that captures a first region including a transparent body having a characteristic of changing a polarization direction of transmitted light and captures a vertically polarized image and a horizontally polarized image. And a mounting table that reflects light having a polarization direction substantially the same as the polarization direction of the third area of the first area that is photographed without including the transparent body. The distribution of the vertical and horizontal polarization degree of the vertical and horizontal polarization degree image based on the vertical polarization image and the horizontal polarization image, which is provided at a position across the pedestal and where the gradation value of one of the vertical polarization image or the horizontal polarization image is stronger than the other. And an image processing device that detects a transparent body based on the image processing apparatus.

  Therefore, since the vertical and horizontal polarization degree images can be taken in, it becomes possible to detect the polarization change of the transparent body without depending on the direction of the polarizing filter on the transparent base, and the calibration work becomes unnecessary. In addition, a rotating mechanism is not required (Problem 3). In addition, since no polarizing filter is disposed, the system is simplified and the cost is reduced.

  According to a third aspect of the present invention, in the transparent body detection system according to the first or second aspect, the image processing device detects the position and orientation of the detected transparent body.

  According to a fourth aspect of the present invention, in the transparent body detection system according to the third aspect, the robot hand includes a robot hand and a robot controller that controls the robot hand, and the robot controller detects the robot hand by detecting the transparent body. The transparent body is picked up based on the detected posture of the transparent body.

  According to a fifth aspect of the present invention, in the transparent body detection system according to any one of the third or fourth aspect, the image processing device determines the presence area of the transparent body based on the detected position and orientation of the transparent body. It determines and compares the existence area | region of two or more transparent bodies, and detects the geometrical defect of a transparent body.

  According to a sixth aspect of the present invention, in the transparent body detection system according to any one of the third to fifth aspects, the image processing apparatus is configured to detect the presence area of the transparent body based on the detected position and orientation of the transparent body. And the appearance defect of the transparent body is detected based on the luminance distribution of the monochrome luminance image based on the vertical polarization image and the horizontal polarization image.

  According to a seventh aspect of the present invention, in the transparent body detection system according to any one of the third to sixth aspects, the image processing apparatus is configured to detect the presence area of the transparent body based on the detected position and orientation of the transparent body. And character and / or symbol information written on the surface of the transparent body is detected based on the luminance distribution of the monochrome luminance image based on the vertical polarization image and the horizontal polarization image.

  The invention according to claim 8 is the transparent body detection system according to any one of claims 3 to 7, wherein the transparent body is a transparent foreign substance that can be mixed into an empty bottle or a bottle containing a transparent liquid. The image processing apparatus determines the presence area of the transparent body based on the detected position and orientation of the transparent body, and the distribution of the vertical and horizontal polarization degrees of the vertical and horizontal polarization degree images based on the vertical polarization image and the horizontal polarization image. Based on the above, it is determined whether or not a transparent foreign matter is mixed in an empty bottle or a bottle containing a transparent liquid.

  According to the present invention, it is possible to easily and accurately detect a transparent body.

It is a schematic block diagram of one Embodiment of the transparent body detection system which concerns on this invention. It is a structural example (1) of an image imaging means. It is a structural example (2) of an image imaging means. It is a structural example (3) of an image imaging means. It is a structural example (4) of an image imaging means. It is an example of a region division type filter. It is an example of the imaging | photography result of a transparent body, Comprising: It is (a) monochrome luminance image and (b) longitudinal and transverse polarization degree image. It is another example of the imaging | photography result of a transparent body, Comprising: (a) Monochrome luminance image, (b) Vertical / horizontal polarization degree image. It is another example of the imaging | photography result of a transparent body, Comprising: (a) Monochrome luminance image, (b) Vertical / horizontal polarization degree image. It is another example of the imaging | photography result of a transparent body, Comprising: (a) Monochrome luminance image, (b) Vertical / horizontal polarization degree image. It is an example of the flowchart of the image process which an image processing apparatus performs. It is an example of an image when a transparent body has a chip. It is an example of an image when a transparent body has a flash. It is an example of an image when there is no dirt in a transparent body. It is an example of an image when there is dirt on a transparent body. It is an example of the conventional transparent body detection system.

  Hereinafter, a configuration according to the present invention will be described in detail based on the embodiment shown in FIGS.

(Transparent object detection system)
The transparent body detection system 1 according to the present embodiment captures a first region including a transparent body 114 having a characteristic that the polarization direction of transmitted light changes, and captures a vertically polarized image and a horizontally polarized image. Including a range of a second area to be photographed including at least the transparent body 114 in the first area, with the camera 12), the mounting table 113 on which the transparent body 114 is mounted, and the mounting table 113 from the image capturing means. A polarizing filter 112 installed at a position, and an image processing device 13 that detects a transparent body based on a distribution of vertical and horizontal polarization degrees of a vertical and horizontal polarization degree image based on a vertical polarization image and a horizontal polarization image.

  FIG. 1 is a schematic configuration diagram of a transparent body detection system 1. The transparent body detection system 1 includes a transparent body placement unit 11, a camera 12 capable of capturing monochrome luminance images and vertical and horizontal polarization degree images, an image processing device 13, a monitor 14, a robot controller 15, and a robot hand 16. It is configured.

  For example, the transparent body detection system 1 is installed indoors in a factory or the like, and is placed under room illumination such as a fluorescent lamp 101. Further, the mounting table 113 of the transparent body mounting unit 11 and the ground are arranged in parallel, and the camera 12 is arranged to image the transparent body 114 from the normal direction of the mounting table 113. Further, a light shielding plate 102 is disposed between the camera 12 and the room lighting (fluorescent lamp 101) so that the room lighting on the transparent body 114 is not directly irradiated. However, partial light shielding is sufficient for the fluorescent lamp 101 directly above. For example, a colored plastic plate may be used as the light shielding plate 102, but the light shielding plate 102 is not limited to this as long as it has a light shielding property. By disposing the light shielding plate 102, unnecessary reflected light on the surface of the transparent body 114 is removed, and the detection accuracy of the transparent body 114 is improved.

  The transparent body placement unit 11 is an area (second area) in which at least the transparent body 114 is photographed among the placement stage 113 on which the transparent body 114 is placed and the other side of the placement stage 113 when viewed from the camera 12 side. A polarizing filter 112 is disposed so as to include the range. A tray 111 is disposed below the polarizing filter 112. The tray 111 has a role of receiving soot from the transparent body 114 that has undergone a cleaning process.

  Further, the mounting table 113 has a mesh structure, and has a mesh occupation ratio that allows the light reflected by the tray 111 to sufficiently pass through the transparent body 114 after passing through the polarizing filter 112. . Specifically, an occupation ratio of about 5% of the transparent body region is desirable. In addition, it is desirable that each mesh is thin, and it is desirable that the mesh be 5% or less with respect to the size of the transparent body 114. By setting it as such a size, since the area | region of the transparent body 114 through which the polarized light passes increases, the background area | region also increases, Therefore The edge detection accuracy of the transparent body 114 improves.

  Moreover, as the polarizing filter 112, for example, an organic film-shaped one is preferably used. Since the organic film has a large area and is inexpensive, the cost of the transparent body detection system 1 can be reduced. Further, since it has elasticity, it can be attached to the tray 111 with a sealing material or the like even if it has a gentle uneven structure.

  In the present embodiment, an example in which the tray 111 is disposed has been described. However, it is sufficient that the polarizing filter 112 is below the mounting table 113, and the lower side of the polarizing filter 112 may be, for example, a floor. The polarizing filter 112 is intended to irradiate the transparent body with light of a specific polarization, and is not particularly limited as long as it fulfills its function. For example, a reflective polarizing filter may be used instead of the transmissive polarizing filter. As the reflective polarizing filter, for example, a resinous wire grid type is known.

  Further, in the present embodiment, a hemispherical transparent body 114 made by injection molding of polycarbonate resin will be described as an example. There are transparent bodies 114 that do not change the polarization characteristics, such as glass, and some resins that do or do not change the polarization characteristics depending on how they are made. Polycarbonate resins have the characteristics that the polarization direction changes due to injection molding or the like. Therefore, this is used.

  The camera 12 is a camera that can capture a luminance monochrome image and a vertical / horizontal polarization degree image. Each pixel signal forming each two-dimensional image is output from the sensor side in the camera 12 to the image processing device 13.

  The image processing apparatus 13 converts each pixel signal from the camera 12 into A / D conversion by binarizing or multileveling the A / D converter 131, and digitally converting the A / D converter 131 into 2 Memory 132 for storing data that has been digitized or multi-valued, a display control unit 133 for display control of the monitor 14, a communication control unit (not shown) that performs communication control of control signals to the robot controller 15, a CPU 135, etc. It is comprised by. The CPU 135 performs various image processing and calculations according to a program stored in a ROM (not shown) or the like.

  Further, the image processing device 13 detects the position, posture, size (existing region), and the like of the transparent body 114 using the light / dark distribution in the image obtained from the camera 12, that is, the image of the data stored in the memory 132. Perform the process.

  In addition, a process for displaying the result such as the position and posture of the obtained transparent body 114 on the monitor 14 through the display control unit 133 is performed. Further, a process of transmitting a control signal obtained from the result of the position and posture of the transparent body 114 to the robot controller 15 is performed.

  The monitor 14 is a display device such as an LCD or a CRT, and the robot controller 15 performs drive control of the robot hand 16 in accordance with a control signal from the image processing device 13. The robot hand 16 picks up the transparent body 114 mounted on the upper surface of the mounting table 113 according to the drive control of the robot controller 15 and stores it in a storage unit such as a tray (not shown).

(Image capturing means)
Next, a configuration of a camera as an image capturing unit that captures a vertical / horizontal polarization degree image will be described. The camera 12 acquires an ambient image including a transparent body 114 having pixels of, for example, a megapixel size by using an image sensor such as a CCD (charge-coupled device) or a CMOS (complementary metal oxide semiconductor) as a light receiving element. The camera 12 in the transparent body detection system 1 according to the present embodiment can acquire a polarization image in addition to the luminance image. The camera 12 for forming this polarized image will be described below.

<First Embodiment of Image Imaging Unit>
As shown in FIG. 2, the camera 12 has a polarizing filter 20a disposed at a position where the vertically polarized light is transmitted and takes a vertically polarized image, and a polarizing filter 20b disposed at a position where the horizontally polarized light is transmitted. And a camera for taking a horizontally polarized image. According to this camera 12, there is no time lag for rotating the polarization filter as in the conventional camera, and a vertically polarized image and a horizontally polarized image can be taken simultaneously.

  In addition, unlike the stereo system, two cameras can be arranged close to each other. In recent years, there is a high need for miniaturization of equipment, so that the camera 12 shown in FIG. 2 is further reduced in size by using a single light receiving element via a lens array and a polarizer filter array as shown in FIG. It is also preferable to use a camera 12 having a configuration for imaging.

  FIG. 3 shows a schematic configuration diagram of the optical system 21 of the camera 12. The optical system 21 has two optical systems for capturing a vertically polarized image and a horizontally polarized image. A lens array 22, a light shielding spacer 23, a polarizing filter 24, a spacer 25, and a solid-state imaging unit 26 are laminated. Is formed.

  The lens array 22 has two lenses 28a and 28b. The two lenses 28a and 28b are formed of a single lens made of, for example, an aspheric lens having the same shape independent from each other, and are arranged on the same plane with the optical axes 27a and 27b being parallel. Here, assuming that the direction parallel to the optical axes 27a and 27b of the lenses 28a and 28b is the Z axis, the direction perpendicular to the Z axis is the X axis, and the direction perpendicular to the Z axis and the X axis is the Y axis, the lenses 28a and 28b. Are arranged on the same XY plane.

  The light-shielding spacer 23 has two openings 29a and 29b, and is provided on the side opposite to the subject side with respect to the lens array 22. The two openings 29a and 29b are penetrated with a predetermined size centering on the optical axes 27a and 27b, respectively, and the inner wall surface is subjected to antireflection treatment of light by blacking, roughing or matting.

  The polarizing filter 24 includes two polarizer regions 30a and 30b having different polarization planes by 90 degrees, and is provided on the side opposite to the lens array 22 with respect to the light shielding spacer 23. The two polarizer regions 30a and 30b are provided in parallel with the XY plane around the optical axes 27a and 27b, respectively. The polarizer regions 30a and 30b convert non-polarized light whose electromagnetic field vibrates in an unspecified direction into linearly polarized light by transmitting only vibration components in the direction along the polarization plane.

  The spacer 25 is formed in a rectangular frame shape having an opening 31 through which regions corresponding to the polarizer regions 30 a and 30 b of the polarizing filter 24 pass, and is provided on the opposite side of the light shielding space 23 with respect to the polarizing filter 24. Yes.

  The solid-state image pickup unit 26 has two solid-state image pickup elements 33 a and 33 b mounted on a substrate 32 having a signal processing unit, and is provided on the opposite side of the polarization filter 24 with respect to the spacer 25. The imaging areas where the subject images are actually formed by the two solid-state imaging devices 33a and 33b are provided on the same plane parallel to the XY plane with the optical axes 27a and 27b as the centers. The solid-state imaging devices 33a and 33b do not have a color filter when performing monochrome sensing, and may be disposed at the previous stage when sensing a color image.

<Second Embodiment of Image Imaging Unit>
In addition, as shown in FIG. 4, an image is picked up by one imaging lens (a plurality of lenses may be coaxially arranged), and a vertical polarization image and a horizontal polarization image are separated at the rear stage of the lens, and the degree of vertical and horizontal polarization from each image. It is also preferable to use a camera 12 having a configuration for forming an image.

  This camera 12, as shown in FIG. 4, captures a vertically polarized image and a horizontally polarized image, as shown in FIG. 4, a half mirror box 34 having 1: 1 transparency, a mirror 35, a vertical polarization filter 36, and a horizontal A CCD 38 that captures a field image through a polarizing filter 37, a vertical polarization filter 36, and a CCD 39 that captures a field image through a horizontal polarization filter 37 are provided. In the camera 12 of the first embodiment, it is possible to simultaneously capture a vertically polarized image and a horizontally polarized image, but parallax occurs. On the other hand, since the camera 12 according to the second embodiment captures images through the same imaging optical system (lens), no parallax occurs. Therefore, the detection area can be small, and processing such as parallax deviation correction is not necessary.

  Instead of the half mirror, a polarization beam splitter that is a prism that reflects laterally polarized light and transmits longitudinally polarized light may be used. When the polarizing beam splitter is used, the vertical polarizing filter 36 and the horizontal polarizing filter 37 are not necessary, so that the optical system can be simplified and the light utilization efficiency can be improved.

<Third embodiment of image pickup means>
In addition, as shown in FIG. 5, a single imaging lens (a plurality of lenses may be coaxially arranged) is used, and a polarizer region that passes only vertically polarized light and a polarizer region that allows only horizontally polarized light to pass through the rear of the lens. It is also preferable to use the camera 12 having a configuration in which the region-divided filter 24 is disposed. Here, as the polarizer region, a region-divided polarizer filter with a clear boundary can be formed by a wire grid method formed with a metal fine concavo-convex shape or an auto-cloning photonic crystal method.

  In the camera 12 of the second embodiment, the optical system is enlarged in order to split the vertically polarized image and the horizontally polarized image into prisms, and two light receiving elements are required. On the other hand, the camera 12 of the third embodiment can acquire a laterally polarized image and a longitudinally polarized image with an optical system coaxial with the imaging lens.

  Here, the configuration of the area division type filter is not limited to the one corresponding to the pixels of the light receiving element, and for example, a configuration as shown in FIG. 6 is also preferable. In FIG. 6, squares arranged in the vertical and horizontal directions indicate the light receiving portions (light receiving element array 40) of the respective light receiving elements, and two types of diagonal bands indicate two types of polarizing filter regions 41 and 42 in the vertical and horizontal directions. Each of the filter regions 41 and 42 has a width of one pixel, that is, a width of one light receiving element in the horizontal direction, and the inclination of the boundary line of the region is 2, that is, two pixels in the vertical direction while proceeding by one pixel in the horizontal direction. It has the shape of an oblique band with a changing angle.

  By combining such a special filter arrangement pattern and signal processing, it is possible to reproduce each filter transmission image on the entire screen even if the alignment accuracy when joining the image sensor array and the area division filter is not sufficient Such an image pickup apparatus can be realized at low cost.

  As described above, the camera 12 in the transparent body detection system 1 according to the present embodiment can acquire a polarization image. The camera 12 preferably acquires a transparent image in real time, and the acquired transparent image is supplied to the image processing device 13.

(Transparent object detection)
The camera 12 in the transparent body detection system 1 according to the present embodiment particularly uses birefringence generated in a plastic transparent body. A film or substrate may be birefringent due to strain or stress, or may exhibit optical anisotropy due to slight structural anisotropy, resulting in a phase difference. When polarized light in a specific direction is incident on such a characteristic, the polarization direction of the transmitted light changes. The camera 12 detects this change.

[When polarization transmission direction of polarizing filter is P direction]
As described above, the camera 12 can shoot a vertical / horizontal polarization degree image, and has gradation in the polarization direction from the horizontal component (S component) to the vertical component (P component) (for example, 256 gradations). ). When S component light is incident, it is displayed as a bright portion, and when P component light is incident, it is displayed as a dark portion.

  FIG. 7 shows an example of the result of photographing by the camera 12 when the polarizing filter 112 is arranged so that the transmission direction is the P component. 7A shows an example of a monochrome luminance image, and FIG. 7B shows an example of a vertical and horizontal polarization degree image.

  As shown in FIG. 7B, the vertical and horizontal polarization degrees of the background portion that does not pass through the transparent body 114 are dark portions. On the other hand, the degree of vertical and horizontal polarization in the region that has passed through the transparent body 114 changes in the polarization direction of transmitted light due to birefringence, and in this case, it appears as a bright portion. On the other hand, as shown in FIG. 7A, in the monochrome luminance image, only the edge portion of the transparent body 114 is slightly white, and it is difficult to separate from the mounting table 113 in image processing. As described above, when the vertical and horizontal polarization degree images are used, the contrast between the transparent body region and other portions can be clearly obtained, and therefore the transparent body region can be extracted by the image processing flow described below.

[When polarization transmission direction of polarizing filter is S direction]
FIG. 8 shows an example of the result of photographing by the camera 12 when the polarizing filter 112 is arranged so that the transmission direction is the S component. 8A is an example of a monochrome luminance image, and FIG. 8B is an example of a vertical and horizontal polarization degree image.

  As shown in FIG. 8B, the vertical and horizontal polarization degrees of the background portion that does not pass through the transparent body 114 are bright portions. On the other hand, the degree of vertical and horizontal polarization of the region that has passed through the transparent body 114 changes in the polarization direction of transmitted light due to birefringence, and in this case, it slightly shifts to the dark part. On the other hand, as shown in FIG. 8A, in the monochrome luminance image, only the edge portion of the transparent body 114 is slightly white, and it is difficult to separate from the mounting table 113 in the image processing. As described above, when the vertical and horizontal polarization degree images are used, the contrast between the transparent body region and other portions can be clearly obtained, and therefore the transparent body region can be extracted by the image processing flow described below.

  As described above, by using the vertical and horizontal polarization degree images, the portion corresponding to the transparent region existing area on the image formed by each pixel value of the memory obtained by imaging becomes a bright portion or a dark portion as a background portion. Therefore, the position, posture and size of the transparent body can be detected.

[Combination with region extraction by luminance image]
Further, detection accuracy can be improved by performing region extraction using a luminance image in addition to region extraction using a vertical / horizontal polarization degree image as described above. In this case, in order to improve the region extraction accuracy in the luminance image, for example, a light source may be disposed between the mounting table 113 and the polarizing filter 112, and the light source may irradiate the polarizing filter. That is, with such an arrangement, the net portion of the mounting table 113 becomes a shadow of light and becomes a dark part, so that it is easy to separate from the transparent body.

  An example of the result of photographing by the camera 12 in this case is shown in FIG. 9A shows an example of a monochrome luminance image, and FIG. 9B shows an example of a vertical and horizontal polarization degree image.

  As shown in FIG. 9B, the vertical and horizontal polarization degrees of the background portion that does not pass through the transparent body 114 are dark portions. On the other hand, the vertical and horizontal polarization degrees of the region that has passed through the transparent body 114 change in the polarization direction of the transmitted light due to birefringence. Further, as shown in FIG. 9A, unlike the case of FIG. 7A, in the monochrome luminance image, the part of the mounting table 113 becomes dark and the edge part of the transparent body 114 is emphasized.

[Configuration without polarizing filter]
In addition, the transparent body detection system according to the present invention may be configured not to include the polarizing filter 112. The transparent body detection system captures a first region including a transparent body 114 having a characteristic that the polarization direction of transmitted light changes, and captures a vertically polarized image and a horizontally polarized image (camera 12); From the image pickup means, which mounts the transparent body 114 and reflects light having a polarization direction substantially the same as the polarization direction of the third area that is imaged without including the transparent body 114 in the first area. A ground (tray 111) provided in a position sandwiching the mounting table 113 in which one of the gradation values of the vertically polarized image or the horizontally polarized image is stronger than the other, and the longitudinal and transverse polarization degree image based on the vertically polarized image and the horizontally polarized image. And an image processing device 13 that detects the transparent body 114 based on the distribution of the vertical and horizontal polarization degrees.

  In such a transparent body detection system that does not include the polarizing filter 112, it is only necessary that the light from the lower side (base) of the mounting table 113 has a specific polarization direction. For example, the reflected light of the tray 111 may be a light whose specific polarization direction is strong. In this case, the materials of the tray 111 and the mounting table 113 are preferably the same.

  FIG. 10 shows an example of the result of photographing by the camera 12 when the polarizing filter 112 is not provided. 10A is an example of a monochrome luminance image, and FIG. 10B is an example of a vertical and horizontal polarization degree image.

  As shown in FIG. 10B, the light transmitted through the transparent body 114 is shifted to the bright portion with respect to the background light having a slightly dark portion. Thus, even when the polarizing filter 112 is not used, it is possible to obtain contrast between the transparent body region and other portions. In the case of such a configuration in which the polarizing filter 112 is not disposed, the system can be simplified and the cost can be reduced.

(First embodiment of transparent body region extraction flow)
FIG. 11 shows a processing flow executed by the image processing apparatus 14 of the transparent body detection system 1 when the polarization direction of the polarizing filter as shown in FIG. 7 is the P component.

  First, the vertical and horizontal polarization degree imaging is performed by the camera 12 (step S101), and the vertical and horizontal polarization degree signals of each pixel obtained by the imaging are A / D converted and digitized by the A / D conversion unit 141 (step S102). ).

  Next, binarization processing is performed (step S103), and each binarized pixel value is stored in the memory 132 as image data (step S104).

  Next, for identification of each bright part, labeling of the bright part block is performed (step S105), and a bright part block having a predetermined area or more is detected. Further, the center of gravity and the second moment are calculated for the bright part lump having a predetermined area or more, and the center of gravity of the obtained bright part lump is obtained as the position of the transparent body 114, and the direction of the second moment is determined for the transparent body 114. The CPU 135 executes processing for obtaining the posture (step S106).

  Thus, first, the transparent body 114 can be detected in the image processing by detecting a cluster of bright portions having a predetermined area or more. Furthermore, the position and orientation of the transparent body 114 can be obtained by calculating the center of gravity and the second moment for the detected bright portion block having a predetermined area or more.

  After the above detection process, the CPU 135 executes a process of transmitting a control signal obtained from the position and orientation data of the transparent body 114 to the robot controller 15, and the robot controller 15 is driven and controlled. Then, it is picked up by the robot hand 16 in conformity with the above-mentioned posture, and aligned and packed in a tray (not shown) (step S107). This clogging is repeated until completion for all the detected transparent bodies 114 (steps S108 and S109).

Note that the vertical and horizontal polarization degrees in the above processing are calculated as follows. The camera 12 acquires the vertical polarization component (P), the horizontal polarization component (S), and polarized RAW image data including the vertical polarization component (P) and horizontal polarization component (S). The vertical / horizontal polarization degree information processing unit uses the acquired vertical polarization component (also referred to as P polarization component or P component) and the horizontal polarization component (referred to as S polarization component or S component) to create a vertical / horizontal polarization degree image and vertical / horizontal polarization. Calculate the degree. Here, the vertical and horizontal polarization degree image data is obtained by the following equation (1).
Vertical / horizontal polarization degree = (P-polarized component−S-polarized component) / (P-polarized component + S-polarized component) (1)

  Further, when calculating the position and orientation, a template may be prepared in advance, and the position and orientation may be calculated based on the template. For example, the transparent body is not limited to the spherical shape as shown in FIG. 7. For example, an image rotated every other degree is prepared in advance as a template, and the rotation state is set based on the template. A corresponding posture may be calculated.

  Note that other image processing by the image processing device 13, that is, each image processing such as binarization processing, labeling processing, center of gravity and second moment calculation processing, may be based on a known or new algorithm, and is particularly limited. Therefore, detailed description is omitted. Moreover, the threshold value used for each image process should just be set according to the magnitude | size, shape, etc. of the transparent body 114. FIG.

  As described above, according to the transparent body detection system 1 according to the present embodiment, the contrast of the portion where the transparent body 114 exists can be increased, and the transparent body 114 can be detected. As a result, since the longitudinal and lateral polarization degree images can be taken in, it becomes possible to detect the polarization change of the transparent body without depending on the direction of the polarizing filter 112 on the transparent base, and the calibration work becomes unnecessary. In addition, a rotating mechanism is not required (Problem 3). Moreover, the light which permeate | transmits the transparent body 114 should just be the light which permeate | transmitted the polarizing filter 112, and it is not necessary to arrange | position an illuminating device directly under the polarizing filter 112 (issue 1). Further, even if the mounting table 113 is installed on the tray 111, the polarizing filter 12 may be provided between the tray 111 and the mounting table 113, and the transparent object detecting device can be mounted from the mounting table 113. No work is required (Problem 2).

  In addition, the position and orientation of the transparent body 114 detected by image processing can be obtained easily and accurately. Furthermore, the transparent body 114 can be accurately picked up by the robot hand 16 based on the obtained position and orientation of the transparent body 114.

(Defect inspection)
Moreover, it is also preferable that the transparent body detection system 1 inspects for the presence or absence of defects such as chipping and flashing in conjunction with the detection of the position and orientation of the transparent body 114.

  First, similarly to the above, the existence area of the transparent body 114 is specified. Here, when the transparent body 114 has a defect of a chip 43, as shown in FIG. 12, the bright portion in the obtained (binary) vertical and horizontal polarization degree image is larger than the existing area of the transparent body in the case of a non-defective product. Get smaller. Further, when the transparent body 114 has a defect of the flash 44, as shown in FIG.

  Therefore, since the number of pixels in the existing region of the transparent body 114 in the case of a non-defective product is known, the normal range defined by the upper limit value and the lower limit value of the number of pixels is used as a reference for determining the presence or absence of a defect from the number of pixels. Is set, and the region where the transparent body 114 is present is specified, and the number of pixels in the bright portion of the transparent body 114 is counted. If this count value is within the normal range, it is determined that there is no defect, and the count value is within the normal range. If the count value is smaller than the lower limit value of the normal range, it can be determined that there is a defect 43 defect. As a result, it is possible to detect a shape defect such as chipping or flashing that may occur in the transparent body 114, and it is possible to discriminate between non-defective products and defective products.

(Transparent object inspection)
Moreover, it is also preferable that the transparent body detection system 1 is used for the inspection of the transparent foreign matter in the bottle for inspecting the presence or absence of the transparent foreign matter inside the translucent bottle. That is, when there is a transparent foreign matter, the vertical and horizontal polarization degree image is a dark or bright block of the transparent foreign matter. Therefore, the presence / absence of a transparent foreign object can be determined according to the presence / absence of a portion that appears as a lump in the image. This makes it possible to detect transparent foreign matters remaining in an empty bottle or a bottle containing a transparent liquid.

(Second embodiment of transparent body region extraction flow)
Another processing flow executed by the image processing apparatus 14 of the transparent body detection system 1 will be described. The second embodiment of the transparent body region extraction flow is different from the first embodiment in that, in addition to the longitudinal and transverse polarization degree images, the brightness image is used to detect defects such as dirt, scratches, and cracks, and character information such as logos and marks. The image information is detected and the classification method at the time of packing is increased.

Generation of monochrome luminance information using polarization information from the camera 12 will be described. The monochrome luminance processing unit creates a monochrome image and calculates luminance information using the acquired P component and S component. Further, the monochrome luminance is obtained by generating and outputting luminance information image data according to the following equation (2).
Monochrome brightness = P polarization component + S polarization component (2)

  When inspecting the presence or absence of defects such as dirt, scratches, and cracks generated on the transparent body 114, the position and orientation of the transparent body 114 are obtained by the processing flow 1 and the existence area of the transparent body 114 is specified.

  Here, when there is no defect in the transparent body 114, the existing area on the image obtained by the monochrome luminance image is an image having a low contrast as shown in FIG. On the other hand, if there is a defect (for example, dirt 45) in the transparent body, as shown in FIG. 15, the part corresponding to the dirt 45 becomes darker than the surroundings, and the area corresponding to the dirt part is partially High contrast image.

  Therefore, in the existence area on the luminance image, it is determined whether the size and degree of change of the luminance distribution of the monochrome luminance image are within a normal range, that is, the number of pixels equal to or less than a predetermined brightness. Measurement may be performed, and the presence or absence of a defect may be determined according to the measurement result.For example, if the number of pixels of the measurement result is equal to or greater than a predetermined inspection threshold, it is determined that there is a defect; Can be determined. As a result, it is possible to detect defects in appearance such as dirt, scratches, and cracks that may occur on the transparent body, and it is possible to discriminate between non-defective products and defective products.

  In addition, after specifying the position and orientation of the transparent body 114 in the vertical and horizontal polarization degree images, based on the size and degree of change of the luminance distribution changing portion of the monochrome luminance image, logos and marks written on the transparent body surface, etc. It is also preferable to detect character information and symbol information. Accordingly, it is possible to pick the transparent body 114 based on the character information and the symbol information together with the outer shape determination.

  The above-described embodiment is a preferred embodiment of the present invention, but is not limited thereto, and various modifications can be made without departing from the gist of the present invention.

DESCRIPTION OF SYMBOLS 1 Transparent body detection system 11 Transparent body mounting part 12 Camera (image imaging means)
13 Image processing device 14 Monitor 15 Robot controller 16 Robot hand 20a, 20b Polarizing filter 21 Optical system 22 Lens array 23 Light blocking spacer 24 Polarizing filter 25 Spacer 26 Solid imaging unit 27a, 27b Optical axis 28a, 28b Lens 29a, 29b Opening 30a , 30b Polarizer region 31 Opening 32 Substrate 33a, 33b Solid-state imaging device 34 Half mirror box 35 Mirror 36 Vertical polarization filter 37 Horizontal polarization filter 38, 39 CCD
40 Light-receiving element array 41 Polarizing filter region (vertical)
42 Polarizing filter area (horizontal)
43 chip 44 beam 45 dirt 101 fluorescent lamp 102 light shielding plate 111 tray 112 polarizing filter 113 mounting table 114 transparent body 131 A / D converter 132 memory 133 display controller 135 CPU

JP 2002-98650 A

Claims (8)

An image imaging means for imaging a first region including a transparent body having a property of changing a polarization direction of transmitted light, and imaging a vertically polarized image and a horizontally polarized image;
A mounting table for mounting the transparent body;
A polarizing filter installed at a position including a range of a second region photographed including at least the transparent body in the first region across the mounting table from the image capturing unit,
A transparent body detection system comprising: an image processing device that detects the transparent body based on a distribution of vertical and horizontal polarization degrees of a vertical and horizontal polarization degree image based on the vertical polarization image and the horizontal polarization image. An image imaging means for imaging a first region including a transparent body having a property of changing a polarization direction of transmitted light, and imaging a vertically polarized image and a horizontally polarized image;
Placing the transparent body, and a mounting table that reflects light having a polarization direction substantially the same as the polarization direction of the third area that is photographed without including the transparent body in the first area;
A ground on which one gradation value of the vertically polarized image or horizontally polarized image is stronger than the other, provided at a position sandwiching the mounting table from the image capturing means;
A transparent body detection system comprising: an image processing device that detects the transparent body based on a distribution of vertical and horizontal polarization degrees of a vertical and horizontal polarization degree image based on the vertical polarization image and the horizontal polarization image.   The transparent body detection system according to claim 1, wherein the image processing device detects a position and a posture of the detected transparent body. A robot hand and a robot controller for controlling the robot hand;
The robot controller moves the robot hand based on the detected position of the transparent body and picks up the transparent body based on the detected posture of the transparent body. The transparent body detection system described. The image processing device determines the presence area of the transparent body based on the detected position and orientation of the transparent body,
The transparent body detection system according to claim 3, wherein a shape defect of the transparent body is detected by comparing existing regions of two or more transparent bodies. The image processing device determines the presence area of the transparent body based on the detected position and orientation of the transparent body,
6. The transparent body according to claim 3, wherein an appearance defect of the transparent body is detected based on a luminance distribution of a monochrome luminance image based on the vertical polarization image and the horizontal polarization image. Detection system. The image processing device determines the presence area of the transparent body based on the detected position and orientation of the transparent body,
The character and / or symbol information written on the surface of the transparent body is detected based on a luminance distribution of a monochrome luminance image based on the vertical polarization image and the horizontal polarization image. The transparent body detection system in any one. The transparent body is a transparent foreign substance that can be mixed in an empty bottle or a bottle containing a transparent liquid,
The image processing device determines the presence area of the transparent body based on the detected position and orientation of the transparent body,
In addition, it is determined whether or not the transparent foreign matter is mixed in the empty bottle or the bottle containing the transparent liquid based on the distribution of the vertical and horizontal polarization degrees of the vertical and horizontal polarization degree images based on the vertical and horizontal polarization images. The transparent body detection system according to any one of claims 3 to 7, wherein: