JP2016108081A - Image processing device for parts feeder and parts feeder - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processing device for parts feeder in which posture determination precision for a work may be highly maintained even in a case of a plurality of works having specific shape which is successively transported.SOLUTION: An image processing device for parts feeder is applied to parts feeder capable of being photographed with an area camera while transporting work W comprising a work main body W1 and an electrode W2 having different light reflection efficiency from the work main body W1, in which an embodiment is adopted where at least 2 surfaces Wf1, Wf1 adjacent each other are arranged at a position where they may be simultaneously photographed, the device comprises image obtaining means for obtaining image data and posture determining means for determining posture of the work W, end part W10a in longitudinal direction of the work main body W1 at a predetermined first region corresponding to either of 2 surfaces Wf1, Wf1 shown on the image data and predetermined second region corresponding to the other one on the basis of color element information in the first region and color element information in the second region to determine posture of the work W.SELECTED DRAWING: Figure 2

Description

  Even if a plurality of workpieces of a specific form are transported on the transport path without any gaps, the present invention accurately discriminates the longitudinal end of one workpiece and increases the posture discrimination accuracy of the workpiece. The present invention relates to an image processing apparatus and a parts feeder that can be maintained.

  2. Description of the Related Art Conventionally, there is known a parts feeder that can transport a workpiece, such as an electronic component, along a transport path to a predetermined supply destination (for example, Patent Document 1). The parts feeder disclosed in Patent Document 1 captures only one surface of a workpiece conveyed along a conveyance path in a state of facing an area camera, and determines the workpiece posture based on the obtained image data. It is configured. It is customary that a workpiece determined to be in a poor posture is rotated (reversed) by 90 ° in a predetermined direction on the conveyance path to be corrected in posture or removed from the conveyance path.

JP 2013-39981 A

  By the way, as the workpiece, for example, as shown in FIG. 14, electrodes W2 are provided at both ends in the longitudinal direction, respectively, and the posture is set on one of the four surfaces Wf1 of the workpiece main body W1 between these electrodes W2, W2. A feature surface W10 is formed for discrimination, and the other three surfaces Wf1 are non-feature surfaces W20 having a color different from that of the feature surface W10.

  When posture determination is performed for such a specific form of the workpiece W, as shown in the partial cross-sectional view of FIG. 15A, one surface of the workpiece W and the Erica camera 200 are not illustrated while facing each other. By illuminating the workpiece W with light from the light source, the workpiece W is imaged, image data in which one surface of the workpiece W and the electrode W2 appear is obtained, and it is determined whether or not the feature surface W10 appears in the image data. It is conceivable to discriminate the posture of the workpiece (1/4 selection). For example, as shown in the plan view of a part of the parts feeder in FIG. 5B, among the works W (WA to WD) conveyed in the X direction, the works WB and WD with the characteristic surface W10 facing upward are shown. If the posture is correct, the workpieces WA and WC with the non-characteristic surface W20 facing upward should be determined as a posture failure.

Here, in order to improve the discharge capability of the workpieces W (the number of workpieces W transferred per unit time to a predetermined supply destination), it is effective to transport a plurality of workpieces W without leaving a gap therebetween. is there.

  However, when a plurality of such specific forms of workpieces W are successively conveyed without any gap, in the obtained image data, the boundary Wa between the electrodes W2 and W2 adjacent to each other is determined as shown in FIG. (In the figure, a boundary that is difficult to distinguish is indicated by a two-dot chain line). In addition, in the figure (c), it has adjusted so that the electrode W2 may appear dark. Moreover, even when the electrode W2 is brightly captured by applying light from the light source as shown in FIG. 4D, it is difficult to determine the boundary Wa between the electrodes W2 and W2. (In the figure, a boundary that is difficult to distinguish is indicated by a two-dot chain line). Therefore, there is a problem that the accuracy of posture determination by pattern matching or the like is not able to be accurately determined for one workpiece W in the image data.

  Such a problem is not limited to the case where the electrode W2 is disposed at the longitudinal end portion Wa of the workpiece W, and the workpiece main body W1 and the light reflection efficiency at the longitudinal end portions Wa of the workpiece W. This may also occur when using a member provided with other members having different values.

  On the other hand, by adopting an area camera with high performance such as resolution, it is possible to capture fine phase shapes, and the cuts between the workpieces W and W are clearly shown in the image data, and the longitudinal direction of the workpiece W Although it is conceivable to make the end portion distinguishable, such a high-performance camera is expensive, leading to an increase in the cost of the parts feeder.

  An object of the present invention is to effectively solve such a problem, and the length of the workpiece can be reduced without using an expensive device even for a specific type of workpiece that is continuously conveyed without a gap between them. An object of the present invention is to provide an image processing apparatus for parts feeder and a parts feeder that can accurately discriminate the direction end and maintain high accuracy of posture discrimination of a workpiece.

  The present invention takes the following measures in view of the above problems.

  That is, in the image processing apparatus for parts feeder according to the present invention, the four substantially rectangular surfaces and the end surfaces at both ends in the longitudinal direction of the four surfaces form a substantially quadrangular prism shape, and any one of the four surfaces is formed. A workpiece main body having a posture distinguishing characteristic surface having a light reflection efficiency different from that of the other three surfaces, and an end member provided on the end surface of the workpiece main body and having a light reflection efficiency different from that of the three surfaces of the workpiece main body Is applied to a parts feeder that can be imaged with a camera while conveying the workpiece having a direction parallel to the longitudinal direction of the workpiece along the conveyance path, and has a plurality of imaging elements as the camera. An image capturing means for capturing image data acquired by the camera, wherein at least two surfaces adjacent to each other among the four surfaces of the work body are provided at positions capable of simultaneously capturing images; and the image capturing means Posture determination means for determining the posture of the workpiece based on the captured image data, a predetermined first region corresponding to one of the two surfaces of the workpiece main body appearing in the image data, and corresponding to the other In the predetermined second area, color element information is acquired over substantially the entire longitudinal direction of the workpiece, and based on the change in the color element information in the first area and the change in the color element information in the second area. The end of the work body in the longitudinal direction is detected, and the posture of the work is determined.

  Here, the light reflection efficiency is a measure of the amount of light provided from a light source placed at a certain position reflected by the surface of the object and incident on the image sensor of the image pickup means placed at another position. Yes, high reflection efficiency means that the amount of light incident on the image sensor is relatively large among the reflected light, and low reflection efficiency means that the light that is reflected to the image sensor. This means that the amount of incident light is relatively small. In addition to luminance and illuminance, the color element information includes information that appears as a difference in image data due to different reflection efficiencies such as brightness and saturation.

  With such a configuration, it is possible to simultaneously acquire image data in which two adjacent surfaces of the work body appear with one camera, and from the image data captured via the image capturing means, posture determination means Thus, the color element information of the first area and the second area is acquired over substantially the entire longitudinal direction of the workpiece. Since at least one of the first region or the second region corresponds to a surface that is not a feature surface, even if the feature surface and the end member have the same degree of reflection efficiency, the color of the first region Based on the change in the element information and the change in the color element information in the second region, the change in the color element information by the end member can be detected, and the end in the longitudinal direction of the work body can be accurately determined. . If the longitudinal end of the workpiece body can be identified, one workpiece can be cut out to obtain image data on which the workpiece has appeared, and the posture can be accurately identified based on this. Can do. Therefore, even for a plurality of workpieces of a specific form that are continuously conveyed without gaps between each other, it is possible to maintain high accuracy of workpiece posture discrimination without using a high-performance and expensive camera.

  When the feature surface has a mark for determining the posture in the front-rear direction of the workpiece at a position biased in the longitudinal direction of the workpiece, as a specific configuration that enables the posture determination as described above, The camera images only a part of the workpiece in the conveyance direction, almost the entire image in a direction substantially orthogonal to the conveyance direction, sequentially images the workpiece conveyed along the conveyance path, A plurality of image data can be acquired for each work, and further comprising preprocessing means for connecting the image data captured by the image capturing means in the order of imaging to generate composite image data, and the end member And a change in color element information corresponding to the feature plane including the mark, and a position of the mark is determined based on the composite image data. It is preferable.

  In particular, in order to achieve the above effect using a general camera and further speed up the posture discrimination processing, the camera is arranged in the workpiece conveyance direction and in a direction substantially orthogonal thereto. A setting unit that employs an area camera having a plurality of image pickup devices and sets only a part of the image pickup devices arranged in a row perpendicular to the transport direction to be usable for imaging among the plurality of image pickup devices of the area camera. The workpiece that is transported along the transport path by the imaging device set by the setting means is only partially in the transport direction, and substantially in the direction substantially perpendicular to the transport direction. It is preferable that images are sequentially captured in the entire imaging range so that a plurality of image data can be acquired for each work.

  On the other hand, the parts feeder of the present invention uses the above-described image processing apparatus for parts feeder, and includes a parts feeder main body having a conveyance path through which the work is conveyed, a plurality of imaging elements, A camera provided at a position where at least two of the four surfaces adjacent to each other can be imaged simultaneously, an image capturing unit that captures image data acquired by the camera, and image data captured by the image capturing unit. Utilizing it, posture discrimination means for discriminating the posture of the workpiece is provided.

  As described above, according to the present invention described above, two adjacent surfaces of the workpiece main body are simultaneously imaged, and based on the change in the color element information of each region corresponding to these two surfaces, the end portion in the longitudinal direction of the workpiece main body Parts feeder that can maintain high accuracy of posture determination of workpieces without using expensive equipment, even for a plurality of workpieces of a specific form that are transported continuously without gaps. An image processing apparatus and a parts feeder can be provided.

The side view which shows the parts feeder which concerns on one Embodiment of this invention. The perspective view which shows the workpiece | work which can be conveyed with the parts feeder. Sectional drawing which cut | disconnects and shows the same parts feeder shown in FIG. 1 by the cutting plane line III-III. The figure which shows the some workpiece | work conveyed along the conveyance path of the parts feeder from upper direction. The figure which shows typically the image data acquired with the parts feeder. The figure for demonstrating the change of the brightness | luminance of the workpiece | work which appears in image data. The graph which shows the brightness | luminance for every site | part of a workpiece | work. The figure for demonstrating the average brightness | luminance of four area | regions of a workpiece | work main body. The figure for demonstrating the change of the brightness | luminance which appears in image data when another workpiece | work is used. The graph which shows the brightness | luminance for every site | part of the work. The figure for demonstrating the average brightness | luminance of four area | regions of a workpiece | work main body. The timing chart for demonstrating operation | movement of the parts feeder. The figure which shows the modification of this invention. The figure which shows an example of the workpiece | work conventionally used. The figure for demonstrating the subject of the conventional parts feeder.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

  As shown in FIG. 1, the parts feeder 100 which is one Embodiment of this invention conveys the several workpiece | work W which is a conveyed product toward the supply destination which is not shown in figure along the conveyance path 10 with which the parts feeder main body 100 is provided. Is.

  As shown in FIG. 2, the parts feeder 100 according to the present embodiment has a work body W1 that is formed into a substantially quadrangular prism shape by four substantially rectangular surfaces Wf1 and substantially square end surfaces Wf2 at both longitudinal ends of the surface Wf1. And a workpiece W configured to have electrodes W2 as end members provided at both ends in the longitudinal direction of the workpiece main body W1, respectively, so that the longitudinal direction is parallel to the conveying direction. is there. The electrode W2 has relatively little unevenness on the surface, while the work body W1 has relatively much unevenness on the surface. Therefore, the surface of the electrode W2 and the surface of the work body W1 have different light reflection efficiencies.

  Here, the light reflection efficiency is a measure of the amount of light given from a light source placed at a certain position reflected by the surface of the object and incident on the image sensor of the imaging means placed at another position. The high reflection efficiency means that the amount of light incident on the image sensor is relatively large among the reflected light, and the low reflection efficiency means the image sensor among the reflected light. The amount of light incident on is relatively small. Even with the same object, the reflection efficiency may be different depending on the position of the light source or the imaging means. The reflection efficiency also varies depending on the surface properties such as the material and shape of the object to which light is applied, the surface roughness, and the color. This difference in reflection efficiency affects the appearance (color element) in the image data obtained by the imaging of the area camera 2. This will be described in detail later.

  Of the four surfaces Wf1 of the work body W1, one surface Wf1 is formed with a posture-determining feature surface W10 having a light reflection efficiency different from that of the other three surfaces W20. The light reflection efficiency can be adjusted by the color and the surface shape. In this embodiment, one surface Wf1 of the workpiece main body W1 with relatively many irregularities is different from the other three surfaces Wf1 in color and surface shape. Surface W10 is formed. On this characteristic surface W10, a substantially rectangular mark M for determining the posture of the workpiece W in the front-rear direction is formed at a position where the longitudinal end portion Wa of the workpiece W is deviated. The mark M and the non-mark portion W11 which is a portion other than the mark M in the feature surface W10 are configured in different colors and have different light reflection efficiencies. The mark M has substantially the same reflection efficiency as the non-characteristic surface W20, which is the other three surfaces, and the non-mark portion W11 has a reflection efficiency substantially equal to that of the electrode W2. The work body W1 and the electrode W2 are slightly different in height. Further, as the workpiece W, a workpiece in which the mark M is not formed on the feature surface W10 may be used.

  The parts feeder main body 100 includes a conveyance path 10 and driving means 11, and conveys a plurality of workpieces W on the conveyance path 10 by vibrating the conveyance path 10 by the driving means 11. As shown in FIG. 3, the conveyance path 10 has a substantially V-shaped cross section perpendicular to the conveyance direction of the workpiece W. The conveyance path 10 contacts the two surfaces Wf1 and Wf1 of the four surfaces Wf1 of the workpiece body W1. A plurality of workpieces W can be transported in the direction parallel to the longitudinal direction without any gaps as shown in the plan view of FIG. 4 while contacting the remaining two surfaces Wf1 and Wf1 obliquely upward.

  As shown in FIG. 1, an area camera 2 is provided above such a conveyance path 10 as an imaging unit, and the area camera 2 is arranged in the conveyance direction of the workpiece W (extending direction of the conveyance path 10) and A plurality of high-sensitivity imaging elements (CMOS sensors (Complementary Metal Oxide Semiconductors)) arranged in a direction orthogonal to the above are provided, and imaging is performed while applying light to the workpiece W from both sides in the longitudinal direction of the workpiece W from a light source (not shown). Therefore, the workpiece W can be imaged without causing a shadow due to the heights of the workpiece body W1 and the electrode W2 to be different at these boundary portions. In the present embodiment, the light source is disposed at a position where light is incident on the upper surface of the workpiece W at a shallow angle, and each part of the workpiece W that appears in the image data depending on the position of the light source will be described later. The color element information to be adjusted can be adjusted. Note that the scan rate of the area camera 2 is about several tens of kHz, and is set to a sufficiently high speed with respect to the work W conveyance speed.

  The area camera 2 can use only a part of the imaging elements (one row in the present embodiment) arranged perpendicular to the conveyance direction of the workpiece W by the setting unit 30 for imaging. The imaging range (imaging line) E of the area camera 2 is set to a range in which only a part of the workpiece W can be imaged in the conveyance direction and substantially the whole can be imaged in a direction orthogonal to the conveyance direction. Therefore, the area camera 2 can simultaneously image the two adjacent surfaces Wf1 and Wf1 of the work main body W1, and can obtain image data a to d as shown in FIG. 5, for example.

  Specifically, the image data a is an image of the electrode W2, and the electrode W2 appears relatively dark in the image data a. Further, the image data b is an image of the vicinity of the boundary with the electrode W2 in the work body W1, the non-marked portion W11 appears relatively dark in the image data b, and the non-of 4 surfaces Wf1 of the work body W1. The characteristic surface W20 appears relatively bright. The image data c is an image of the mark M, and the non-marked portion W11 appears relatively dark in the image data c, and the mark M and the non-characteristic surface W20 appear relatively bright. Further, the image data d is an image of a portion of the mark body W1 that does not correspond to the mark M, and the non-marked portion W11 appears relatively dark and the non-characteristic surface W20 appears relatively bright in the image data d. Depending on the posture of the workpiece W, two non-characteristic surfaces W20 may appear in the image data, or the characteristic surface W10 may appear on the lower side of the figure. It is possible to make the electrode W2 appear brighter depending on how the light from the light source is applied.

  The area camera 2 shown in FIG. 1 operates so as to continuously capture images at regular intervals before the workpiece W reaches the imaging position P1, and the workpiece W conveyed toward the downstream side passes through the imaging position P1. In the meantime, imaging is performed a plurality of times, and a plurality of pieces of image data in which different positions of the workpiece W appear over substantially the entire longitudinal direction of the workpiece W are acquired. The acquired image data is transferred to a control device (controller) 3 to be described later every time an image is taken. The image data as described above acquired by the area camera 2 has a smaller number of pixels and a smaller amount of data than image data acquired by imaging using almost all imaging elements of the area camera 2, and therefore, FIG. Can be immediately captured in the control device 3 via the image capturing means 31 shown in FIG.

  The control device 3 shown in FIG. 1 is composed of a normal microcomputer unit having a CPU, a memory, an interface, etc. (not shown). An appropriate program is stored in the memory, and the CPU sequentially stores the program. In cooperation with the peripheral hardware resources, it plays the role of the setting means 30, the image capturing means 31, the preprocessing means 32, the posture determination means 33, and the command output means 34.

  The image capturing means 31 captures image data acquired by the area camera 2 into the control device 3 and captures the image data immediately each time the area camera 2 captures an image.

  The preprocessing unit 32 includes a binarization processing unit 32a, a detection unit 32b, and a composite image data generation unit 32c. When image data is captured via the image capture unit 31, the binarization processing unit 32a 6, the luminance data as the color element information of the first area 61 corresponding to one of the two surfaces Wf1 and Wf1 of the work body W1 appearing in the image data, and the first area 61 The brightness data of the second region 62 corresponding to the other of the two surfaces Wf1 and Wf1 are sequentially acquired while being separated in a direction substantially perpendicular to the conveyance direction of the workpiece W.

  Then, the brightness change data in the longitudinal direction of the workpiece W is generated as UP (1) if the obtained brightness value is higher than the threshold A, and as DOWN (0) if it is lower. In the present embodiment, the luminance value is DOWN (0) at the electrode W2 and the non-marked portion W11, and UP (1) at the mark M and the non-characteristic surface W20.

  Therefore, as shown in FIG. 7, in both the first region 61 and the second region 62, the luminance value of the electrode W2 becomes DOWN (0), and the OR output becomes zero. Further, in the range that does not correspond to the mark M in the workpiece main body W1 (non-mark range N (see FIG. 6)), the first region 61 located on the characteristic surface W10 becomes DOWN (0) and is located on the non-characteristic surface W20. In the second area 62, the output becomes UP (1) and the OR output becomes 1. Furthermore, in the range corresponding to the mark M in the workpiece main body W1 (mark corresponding range M1 (see FIG. 6)), the first area 61 becomes UP (1) and the second area 62 becomes UP (1). The output becomes 1. Note that the positions of the first region 61 and the second region 62 in the direction orthogonal to the conveyance direction of the workpiece W are appropriately set so as to overlap the mark M when the feature surface W10 appears in the image data.

  The detection unit 32b determines a change in luminance corresponding to the work body W1 based on the luminance change data. Specifically, when the OR output of one region 61 (62) and the other region 62 (61) changes from 0 to 1, the detection unit 32b is on the downstream side in the conveyance direction of the workpiece W (front side in the traveling direction). It is determined that it corresponds to the boundary between the electrode W2 and the work body W1 (the end W10a on the downstream side in the transport direction of the work body 1). In addition, when the OR output of one region 61 (62) and the other region 62 (61) changes from 1 to 0, the detection unit 32b is an electrode on the upstream side in the conveyance direction of the workpiece W (the rear side in the traveling direction). It is determined that it corresponds to the boundary between W2 and the workpiece main body W1 (the end W10a on the upstream side in the transport direction of the workpiece main body W1).

  When the end W10a of the workpiece main body W1 is detected in this way, the composite image data generation unit 32c connects the image data in the order of imaging, and obtains two-dimensional image data in which almost one workpiece W appears. Composite image data is generated. The composite image data generation unit 32c estimates the acquired images of the electrode W2 portions before and after the workpiece main body W1 from the set values and the scan rate, the lens magnification, and the CMOS light receiving element size that are the actual dimensions of the workpiece main body W1 and the electrode W2. It is possible to synthesize an image for one workpiece. Note that, if there is image data in which substantially the entire work body W1 appears, posture determination processing described later is possible, so the composite image data generation unit 32c does not show the electrode W2 as composite image data, and the work body W1. It is also possible to generate the one in which substantially the whole appears.

  The posture determination means 33 determines the posture of the workpiece W based on such composite image data. Specifically, the posture determination means 33 divides the work main body W1 appearing in the composite image data into two at the center in the transport direction and into two in the direction orthogonal to the transport direction, for a total of four areas (first number). 1-1 region, 1-2 region, 2-1 region, 2-2 region). Then, a threshold value is set so that the average luminance of the area where the mark M exists is UP (1) and the average luminance of the other areas is DOWN (0), and each area is binarized, and the four areas Based on these correlations, the quality of the workpiece W with respect to a preset posture is determined. FIG. 8 shows all the correlation patterns of the four areas when the workpiece W is used. Specifically, the first area 61 is transported downstream (front side) or upstream (transport direction) ( If the mark M appears on the rear side) is set to the correct posture, the pattern a and b in FIG. 8 can be determined as non-defective products, and the quarter selection can be performed. If the correct posture is shown on the downstream side (front side) in the transport direction, it is possible to select only the pattern a in FIG. Note that even when a workpiece on which the mark M is not formed on the feature surface W10 is used, it is possible to perform the quarter selection with the same idea.

  Further, as shown in FIG. 9, the pattern matching by the detection of the work main body W1 and the region division using such luminance has the same light reflection efficiency between the mark M and the electrode W2, and the non-marked portion W11. Even a workpiece W ′ having a reflection efficiency comparable to that of the non-characteristic surface W20 can be accurately detected.

  Specifically, as shown in FIG. 10, the work W ′ becomes UP (1) in the first area 61 in the range not corresponding to the mark M (non-mark range N) in the work body W1, and the second area 62. As a result, UP (1) and the OR output becomes 1. Further, in the mark corresponding range M1, DOWN (0) is obtained in the first area 61, UP (1) is obtained in the second area 62, and the OR output becomes 1. Therefore, when the OR output of both the regions 61 and 62 changes from 0 to 1, the detection unit 32b detects the boundary between the electrode W2 and the workpiece body W1 on the downstream side in the conveyance direction of the workpiece W (front side in the traveling direction). The main body W1 is determined to correspond to the end W10a) on the downstream side in the transport direction. In addition, when the OR output of both the regions 61 and 62 changes from 1 to 0, the detection unit 32b detects the boundary between the electrode W2 and the workpiece body W1 on the upstream side in the conveyance direction of the workpiece W (the rear side in the traveling direction). It is determined that it corresponds to the upstream end W10a) of the work body W1 in the transport direction.

  In addition, when the posture discriminating unit 33 assumes that the mark M appears on the downstream side (front side) in the transport direction or the upstream side (rear side) in the transport direction of the first region 61, the pattern shown in FIGS. If the mark M appears on the downstream side (front side) in the transport direction of the workpiece body W and is in the correct posture, the pattern of FIG. Only one-eighth selection can be performed by judging only the non-defective product.

  As described above, the image capturing means 31, the preprocessing means 32, and the posture determination means 33 constitute the part feeder image processing apparatus 8 of the present invention for determining the posture of the workpiece W.

  When the command output unit 34 determines that the posture determination unit 33 is in an inappropriate posture (incorrect posture), the command output unit 34 is at the posture correction position P2 set in the conveyance path 10 with respect to the posture correction unit 5 illustrated in FIG. A command for rotating (reversing) the workpiece W by 90 ° around the axis parallel to the conveying direction on the conveying path 10 is output. Since this command is transmitted after a predetermined time has elapsed with reference to the time point when the detection unit 32b detects a specific part of the workpiece W (end W10a in the longitudinal direction, etc.), a trigger for command output becomes unnecessary. The fiber sensor for the synchronization trigger and the hole processing for the synchronization are not necessary. The posture correction means 5 includes an air injection nozzle 50 as an urging force applying unit that injects compressed air toward the posture correction position P2 set downstream of the area camera in the transport direction. The urging force is applied to the workpiece W by the jetted compressed air, and the workpiece W is rotated 90 ° on the conveyance path 10 to change the posture. The air injection nozzle 50 is formed by, for example, a hole provided in the side wall 10a of the conveyance path 10, and compressed air is injected by inputting an energization command as the command.

  The operation of the parts feeder 100 configured as described above will be described with reference to the timing chart shown in FIG. In the following, an operation from when one work W having an inappropriate posture is imaged by the area camera 2 until the posture is corrected by the posture correcting means 5 is described.

  When the workpiece W transported on the transport path 10 is imaged at time t01, the image data acquired thereby is immediately captured (transferred) via the image capturing means 31, and preprocessing is performed on the image data. Means 32 performs preprocessing. In this preprocessing, the detection unit 32b detects a location corresponding to the boundary between the work main body W1 and the electrode W2 (the end W10a in the longitudinal direction of the work main body W1) from the luminance change data at time t02. Even after imaging at time t01, imaging is sequentially performed at predetermined intervals, and image data capturing and preprocessing are immediately performed each time. The composite image data generation unit 32c starts generating composite image data at time t03, and performs posture determination processing (pattern matching by region division) by the posture determination means 33 until time t04 based on the composite image data. Note that the processing up to time t03 can be performed by hardware (for example, a field-programmable gate array (FPGA)) in addition to being performed in software, and the processing after time t03 is a program stored in the memory. This is done in software by executing Thereafter, the command output means 34 is controlled so that the energization command is output at time t05 when the standby time tα has elapsed from time t02. As a result, compressed air is jetted from the air jet nozzle 50 of the posture correcting means 5, and the urging force of air actually acts on the workpiece W at time t06 when the transmission time td has elapsed from time t05. If the workpiece W that has been subjected to posture determination processing is in an appropriate posture and is determined to be in a predetermined posture by the posture determination processing, the posture for correcting the posture of the workpiece W on the transport path 10 is corrected. Processing (output of energization command and injection from the air injection nozzle 50) is not performed.

  In this way, the posture of the workpiece W is corrected and supplied to the supply destination.

  As described above, the image processing apparatus 8 for parts feeder according to the first embodiment has a substantially quadrangular prism shape with the four substantially rectangular surfaces Wf1 and the end surfaces Wf2 at the longitudinal ends of the four surfaces Wf1. Of the workpiece main body W1 and the end surface Wf2 of the workpiece main body W1 are provided respectively on the workpiece main body W1 and the above-mentioned 3 of the workpiece main body W1. Parts that can be imaged by the area camera 2 while conveying the workpiece W having the surface Wf1 and the electrode W2 as an end member having different light reflection efficiency along the conveyance path 10 in a direction parallel to the longitudinal direction of the workpiece 10 The area camera 2 includes a plurality of image sensors, and at least two surfaces Wf1 and Wf1 adjacent to each other among the four surfaces Wf1 of the work body W1 are the same. The position of the work W is determined based on the image capture means 31 that captures image data acquired by the area camera 2 and the image data captured by the image capture means 31. A posture discriminating means 33 for discriminating, in a predetermined first area 61 corresponding to one of the two surfaces Wf1 and Wf1 appearing in the image data, and a predetermined second area 62 corresponding to the other, the workpiece W Color element information is obtained over substantially the entire length direction of the workpiece W1, and the longitudinal end W10a of the work body W1 is determined based on the color element information of the first region 61 and the color element information of the second region 62. It detects and discriminate | determines the attitude | position of the workpiece | work W.

  With such a configuration, the image data on which the two surfaces Wf1 and Wf1 adjacent to each other of the work body W1 appear can be simultaneously acquired by one area camera 2, and the image captured via the image capturing means 31 is obtained. From the data, the posture determination means 33 acquires the color element information of the first area 61 and the second area 62 over almost the entire length of the workpiece W. Since at least one of the first region 61 or the second region 62 corresponds to the non-characteristic surface W20, the non-marked portion W11 of the characteristic surface W10 and the electrode W2 have the same degree of reflection efficiency. Even so, based on the change in the color element information in the first area 61 and the change in the color element information in the second area 62, the change in the color element information due to the electrode W2 is detected, and the longitudinal direction of the work body W10 is detected. It is possible to accurately determine the end portion Wa. If the longitudinal end of the workpiece W can be identified, one workpiece W can be cut out to obtain composite image data in which one workpiece W appears, and the posture can be accurately determined based on the composite image data. A determination can be made. Therefore, the posture determination accuracy of the workpiece W can be maintained high without using a high-performance and expensive camera even for a plurality of workpieces W of a specific form that are continuously conveyed without gaps.

  The feature surface W10 has a mark M for determining the posture of the workpiece W in the front-rear direction at a position offset in the longitudinal direction of the workpiece W. Only a part can be imaged substantially in the direction substantially perpendicular to the conveyance direction, and a plurality of image data can be acquired for each workpiece W by sequentially imaging the workpiece W conveyed along the conveyance path 10. And a pre-processing means 32 for connecting the image data acquired by the image capturing means 31 in the order of imaging to generate composite image data, and a change in color element information corresponding to the electrode W2, and a mark M Since the change of the color element information corresponding to the feature plane W10 including each is specified and the position of the mark M is determined based on the composite image data, the resolution of the area camera 2 is increased. And without the position of the longitudinal end portions Wa and the mark of the work W to accurately determine the attitude of the workpiece W can be accurately determined.

  In particular, the setting unit 30 further includes a setting unit 30 configured to set only a part of the plurality of imaging devices included in the area camera 2 to be used for imaging in a row orthogonal to the transport direction. The workpiece W transported along the transport path 10 by the set image sensor is only partly in the transport direction of the work W, and almost the entire imaging range in the direction substantially orthogonal to the transport direction. A plurality of image data can be acquired for each work by sequentially capturing images with E.

  With such a configuration, the number of pixels of the image data acquired by the area camera 2 in one imaging operation can be reduced, so that the capturing speed (transfer speed) by the image capturing means 31 is improved, and one work piece is obtained. It is possible to shorten the time from imaging to posture determination processing with respect to W and to speed up the transfer of the workpiece W. Therefore, by using the general-purpose area camera 2, the above-described effects that can maintain the posture discrimination accuracy at a high level can be achieved, and the speed of the posture discrimination process can be further increased. Can be transported without any gaps, and the processing capability of the workpiece W can be greatly improved.

  Furthermore, the parts feeder 1 of the present invention uses the part feeder image processing apparatus 8 and includes a parts feeder main body 100 having a conveyance path 10 through which the workpiece W is conveyed, and a plurality of imaging elements. An area camera 2 provided at a position where at least two surfaces Wf1 and Wf1 adjacent to each other among the four surfaces Wf1 of the work body W1 can be simultaneously imaged, and an image capturing unit 31 for capturing image data acquired by the area camera 2; In addition, since the image capturing unit 31 includes the posture determination unit 33 that determines the posture of the workpiece W using the image data captured, the plurality of specific forms that are continuously conveyed without gaps. Even for the workpiece W, the posture discrimination accuracy of the workpiece W can be maintained high without using a high-performance and expensive camera.

  As mentioned above, although one Embodiment of this invention was described, the specific structure of each part is not limited only to embodiment mentioned above.

For example, in the parts feeder 100 of the above embodiment, only one row of imaging elements is used for imaging in the area camera 2, but as shown in FIG. 13, the first that forms a row orthogonal to the conveyance direction of the workpiece W. An imaging element group and a second imaging element group forming a row perpendicular to the conveyance direction on the downstream side in the conveyance direction from the first imaging element group are set, and an imaging range (first The imaging line) E _L 1 or the work W located in the imaging range (second imaging line) E _L 2 of the second imaging element group is configured to be imaged. Incidentally, E _E shows the imaging range of the area camera 2 in the case of using all of the imaging element. Further, the posture correcting means 5 has two air injection nozzles 50a and 50b, and one of the air injection nozzles 50a includes an imaging range E _L 1 of the first imaging element group and an imaging range E _L 2 of the second imaging element group. And the other air injection nozzle 50b may be provided on the downstream side in the transport direction from the imaging range E _L2 of the second imaging element group. As a result, the posture determination means 33 (see FIG. 1) performs posture determination processing based on the image data acquired by the first image sensor group, and as a result, for the workpiece W determined as an inappropriate posture. At the same time, posture correction is performed using one of the air injection nozzles 50a, and posture determination processing is performed again based on image data acquired by the second imaging element group. The workpiece W that has been determined to be in an inappropriate posture in the posture determination processing again is subjected to posture correction processing using the other air injection nozzle 50b.

For example, if the resolution of the area camera 2 is increased in order to clearly show the gap between the workpieces W and W, the field of view becomes narrow. However, the present invention detects the end W10a in the longitudinal direction of the workpiece body W1. Since there is no need to increase the resolution, it is possible to perform the posture determination processing for one workpiece W a plurality of times using one area camera 2. Further, even when a plurality of camera installation spaces cannot be secured, it is possible to suitably perform a plurality of posture determination processes for one workpiece W.

  In addition, the area camera 2 includes an area scan mode in which all of the image sensors (or a part of the image sensors) provided in the area camera 2 are used for imaging, and a part of the image sensors arranged in the direction orthogonal to the transport direction (this embodiment). In this case, it is possible to switch to a line scan mode in which only one column) is used for imaging. When the detection unit 32b (see FIG. 1) detects the longitudinal end portion Wa of the workpiece W when the line scan mode is set, an area scan is performed. An image of substantially the entire work W may be captured in the mode, and the posture determination process may be performed based on the image data acquired in the area mode instead of the composite image data.

  In addition, from the two adjacent surfaces Wf1 appearing in the image data, an expected number of inversions required to obtain a predetermined posture may be obtained, and posture correction processing may be performed accordingly. For example, assuming that the feature surface W10 appears in the first region 61 is a correct posture, the posture correction processing is performed once if the feature surface W10 appears in the second region 62, and at least twice if the feature surface W10 does not appear. Since it is understood that it is necessary, the number of times of imaging is increased from 3 to 2 times by performing processing such as rotating the workpiece W twice until the next posture determination. Can be reduced.

  In the above-described embodiment, the preprocessing unit 32 performs preprocessing such as immediate binarization processing every time image data is captured by the image capturing unit 31, but the capturing of one workpiece W is completed. Then, the binarization process, which is the pre-processing, and the image combination may be performed on all image data in which the workpiece W appears.

  Furthermore, in the above embodiment, the image data acquired by the area camera 2 is synthesized, but the determination may be made without being synthesized.

  Furthermore, the image pickup device group is not limited to one in which the image pickup devices are arranged in one row, and two or more rows of image pickup devices adjacent to each other along the workpiece W conveying direction are arranged within the range in which the effect of the present invention is exhibited. It may be a thing. Further, a line camera may be used instead of the Erica camera 2 as the imaging means. Furthermore, the workpiece W is not only provided with electrodes W2 on both end faces Wf2 of the work body W1, but is provided with other members having light reflection efficiency different from that of the work body W1 on both end faces Wf2 of the work body W1. Even if it is a thing, it can perform an attitude | position discrimination | determination process suitably. In the above embodiment, the luminance is used as the color element information. However, other detection values such as saturation and illuminance may be used. Furthermore, as the workpiece W, a workpiece having a workpiece body W1 whose longitudinal direction is a direction orthogonal to the conveyance direction may be used.

  Other configurations can be variously modified without departing from the spirit of the present invention.

DESCRIPTION OF SYMBOLS 1 ... Parts feeder 2 ... Area camera 8 ... Parts feeder image processing apparatus 10 ... Conveyance path 30 ... Setting means 31 ... Image taking-in means 32 ... Pre-processing means 33 ... posture determination means 61 ... first area 62 ... second area E ... imaging range M ... mark W, W '... work W1 ... work body W2 ... end Parts (electrodes)
W10: Characteristic surface W10a: Workpiece body end Wa: Workpiece longitudinal end Wf1: Surface Wf2: End surface

Claims (4)

The substantially rectangular four surfaces and the end surfaces at both ends in the longitudinal direction of the four surfaces form a substantially quadrangular prism shape, and any one of the four surfaces is discriminated in posture with different light reflection efficiency from the other three surfaces. A workpiece body having a characteristic surface for use, and a workpiece provided on each of the end surfaces of the workpiece body and having end members having different light reflection efficiencies from the three surfaces of the workpiece body are arranged along the conveyance path. It is applied to a parts feeder that can be imaged with a camera while being conveyed in a direction parallel to the longitudinal direction,
As the camera, a camera having a plurality of image sensors, and being provided at a position where at least two surfaces adjacent to each other among the four surfaces of the work body can be simultaneously imaged, is adopted.
Image capturing means for capturing image data acquired by the camera;
Posture determination means for determining the posture of the workpiece based on the image data captured by the image capture means;
In a predetermined first area corresponding to one of the two surfaces of the work main body appearing in the image data and a predetermined second area corresponding to the other, color element information is provided over substantially the entire longitudinal direction of the work. Acquiring each, and detecting the end of the work body in the longitudinal direction based on the change in the color element information in the first area and the change in the color element information in the second area to determine the posture of the work An image processing apparatus for a parts feeder characterized by the above. The characteristic surface has a mark for discriminating the posture of the workpiece in the front-rear direction at an offset position in the longitudinal direction of the workpiece,
The camera images only a part of the workpiece in the conveyance direction, almost the entire image in a direction substantially orthogonal to the conveyance direction, sequentially images the workpiece conveyed along the conveyance path, Multiple pieces of image data can be acquired for each work.
Pre-processing means for generating composite image data by connecting the image data captured by the image capturing means in the order of imaging;
A change in color element information corresponding to the end member and a change in color element information corresponding to the feature plane including the mark are specified, and the position of the mark is determined based on the composite image data. Item 2. The image processing apparatus for parts feeder according to Item 1. As the camera, an area camera having a plurality of image sensors arranged in the workpiece conveyance direction and in a direction substantially perpendicular thereto is adopted.
Of the plurality of image sensors that the area camera has, further comprising setting means for setting only a part of the image sensors that form a row perpendicular to the transport direction to be usable for imaging,
With the imaging device set by the setting means, a part of the workpiece conveyed along the conveyance path is substantially the entire imaging range in the direction substantially perpendicular to the conveyance direction. The image processing apparatus for parts feeder according to claim 1 or 2, wherein a plurality of image data can be acquired for each work by sequentially imaging. Using the image processing apparatus for parts feeder according to any one of claims 1 to 3,
A parts feeder main body having a conveyance path through which the workpiece is conveyed;
A camera provided with a plurality of image sensors, and provided at a position capable of simultaneously imaging at least two surfaces adjacent to each other among the four surfaces of the work body;
Image capturing means for capturing image data acquired by the camera;
A parts feeder comprising: posture determination means for determining the posture of the workpiece using image data captured by the image capture means.

Images