JPH03167896A - Recognition of component in component mounting apparatus by image recognition - Google Patents

Abstract

PURPOSE:To detect the existing position of a component at high speed by using a low-speed computer by a method wherein, with regard to an image of a component mounting face, it is judged whether the component is good or not on the basis of the brightness level of picture element at positions on a grid being set for every prescribed number of picture elements. CONSTITUTION:A first control means 1 judges whether a flag F is equal to 1 or not. When F=0, picture-element data which are supplied one after another from an A/D converter 10 are taken into a RAM collectively. When F=1, window data for next operation use are taken in. Grids Gi are set at prescribed intervals of picture elements for an image of one frame; when a black grid level exists at a bonding of a window W set by using the black-level grid as a reference, it is judged that components are close to each other and that the components are not suitable for a suction operation; a next detection is executed. When the black level does not exist on the boundary and the number of black grid levels inside the W is within a prescribed range, it is judged that they are situated within a range of target components as viewed from a TV camera 8. An interval between black-level picture elements is checked. When the interval is within a prescribed range, coordinates at the center of gravity are computed and it is set as F=1.

Description

[Detailed Description of the Invention] Technical Field The present invention relates to a component mounting device that automatically and sequentially mounts components onto a so-called printed circuit board (hereinafter simply referred to as a board).
In particular, the present invention relates to a component recognition method in an image recognition component mounting apparatus that grasps a component and mounts it on a board using image recognition.

BACKGROUND ART A component mounting device operated by such image recognition supplies electronic components randomly (in an unaligned state) onto a conveyor in order to correctly mount electronic components on a board, which are difficult to supply in an aligned manner because they have weak terminals. The mounted electronic components are imaged by a television camera, and the resulting image data is processed to determine the shape, position, and orientation of each component.

In the conventional image recognition component mounting apparatus, the image data processing program is complicated, and if the component mounting speed is to be increased, a high-speed computer is required, resulting in high cost and impractical problems.

Summary of the Invention [Object of the Invention] An object of the present invention is to provide a component recognition method in an image recognition component mounting apparatus that is formed by a low-speed computer and can detect the location of a component at high speed.

[Structure of the Invention] The image recognition method in the image recognition component mounting apparatus according to the present invention includes capturing images of electronic components randomly supplied on a component placement surface, generating image signals, and processing the image signals. 2
A component recognition method in an image recognition component mounting apparatus that uses pixel data constituting a dimensional screen and recognizes the presence of the electronic component on the component placement surface based on the pixel data, the method comprising: a determination storage step of storing a black level grid by determining that the luminance level of the pixel data in the grid as coordinate positions of every predetermined pixel is at a black level indicating the presence of a component; and using the black level grid as a reference. a boundary determination step that determines that no black level grid exists on the boundary of the window area and generates a determined output; and a predetermined number of black level grids stored by the storage means in accordance with the determination in the boundary determination step. a simple area determination step that generates a determination output by determining whether the number is greater than or equal to a predetermined number; and determining that the number of black level pixels within the name of the window area exceeds a predetermined number according to the determination in the simple area determination step. The method is characterized by comprising an area detailed determination process.

[Operation of the Invention] According to the component recognition method of the present invention as described above, instead of determining the level of all data of pixel data obtained from an image signal, coordinate positions at predetermined intervals of a constant number of pixels, that is, The presence of parts is roughly determined by determining the level of pixel data in the grid.

Embodiments Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

First, with reference to FIG. 1, an image recognition component mounting apparatus that executes a component recognition method according to the present invention will be described.

In this device, a first control means 1 drives a motor 3 for operating a component supply device 2, and several electronic components 5 (in this case, light emitting elements) are randomly transferred from the component supply device 2 onto a conveyor 4 ( In other words, they are dropped in an unaligned state. Then, a motor 6 for operating the conveyor 4 is driven, and the light emitting element 5 dropped onto the conveyor 4 is brought to a component supply position 9 where it can be photographed by a television camera 8 as an imaging means. An image signal generated by the television camera 8 is converted into pixel data via an A/D converter 10 and sent to the first control means 1.

The first control means 1 learns the shape, placement and posture of each light emitting element 5 at the parts feeding position 9 based on this pixel data, and drives the motor 13 for operating the X-Y table 12 to drive the parts transport chuck. 14, the light emitting element determined to be transported first is brought to a position where it can be gripped, and the light emitting element is gripped.

Then, this gripped light emitting element 5 is turned over by a front-back reversing device 15.
If the delivered light-emitting element is turned upside down, the front-back reversing device 15 is operated to correct the posture so that the light-emitting element 5 always faces up, and the index table 1
Transfer it to the index chuck 18 above 7. The index table 17 is connected to the cam mechanism 2 by the main motor 19.
The first control means 1 controls the operation of the cam mechanism 20 to rotate the cam mechanism 20 intermittently by 36 degrees. The index chucks 18 are mounted on the index table 17 at equal intervals (36° pitch).
It is provided in 0 locations. The index chuck 18 is controlled by a cam mechanism 20 by a second control means 22 at a component supply location by the above-mentioned front/back reversing device 15 and at a component discharge location described below.
The gripping claws are opened by applying mechanical force from the index chuck drive devices 23 and 24, which are controlled via the index chuck drive devices 23 and 24. When this mechanical force is released, the gripping claws are closed.

As mentioned above, the index table 17 has 36″ as 1
The television camera 26, the terminal handling device 27, and the terminal bending device 28 are rotated intermittently as pitches, and the television camera 26, the terminal handling device 27, and the terminal bending device 28 correspond to the positions advanced by one pitch from the above-mentioned parts supply point.
, two television cameras, and a component mounting chuck 30 are provided. The position where this component mounting chuck 30 is placed is the above-mentioned component ejection location, and the index chuck 18 that grips the light emitting element 5 is applied with mechanical force by the index chuck drive device 24 at this position to grip the light emitting element. The state is released and the part is transferred to the component mounting chuck 30.

The index chuck drive devices 23 and 24, the terminal handling device 27, the terminal bending device 28, and the component mounting chuck 30 are operated by the main motor 19 via a cam mechanism 20 and a clutch 21 whose operations are controlled by the second control means 22.
Driven by. Further, the timing of driving the terminal handling device 27, the terminal bending device 28, and the component mounting chuck 30 is based on a signal from the encoder 31. The second control means 22 first recognizes the state of the terminal of the light emitting element 5 gripped by the index chuck 18 based on the image signal from the television camera 26, and operates the terminal handling device 27 accordingly. Therefore, the terminals 5a of the light-emitting element 5, which had been deformed as shown in FIG. 2, are handled with their tips facing forward, and are molded parallel to each other as shown in FIG.

The second control means 22 then actuates the terminal bending device 28 to bend the terminal 5a approximately 90'' as shown in FIG.

Thereafter, the second control means recognizes the state of the terminal 5a of the light emitting element 5 based on the image signal from the television camera 29, and if it determines that the molding state of the terminal 5a is good, it operates the index chuck drive device 24. Then, the light emitting element 5 is released from being held by the index chuck 18, and the light emitting element 5 is held by the component mounting chuck 30. Then, a motor 34 for operating the X-Y table 33 is driven, and a printed circuit board 35 is brought to be mounted with the light emitting element 5.

Thereafter, the second control means 22 rotates the component mounting chuck rotation motor 37 to rotate the component mounting chuck 30 to an appropriate angular position, thereby determining the angular position of the light emitting element 5 with respect to the printed circuit board 35. Then, the component mounting chuck 30 is lowered, and as shown in FIG.
5, the luminous cord 5 is attached to it. At this time, the terminal 5a of the light emitting element 5 is inserted into the terminal insertion hole 35 formed in the printed circuit board 35.
a, and its tip portion protrudes to the lower surface side of the printed circuit board 35. The second control means 22 inspects the conductivity of the terminal 5a using three continuity checking/terminal cutting devices, and if the continuity is good, cuts the tip of the terminal 5a using the continuity chuck/terminal cutting device 39. and cut it to the specified length.

On the other hand, when the second control means 22 recognizes that the terminal continuity test result by the continuity chuck/terminal cutting device 39 is defective, the second control means 22 uses the component mounting chuck 30 to remove the light emitting element 5 from the printed circuit board 35 and remove the light emitting element 5 from the printed circuit board 35. This is stored in the defective parts storage section 40 for recovery, and the subsequent light emitting element is immediately mounted.

Thereafter, the above series of operations are performed for each light emitting element to be mounted on the printed circuit board 35.

Next, the component recognition operation in the first control means 1 will be explained with reference to FIG.

The means for performing the component recognition operation by the first control means is a CPU.
, RAM, and ROM. FIG. 6 is a flowchart showing a program executed by this microcomputer, which is executed at a desired timing by interrupting the program controlling the control operation of the motor 13 and component transport chuck 14 by the first means.

In this program, first, it is determined whether the flag F is equal to 1 (step Sl), and when F is equal to 0, the pixel data sequentially supplied from the A/D converter 10 is collected for one frame. The data is imported into a predetermined area in the RAM (step S2).

Note that when it is determined that it is F-1, step 818 described below is performed.
to go into. FIG. 7 shows an example of a one-frame image represented by pixel data, and in this image it is shown that electronic components dl, d2, and d3 are placed on the conveyor 4 in a non-aligned state. ing. As shown in FIG. 8, according to the present invention, grids Gt (t-1, . . . 196) are set at predetermined pixel intervals (for example, 100 pixels) in the vertical and horizontal directions in one frame of image. Then, the area outside the existing window is checked in numerical order starting from the lowest number Gε (step S3),
When a black level grid is found, a region of a predetermined size, that is, a window W is created based on this detected black level grid.
(see FIG. 9) (step S4). Then,
As soon as a black level grid exists on the boundary of this window W, it is determined (step S5). When a black level grid exists on the window boundary, for example, as shown in Figure 10, the parts are close to each other and it is not suitable for parts pick-up. The process returns to step S3 to perform the black level grid detection operation in the direction of the black level grid.

On the other hand, if there is no black level grid on the window boundary, it is determined whether the number of black level grids within the window is within a predetermined value range (step S7).

If the number of black level grids in the window is not within the specified value range, the size of the target component is outside the specified range, or the component is placed at an angle to the mounting surface and cannot be suctioned. It is determined that it is not suitable and the process returns to step S3. On the other hand, when it is determined that the number of black level grids in the window is within the predetermined value range, it is assumed that the area of the target component as seen from the television camera 8 is within the specified range as shown in FIG. The determination operation begins in step S8. Step S8
, the maximum distance X between black level pixels in the horizontal direction and the maximum distance Y between black level pixels in the vertical direction within the window are predetermined maximum values XMX and YMx.
It is determined whether the value is smaller than or not. As shown in FIG. 12, if both XxXMx and Y<YMz are established, it is assumed that the size of the target part is within the specified range, and the process proceeds to step S9 for the next area check. In step S9, the number of black level pixels within the window is set to a predetermined minimum value NMN.
and a predetermined maximum value NMX. If the number of black level pixels in the window is not between NMN and NMx, the component area is outside the specified range, so the shape of the component is abnormal or the orientation of the component is incorrect, and step S3
Back to Ha. On the other hand, the number of black level pixels in the window is NMN
If it is determined that the position is between NMx and NMx, the coordinates of the center of gravity of the target component are calculated in preparation for calculating the component suction or gripping position (step S10). Since the method for calculating the coordinates of the center of gravity position is well known, it will not be described in detail here.

Next, in step S11, it is determined whether or not a flag F indicating that the operation up to the calculation of the suction position has already been performed for one component is set for the captured one frame image. When it is determined that the flag F is 1, the suction position calculation has already been completed for one component, so the center of gravity position coordinate data for the component in the current setting window is stored (step S12).
, return to the main routine. On the other hand, when F-0 is determined, black level data on the equivalent ellipse and principal axis of inertia of the target part d1 is calculated as shown in FIG.
13). This operation is also well known and will not be described in detail here.

Next, in step S14, as shown in FIG. 14, a long sleeve line g1 and a short axis line p2 connecting the farthest and closest points on the contour from the center of gravity of the part to the center of gravity are determined. Next, assuming that the counterclockwise angles of D1 and g2 with respect to the horizontal right direction reference line γ are 01 and θ2, Δθ - θ1 - θ2
Find Δθ - Δθ−3 when Δθ > 180@
60＠ Binding, when Δθ<-180°, Δθ−Δθ+
It is assumed to be 360°. In this way, -l80° and Δθ<l80
The front and back sides are determined based on the sign of Δθ (step S15). Next, in step S17, the position where the target part should be picked up by the suction means that picks it up from the placement device, that is, the gripping position is calculated. The coordinates (Px, Py) of this suction position can be obtained by calculating the following equation (1) or (2).

Here, Gx and Gy are the coordinates of the center of gravity position previously calculated, and GP is the distance between the center of gravity and the suction position, which is determined in advance for each type of component. θGP is the distance between the center of gravity and the suction position, and θcp is the angle between the line connecting the center of gravity and the suction position and the principal axis of inertia, and is determined in advance for each type of component. Further, θM is the angle formed between the principal axis of inertia obtained in step S13 and the horizontal direction.

In addition, in the case of the R state (table) in which the longitudinal part of the part extends to the right as shown in Fig. 15 (A), using the above equation (1), the longitudinal part of the part extends as shown in Fig. 15 (B) In the case of the L state (back) extending to the left, formula (2) is used.

When the suction position coordinates Px, Py are obtained in this way, the flag F is set to 1 (step S17), and then the process returns to step S3.

Note that when F-1 is determined in step S1, the next window data stored in step S12 is imported as the current one (step S18) without importing image data, and the flag F is After setting it to zero (step S19), the process proceeds to step S13.

In this way, steps 313 to S17 are executed again, and step S3 is entered to search for the next part. If no part is found, the process returns to the main routine, and at this time the conveyor is moved one pitch (step S20).

Effects of the Invention As is clear from the above, according to the component recognition method in an image recognition device according to the present invention, when determining the presence or absence of a component from an image signal obtained by capturing an image of a component placement surface, all pixel data levels are Instead of determining the presence or absence of parts, we decided to determine the brightness level of the pixel at the position of the grid set every predetermined pixel, so even if a computer with slow calculation speed is used, it is possible to quickly determine the presence or absence of a component. Part recognition speed is improved.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an image recognition component mounting apparatus that executes the component recognition method of the present invention, and FIGS. 2 to 5 are
FIG. 6 is a perspective view illustrating the operation when aligning the terminals of the target parts, and FIG. 6 is a flowchart showing a subroutine of the control means controlling the image recognition operation in the apparatus of FIG. 1, and FIGS. 7 to 15 (A) , (B) is a diagram showing the calculation contents of the subroutine operation shown in FIG. 6.

Claims (2)

[Claims] (1) Image the electronic components randomly supplied on the component placement surface to generate an image signal, process the image signal, and
A component recognition method in an image recognition component mounting apparatus that generates pixel data constituting a dimensional screen and recognizes the presence of the electronic component on the component placement surface based on the pixel data, the method comprising: and a determination storage step of storing a black level grid by determining that the brightness level of the pixel data in the grid as coordinate positions of every predetermined pixel in the vertical direction is at a black level indicating the presence of a component; a boundary determination step that determines that a black level grid does not exist on the boundary of a reference window area and generates a determined output; and a black level grid that is stored by the storage means in accordance with the determined output from the boundary determination step. a simple area determination step that determines that the number of pixels in the window area is within a predetermined range and produces a determined output; A component recognition method in an image recognition component mounting apparatus, characterized in that the method comprises: a step of determining area details for determining the location of the component; (2) The image according to claim 1, further comprising a range determination step for determining whether the existence range of black level pixels in the window area is within a predetermined range according to the determination in the simple area determination step. A component recognition method in a recognition component mounting device.