JP2596158B2 - Component recognition device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a component recognition device, and more particularly, to a device that processes an image signal to recognize a position of a component in a field of view of an imaging device. .

2. Description of the Related Art An apparatus as shown in FIG. 5 is disclosed in Japanese Patent Application Laid-Open No. 63-18206 for inspecting the mounting state of electronic components mounted on a printed circuit board.
No. 1993. In the figure, (1) is a substrate, for example, a printed board, (2) is a test component having a side, for example, an electronic component, and (3a), (3b) are shadows on different sides of the electronic component (2). 2 so that
A light source that emits light from a direction, (4) is a light source (3a), (3
b) A photoelectric converter that captures an image of an electronic component by the light irradiation from step b) and converts the image into an electric signal.
(5a) and (5b) are first and second image memories for storing image signals output from the photoelectric converter (4), and (12) is an image processing device for performing image processing, and a subtraction device (11), An arithmetic unit (6) and a determination unit (7) are provided. Light source (3a),
(3b) and the photoelectric converter (4) are arranged such that an extension line connecting the light sources (3a) and (3b) and the photoelectric converter (4) is inclined with respect to the disposing direction of the electronic component (2). Has been established.

The conventional device is configured as described above, and detects the position and shape of a component by comparing the obtained image with a reference image. That is, first, the component to be inspected (2) is set, and one light source (3a) is turned on to irradiate the component to be inspected (2). Next, the image of the light irradiation surface is picked up by the television camera (4), converted into an electric signal, and stored in the first image memory (5a). Next, the light source (3a) is turned off, the other light source (3b) is turned on, and an image of the light irradiation surface is captured by the television camera (4). This is converted into an electric signal and stored in the second image memory (5b). The difference between the signal values of the pixels corresponding to the two image signals stored therein is calculated by a subtraction device (11) to obtain a difference image. The arithmetic unit (6) compares the difference image with a pre-stored reference image to determine the position of the component (2), and inputs the result to the determination unit (7), where the pre-stored reference value Is determined.

[Problems to be Solved by the Invention] In the conventional component recognition device as described above, two image signals for obtaining a difference image can always obtain an image in which the component surface of any component is reflected in the same manner. Was needed. Otherwise, for example, when the surface of the component has irregularities, and when different reflections occur for each component, such as when light or shadow is partially formed, unevenness on the component remains in the difference image, and Could not be accurately extracted. This will be described with reference to the drawings. FIG. 6A is an explanatory view showing a side surface of the component. When the surface of the component (2) has irregularities as described above, for example, in an image illuminated by the light source (3a), the signal level of the component (20a) is higher than that of a flat portion, and the signal level of the portion (20b) is flat. The image becomes lower than the normal part, and the image is taken as shown in FIG. 6 (b). On the other hand, in the image illuminated by the light source (3b),
Is lower than the flat part and the signal level of the part (20b) is higher than the flat part, and the image is taken as shown in FIG. 6 (c). Accordingly, when a difference image between the image of FIG. 6B and the image of FIG. 6C is created, the part (20a) and the part
0b) is not a shadow, but the signal level difference is large, and therefore, as shown in FIG.
The images were similar to (a) and (30b), and it was difficult to identify them by image processing. As described above, when obtaining a difference image between two image signals, unevenness in the images is added. Particularly, when the light source is disposed axially symmetric with the optical axis of the camera lens (4). Is amplified by a factor of about two, resulting in a level equivalent to the signal of the shadow on the side, and there is a problem that the position and shape of the component are erroneously recognized.

The present invention has been made to solve the above problems, and has as its object to provide a component recognition device capable of accurately recognizing the shape and position of a component.

[Means for Solving the Problems] A component recognition device according to the present invention provides three or more light sources capable of forming shadows on different sides of a component having a side, an image of a side of the component by irradiation of the light source. A photoelectric converter that captures an image and converts it into an electric signal, an image memory that stores the electric signal output from the photoelectric converter as an image signal, and two image signals of a shadow other than the recognized side of the side of the component. A maximum value image operation device for obtaining the maximum value image signal by comparing the signal values of the corresponding pixels of the image signal and leaving the larger one, a signal between the image signal of the recognized side and the maximum value image signal The apparatus includes a difference image calculation device that calculates a difference between values to obtain a difference image, and a determination device that recognizes a position of a component from a calculation result of the calculation device.

[Operation] In the present invention, the maximum of the two signals obtained by sequentially lighting the light sources capable of obtaining the shadows of the sides other than the recognized side is the signal having the larger output signal value. By creating a value image, it is possible to obtain an image in which unevenness of the image value due to unevenness of the component surface is eliminated. Furthermore, by taking the difference between the maximum value image and the image illuminated obliquely so that the shadow of the recognized side of the component is formed, the recognized side of the component can be accurately measured.

Embodiment An embodiment of the present invention will be described below with reference to FIG. FIG. 1 is a block diagram showing a configuration of a component recognition apparatus according to one embodiment of the present invention. This device is a mounting state inspection device that recognizes components mounted on a printed circuit board. In the drawing, (1) is a printed board, and (2) is a test component having a side portion, which is mounted on the printed board (1). (3) is a light source for irradiating the part (2) to be inspected,
(3a), (3b), (3c) and (3d) are printed circuit boards (1)
Above each at a different angle, eg light source (3a) and light source (3c)
Are installed obliquely so as to be symmetrical with respect to the component to be tested (2), and the light source (3b) is 90 ° horizontally from the light source (3a) and the light source (3c), respectively, and the component ( The angle and distance to 2) are set at the same position. Similarly, the light source (3
d) is installed obliquely so as to be symmetrical with the light source (3b) with the part to be tested (2) interposed therebetween, and at 90 ° in the horizontal direction from the light source (3a) and the light source (3c), respectively. The angle and the distance to the component (2) are set at the same position. That is, 4
Light sources (3a), (3b), (3c), and (3d) are components (2)
Are arranged so that shadows can be formed on different sides of the component (2).
° and are installed at two opposing positions with the part (2) interposed therebetween. (4) is the light source (3a), (3b), (3c),
A photoelectric converter that captures a shadow image of the side of the component (2) by the irradiation of (3d) and converts it into an electric signal, such as a television camera. (5) is an image memory for storing the image signal output from the photoelectric converter (4), for example, a first image memory (5a) for storing an image signal when the light source (3a) is irradiated, and a light source (3b). A second image memory (5b) for storing an image signal when illuminating the light source, a third image memory (5c) for storing an image signal when illuminating the light source (3c), and an image when illuminating the light source (3d) A fourth image memory (5d) for storing signals. (6)
Is a computing device that recognizes the side of the component (2) by computing the maximum value image and the difference image, and (7) is a decision device that outputs and displays a decision signal from the shape signal calculated by the computing device (6). is there.

The operation of the component recognition device configured as described above will be described. First, the printed circuit board (1) is installed at a position where an image can be captured by the photoelectric converter (4), the light source (3a) is turned on, and the printed circuit board (1) is irradiated. The image of the irradiation surface is picked up by the photoelectric converter (4) and stored in the first image memory (5a). The image at this time is shown in FIG. This light source (3
(a) is for irradiating the component (2) from the left side, and the shadow (8a) of the component is indicated by oblique lines in the figure. In the figure, (13) indicates a processed image, and the component (2) has four sides (2a), (2b), (2c), and (2d).
Next, with the light source (3a) turned off and the light source (3b) turned on from the near side toward the component (2), an image of the irradiated surface is captured by the photoelectric converter (4), and the second image memory (5b) is taken. ). The image at this time is shown in FIG. Next, the light source (3
Turn off the light from b) and turn on the light source (3
With c) turned on, the image on the irradiation surface is converted to a photoelectric converter (4).
And the image is stored in the third image memory (5c). The image at this time is shown in FIG. Next, with the light source (3c) turned off and the light source (3d) turned on from the side facing the component (2), an image of the irradiated surface is captured by the photoelectric converter (4) and the fourth image memory (5d) is taken. To memorize. The image at this time is
It is shown in FIG.

Next, a case where the side portion (2b) of the component (2) is recognized will be described. The arithmetic unit (6) selects two image signals for the sides other than the side (2b). For example, the first image memory (5
a), comparing the signal values for each image in the image stored in the third image memory (5c), and calculating the larger signal level to create a maximum value image. This is shown in FIG. Next, the signal value of this maximum value image and the second image memory (5
The difference from the signal value stored in b) is obtained and used as a difference image.
This is shown in FIG. 3 (b).

That is, the maximum value of the image signal fa (x, y) when the light source (3a) is irradiated with light and the image signal fc (x, y) when the light source (3c) is irradiated with light is calculated for the shadow image in FIG. Determined by device (6). Next, the difference between the maximum value and the image signal fb (x, y) when the light is emitted by the light source (3b) is obtained by the arithmetic unit (6).
Let g (x, y) = max (fa (x, y), fc (x, y)) − fb (x, y).

Here, when the difference image g (x, y) shown in FIG. 3B is examined, when compared with the signal value of each pixel in the maximum value image shown in FIG. 3A, the light source (3b) The part where the shadow (8b) of the part is formed in the image illuminated with the part (2) is dark. Accordingly, the signal value at the corresponding pixel is small, and the value of g (x, y) becomes a positive value, and remains as a shadow (8e). However, in the image illuminating the component (2) with the light source (3a), the shadow (8
Parts (2) depending on the part where a) was made and the light source (3c)
In the image illuminated with, the part where the shadow (8c) of the part is formed is replaced by the signal value when no shadow is generated when the maximum value image is obtained, and g (x, y) of the part is close to zero Value. 6 (b) and 6 (c) (20
As shown in (a) and (20b), the part where the signal level appearing on the component surface has unevenness is max (fa (x, y), fc (x, y))
Is replaced by the brighter signal level, the variation in the brightness of the entire component surface is fa (x,
y) and fc (x, y) are smaller for the maximum value image than for the single image. In addition, since the direction of irradiating the light source (3b) is different from the direction of irradiating the light sources (3a) and (3c), the position of the unevenness of the signal level caused by the unevenness of the component surface in the image signal fb (x, y). Is the image signal fa
It is different from the position of the unevenness appearing in each of (x, y) and fc (x, y). Therefore, the difference image g (x, y) (= max (fa (x,
y), fc (x, y)) − fb (x, y)) The signal value of the pixel corresponding to the unevenness does not change much from the value of fb (x, y), and is different from that of the prior art. Image (| fd (x, y) -fb
(X, y) |). Therefore, in the difference image g (x, y), the image signal fb (x, y)
The shaded part (8b) of the part has a larger value than the other parts including the unevenness. In this way, the side of the part (2b)
Only the shadow part (8e) can be extracted.

Next, a boundary line on the component side in the shadow (8e) is obtained by, for example, binarizing the image. For example, in FIG.
A point that changes from black to white when scanning in the vertical direction from the top to the bottom of the figure is extracted. That is, g (x, y) is converted to y
When the point where black changes to white in the upper half of the figure is extracted along the axis, the boundary line (9) in FIG. 4A is obtained.

Next, the boundary (9) in FIG. 4 (a) is linearized by performing Hough transform. This is shown in FIG. 4 (b). In the figure, (10) is a linearized boundary line. Next, the determination device (7) recognizes that the boundary line (10) corresponds to the side portion (2b) of the component.

Similarly, the right side (2a) and the left side (2
c) By recognizing each side of the lower side (2d) by the determination device (7), the position, center position, or rotation angle of the component (2) can be obtained. Further, the result is compared with reference values such as a reference position, a rotation angle, and a degree of coincidence between a side portion and a point sequence obtained by the Hough transform, which are set in advance by the determination device (7), to determine pass / fail. The result of the determination is displayed. The reference value stored in advance in the determination device (7) is, for example, an allowable deviation amount of the component position, an allowable rotation amount, a constraint value regarding the position of the component ridge line, or the like.

As described above, the maximum value image can be used to obtain an image in which the image values are not uneven due to the unevenness of the surface of the component (2). By taking the difference, it is possible to accurately measure the shadow of the recognized side of the component. Since the position, shape, or other geometrical feature of the component is recognized from the shadow, the component can be reliably and accurately recognized.

In the above embodiment, the difference image is binarized by the arithmetic unit (6) to obtain a binarized image, Hough transform is performed only on the component-side edge of the binarized image, and the edge of the component (2) is obtained. Although the candidate points of the part (2d) are linearized and the boundary between the part (2) and the background is determined to recognize the position, shape or other geometric feature of the part (2), the Hough transform is used. What can be recognized without doing so is not limited to this.

Further, the number of light sources is not limited to four, and the positions at which the light sources are installed do not have to be 90 ° as in the above-described embodiment. This can be achieved by arranging the shape of the component to be inspected so that its side can be easily recognized.

In the above embodiment, the maximum value image calculation for obtaining the maximum value image while leaving the larger one of the two image signals, and the difference image calculation for obtaining the difference image between the image signal of the image of the recognized side and the maximum value image signal are performed. Although they are configured to be performed by the same arithmetic unit, the same effects as when each is performed by a dedicated device can be obtained.

[Effects of the Invention] As described above, according to the present invention, three or more light sources capable of forming shadows on different sides of a component having a side, an image of a side of the component due to irradiation from the light source A photoelectric converter that captures an image and converts it into an electric signal, an image memory that stores the electric signal output from the photoelectric converter as an image signal, and two image signals of a shadow other than the recognized side of the side of the component. A maximum value image calculation device that obtains the maximum value image signal by comparing the signal values of the corresponding pixels of the image signals and leaving the larger one, the density of the image signal of the recognized side and the density of the maximum value image signal By providing a difference image calculation device that calculates a difference and obtains a difference image to recognize a recognized side portion, and a determination device that recognizes the shape of a component from a calculation result of the calculation device,
It is possible to obtain an image in which the unevenness of density due to the unevenness of the component surface has a smaller signal value than in the past, it is possible to accurately measure the shadow of the component, and from this shadow, the position of the component,
Since the shape or other geometrical features are recognized, there is an effect that a component recognition device that can reliably and accurately recognize the position of the component can be obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram showing the configuration of a component recognition apparatus according to an embodiment of the present invention, and FIGS. 2 (a) to 2 (d) relate to an embodiment of the present invention when components are illuminated by respective light sources. FIG. 2 (a) is an image of a component illuminated by a light source (3a), FIG. 2 (b) is an image of a component illuminated by a light source (3b), FIG. FIG. 2C is an explanatory diagram showing an image when the component is irradiated by the light source (3c), and FIG. 2D is an explanatory diagram showing an image when the component is irradiated by the light source (3d). FIG. 3A is an explanatory diagram showing a maximum value image according to this embodiment, FIG. 3B is an explanatory diagram showing a difference image, and FIG. 4A is a side portion of a component according to this embodiment. FIG. 4 (b) is an explanatory diagram in which sides of FIG. 4 (a) are linearized, FIG. 5 is a block diagram showing a configuration of a conventional component recognition device, and FIG. 6 (a). ) ~
FIG. 6D is an explanatory view of a conventional apparatus when a component is illuminated by a light source. FIG. 6A is an explanatory view showing a side surface of the component.
Figure (b) shows the image when the part is illuminated by the light source (3a).
Figure (c) shows the image when the part is illuminated by the light source (3b).
FIG. 4D is an explanatory diagram showing a difference image. (2) ... parts, (2a), (2b), (2c), (2d) ...
Side, (3a), (3b), (3c), (3d) ... light source,
(4) ... photoelectric converter, (5a), (5b), (5c), (5
d) image memory (6) arithmetic unit (7)
Judging device, (8a), (8b), (8c), (8d), (8e) ...
… Shadow of the part, (9) …… Candidate point of the side of the part, (10)…
... Side of the linearized part. In the drawings, the same reference numerals indicate the same or corresponding parts.

Claims (1)

(57) [Claims] 1. A light source comprising: three or more light sources capable of forming shadows on different sides of a component having a side portion; and a photoelectric device for capturing a shadow image of the side portion of the component by irradiation of the light source and converting the image to an electric signal. A converter, an image memory for storing an electric signal output from the photoelectric converter as an image signal, and two image signals obtained by capturing an image of a side of the side of the component other than the recognized side. A maximum value image calculation device for obtaining a maximum value image signal by comparing signal values of corresponding pixels of the image signals and leaving a larger one, an image signal of the image of the recognized side and the maximum value image signal And a component recognizing device comprising: a difference image calculating device that obtains a difference image by calculating a difference between signal values of the component and a determining device that recognizes a position of the component from a calculation result of the calculating device.