JPS62287108A - Measuring method for position of electronic parts - Google Patents

Abstract

PURPOSE:To eliminate the need for rough positioning by extracting a characteristic pattern corresponding to the characteristic shape of the pin array of an electronic parts from an image obtained by a visual device and measuring the position of the electronic parts on a substrate from the position of the pattern. CONSTITUTION:A flat package IC10 and the substrate 12 are put in the visual field of the visual device 13 to detect the array of pins 14 of the flat package IC10 and an array of lands 15 of the substrate 12, and straight lines 16, 17, 18, and 19 are found from the pin array or land array to perform the positioning. Then the image of the substrate 12 including the lands 15 is picked up by a camera 26 for the substrate 26 and processed by the image processing part 24 to find the positions of the lands 15 by a robot coordinate system. Then, a hand 25 is moved onto a camera 27 for the IC and pick its image up and the image is processed by the processing part 24 to find the position shift between the IC10 and hand 25 by the robot coordinate system. Lastly, the hand 25 is moved and lowered vertically, thereby mounting the IC10 on the lands 15.

Description

[Detailed Description of the Invention] V. Detailed Description of the Invention [Field of Industrial Application] The present invention relates to an automatic component mounting technique for imposition, and in particular, an electronic component position measuring method for measuring the position of a component. Regarding.

[Conventional technology]

When mounting electronic components such as parts on a printed circuit board automatically, imposition measures the position of the electronic component held by the robot hand and the position of the printed circuit board into which it will be inserted, and aligns the two. while inserting electronic components into predetermined locations on the printed circuit board.

Conventionally, as shown in FIG. 8(b), for example, an upright mark 2 is provided at a predetermined location on a printed circuit board 10, and the printed circuit board 1 is observed with a visual device 3 such as a camera.
As shown in a), by measuring the position of the positioning mark 2 within the frame of the measurement range 4, alignment with the electronic component is performed.

Regarding the conventional electronic component position measurement method,
For example, there is Japanese Unexamined Patent Publication No. 1988-1900.

[Problem that the invention seeks to solve]

The idea of attaching positioning marks to the board and measuring its position assumes that the position of the measurement target is roughly determined in advance, and that the measurement target will always fall within the fixed measurement range in the image. .

However, the above premise is that as electronic components become smaller,
It becomes difficult to establish. Furthermore, mechanically regulating the position of electronic components and circuit boards involves difficulties in terms of precision, but as electronic components become smaller, it becomes increasingly impossible to mechanically regulate the position. Therefore, in order to promote automatic insertion of electronic components, which are becoming increasingly smaller, there is a desire to develop a method for measuring the position of electronic components that is different from conventional methods, that is, does not require rough positioning.

An object of the present invention is to provide an electronic component position measuring method that does not require rough positioning.

[Means for solving problems]

The above object is to provide an automatic electronic component mounting device that is equipped with a visual device and automatically inserts electronic components onto a board based on images of pin rows of electronic components and land rows of a board captured by the visual device. This is achieved by extracting a unique pattern corresponding to the pin row shape of the electronic component from the captured image, and aligning the electronic component and the insertion location of the board based on the extracted pattern position.

[Effect]

An image of an electronic component is captured by a visual device, and the unique shape of the pin array of the electronic component (for example, a flat package IC is
It has a shape with rows of contact pins lined up on all four circumferences. ), the position of the electronic component is measured. On the other hand, the land array on the board is captured by a visual device, a pattern with a shape corresponding to the shape of the electronic component is extracted from the image, and the position of the board is measured from the pattern. With this, the relative position of the electronic component with respect to the board is measured, and by aligning them vertically, the electronic component can be smoothly inserted into the corresponding location on the board.

〔Example〕

An embodiment of the present invention will be described below with reference to FIGS. 1 to 7.

In the following description of the embodiment, a case will be described in which a flat package IC having a row of connection bins protruding from its four sides is automatically inserted into a board.

In an image of a flat package IC or a land on a substrate, pins and contacts are bright areas 11, as shown in FIG. 2, and are easy to detect. These regions 11 have no distinctive shape when used alone, and are difficult to identify because they all have the same shape. Further, since other patterns such as board wiring (not shown) are included in the image, identification is even more difficult. However, when viewed as a whole, it can be seen that the regular arrangement of the regions 11 is characteristic. That is, the center of gravity of the area 110 (
The rows of lines (indicated by x marks) are parallel or orthogonal. Therefore, in this embodiment, as shown in FIG.
The rows of lands 15 of No. 2 are detected, and straight lines i6, 17, 18, 19 are determined from the obtained pin rows or land rows, and alignment is performed.

FIG. 3 is an external view of the electronic component mounting device. This device is
Frame 20. Orthogonal robot, 7) 21, board transfer unit 2
2, a parts supply section 23, and an image processing section 24. The orthogonal robot 21 has a hand 25, and a substrate camera 26 is attached to the hand 25 facing downward, and an IC camera 27 is attached to the upper part of the pedestal 20 facing upward.

The board transport section 22 and the component supply section 23 each supply a pallet containing a board and an IC, and fix the pallet at a fixed position. The substrate is fixed on the pedestal 1, and the IC is vacuum-adsorbed by the hand 25. Therefore, the fixed position of the board is precisely measured by the image processing section 24 from an image captured by the board camera 26, as will be described in detail later.

On the other hand, after the hand 25 chucks the IC, the hand 2
5 on the IC camera 27 (i), and the relative position of rC with respect to the hand 25 is determined by the camera 2 as described in detail later.
The image processing unit 24 performs precise measurements based on the images from 7.

An overview of the operation of this electronic component mounting apparatus will be explained with reference to the flowchart shown in FIG.

First, a substrate is transported by the substrate transport section 22 and mechanically fixed in a fixed vertical position. Next, the hand 25 moves,
It is stopped at a position where the land on the board can be photographed by the board camera 26. Next, an image of the board including the land is captured by the board camera 26, and this image is processed by the image processing unit 24 as will be described later.
(, +θL) are found in the robot coordinate system.

Next, the hand 25 is moved onto the pallet and chucks the IC. Next, the hand 25 is moved onto the IC camera 27, and an image of the IC is taken here.

This image is processed by the image processing unit 24 as described later, and the positional deviation (x, yx, θr) between the IC and the hand is determined in the robot coordinate system. Finally, when the hand 25 is moved to (x, -x, YL-Yl, θ, -θ1) and lowered vertically, the IC is correctly mounted on the land.

The details of the image processing in the image processing section described above will be explained with reference to the flowchart of FIG.

First, an image is input and regionalized so that the region 11 becomes clear as shown in FIG. Next, find the center of gravity of all regions and apply Hough K transformation to the coordinates of the center of gravity.

When there is a subset of points aligned on a straight line in a plane,
Hough transformation is a method for finding this straight line. below,
Let me explain this. The equation on the straight line shown in Figure 6 is cosθ-X + sinθ#y-r = o
--Represented by (1). As can be seen from this formula,
The equation of a straight line can be expressed by two parameters, r and θ. The condition for this straight line to pass through the point (XI'yj) is cosθx, +5in19ly, -r=o
-...(2). This is X1+'f
When drawn in r-θ space with l as a constant, it becomes as shown in FIG. 7(b). If the X-Y coordinate system in the screen is defined as shown in Figure 7(a) -6sx<r<23X, -180°
〈θ = 180°, it can be contained in a finite space. This transformation from coordinate space to parameter space is called Hough transformation. References include ``Image Recognition Theory'' by Makoto Nagao (
(Corona Inc.) p72-74. Points in parameter space correspond to straight lines in coordinate space. Figure 7 (C)
A sequence of points A and B that can be considered as a straight line in the coordinate space.

When there are C and D, the curves A, B, C, and D in the parameter space corresponding to each point intersect at one point P as shown in FIG. 7(d). Therefore, by searching for points where many curves intersect in the parameter space, we can create a linear point sequence in the coordinate space,
You can find the equation of that straight line. To find the intersection of curves, the parameter space is divided into fine grids, the number of times the curve passes through each grid section is counted, and the section with the most number of passes can be determined to correspond to a linear point sequence.

Therefore, returning to FIG. 5, after Rough transformation is applied to the barycenter coordinates, the parameter space is divided into fine grid sections, and sections in which the number of curve passages is greater than a predetermined value are extracted. However, the predetermined value is predetermined from the number of bins and the number of land contacts for each column. Median value of the determined interval (rl, θ,) (1=1.
2. ...) is a parameter of a straight line. Among the many straight lines thus determined, there are certain conditions for the IC bin rows and land contact rows.

In other words, there are two parallel lines, and the sets of parallel lines are orthogonal to each other. Also, the distance between parallel lines = d! /i becomes a constant value, etc. According to these conditions, only bin rows or contact rows are selected from among the straight lines. The position has now been determined, but in order to use it as data, it must be determined in the form of coordinates of the representative point. So, for example, the intersection of the median lines of parallel lines (X.

Y), and the angle θ that the bisector of the median line forms with the X axis is determined as the overall coordinates.

In the explanation of Fig. 4, the positions of the lands and ICs are determined in the robot coordinate system, but in order to do so, they must first be determined in the visual coordinate system, and then coordinate transformation is performed based on the position constants and optical constants specific to each camera. It is necessary to do this.

As explained above, according to this embodiment, the position can be measured no matter where the image of the IC or land, which is the object of vertical measurement, is located on the screen. There is no need to perform rough positioning in advance for positioning, and the position can be measured even when it is difficult to place the object to be measured within the measurement range by coarse positioning.

〔Effect of the invention〕

According to the present invention, since the position is measured by extracting a pattern corresponding to the unique shape of the electronic component from the photographed image, rough positioning is not necessary, and high accuracy is not required for the mechanical fixing mechanism.

[Brief explanation of the drawing]

Fig. 1 is an explanatory diagram of an embodiment of the present invention, Fig. 2 is a captured image of a flat package IC, Fig. 3 is an external view of an electronic component mounting device, Fig. 4 is a flowchart of a component mounting process, and Fig. 5 is an image processing flowchart, FIG. 6 is a graph defining a straight line, and FIG. 7(a). (b), (c), (d) are Hough
Conversion explanatory diagram, FIG. 8(a). (b) is an explanatory diagram of a conventional method. 10... Flat package IC, 11... Area,
12... Board, 13... Camera, 14... Bin,
15... Land, 16.17, 18.19... Straight line, 20... Frame, 21... Cartesian robot, 22.
... Board transport unit, 23... Component supply unit, 24... Image processing unit, 25... Hand, 26... Board camera, 27... Camera for work C.

Claims (1)

[Scope of Claims] 1. An electronic component automatic mounting device that includes a visual device and automatically inserts electronic components onto a board based on images of pin rows of the electronic component and land rows of the board captured by the visual device, Electronic component positioning, characterized in that a unique pattern corresponding to the unique shape of a pin row of an electronic component is extracted from an image captured by a visual device, and the position of the electronic component with respect to a board is measured based on the position of the unique pattern. Measuring method. 2. The electronic component position according to claim 1, characterized in that the image is Hough-transformed to extract straight line patterns, and from among the set of straight line patterns, a set that satisfies a predetermined condition is extracted as the unique pattern. Measuring method.