JPH05187826A - Detecting method of position - Google Patents

Abstract

PURPOSE:To dispense with a frame memory and to make it unnecessary to connect partial images by the frame memory, by obtaining a circumscribed quadrilateral of the whole measuring object on the basis of the data of the circumscribed quadrilaterals of every divided image. CONSTITUTION:A measuring area S where an electronic part 3 is mounted on an XY table 2 is divided according to a specific formula so that frames F of the size of a view field of a photographing camera 5 are overlapped in the X, Y two directions. The divided positions are preliminarily registered in a computer 8. The table 2 is moved in accordance with the registered position thereby to move the camera 5 relatively to each of the divided positions of the area S. An image recognizing device 7 operates a circumscribed quadrilateral of the partial pattern of the photographed screen through image processing, and the computer 8 inputs tone circumscribed quadrilateral as the data of partial circumscribed quadrilaterals. This procedure is repeated for the number of the divided positions. Then, a circumscribed quadrilateral of the whole of the subject pattern is operated from the circumscribed quadrilaterals at each photographing position. The center of circumscription is obtained from this circumscribed quadrilateral.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a position detecting method,
More specifically, the present invention relates to a position detection method of capturing an image of a measurement target on a plurality of screens and performing image processing on a pattern of the measurement target captured on the plurality of screens to detect a circumscribing center.
It is used for visual inspection and measurement of electronic components before they are mounted on the circuit board and electronic components after they are mounted on the circuit board.

[0002]

2. Description of the Related Art In a conventional position detecting method, an object to be measured is sequentially divided and imaged by relative movement with an image pickup device, and an image picked up each time is taken into a frame memory, and an image of the object to be divided is picked up. When all are stored in the frame memory, the images captured in each frame memory are joined together to obtain the pattern of the entire measurement target, and the circumscribing center is detected by performing image processing on this pattern.

[0003]

However, the above-mentioned conventional method is expensive because it requires as many frame memories as the number of divided images of the measurement object.

There is also a problem that the processing time becomes long because the images of the number of divided images are joined together.

In view of this, the present invention obtains a circumscribed quadrangle for each image obtained by divisional imaging, and on the basis of this information, the circumscribed quadrangle of the entire measurement target is obtained so that position detection can be performed. An object of the present invention is to provide a position detection method that can omit the process of joining a memory and partial images by the memory.

[0006]

In order to achieve the above-mentioned object, the present invention takes an image of a measurement target on a plurality of screens, performs image processing on the pattern of the measurement target imaged on the plurality of screens, and performs circumscribed centering. In the position detection method for detecting, the measurement range for position detection is sequentially divided and imaged by relative movement with the image pickup device, and information of each image pickup position and a partial image on the screen imaged at each image pickup position. The circumscribed quadrangle of the entire measurement target pattern is calculated by using the information of the circumscribed quadrangle obtained by performing image processing on the measurement target pattern, and the circumscribing center of the measurement target is calculated from the circumscribed quadrangle of the entire measurement target pattern. It is characterized by having done.

The measurement range can be set according to the size and shape of the object to be measured, or the amount of deviation such as the allowable position and orientation.

Imaging at the imaging position where the polygonal corner of the measurement object exists in the measurement range is given priority over the other, and the size error of the measurement object is detected from these image processing information, and there is a size error. It is also possible to omit imaging at the remaining imaging positions.

[0009]

According to the above configuration of the present invention, the measurement range for position detection is sequentially divided and imaged by relative movement with the image pickup device, but partial measurement on the screen imaged at each image pickup position is performed. Image processing is performed on the target pattern to obtain a partial circumscribed quadrangle, and the circumscribed quadrangle of the entire measurement target pattern is obtained by calculation based on the information of each photographing position and the information of each partial circumscribed quadrangle. The circumscribed quadrangle of the entire measurement target can be obtained only by storing the imaging position and the partial circumscribed quadrangle information, and the circumscribed center of the measurement target can be detected from the circumscribed quadrangle of the entire measurement target.

By setting the measurement range according to the size and shape of the object to be measured, or the amount of deviation such as the allowable position and orientation, it is possible to handle objects of any size and shape. Even if there is a deviation in position or orientation, it can be accommodated within the allowable range, and position detection can be efficiently and appropriately achieved by the number of imagings that is sufficient.

Imaging at the imaging position where the polygonal corner of the measurement object exists in the measurement range is given priority over the other, and the size error of the measurement object is detected from these image processing information, and there is a size error. By omitting the imaging at the remaining imaging positions, it is possible to avoid useless measurement of the measurement target having a size error.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Some embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 shows an example of a position detecting device for detecting a position by the method of the present invention, in which an XY table 2 supported by an orthogonal robot 1 having a drive mechanism in orthogonal X and Y directions.
The electronic parts 3 and the like to be measured are placed on the XY table 2.

An image pickup camera 5 is installed together with an illumination lamp 6 on the XY table 2, and X, Y2 of the XY table 2 are arranged.
Due to the movement in the direction, the imaging camera 5 and the illumination lamp 6 are relatively moved so as to face any position on the upper surface of the XY table 2.

The image pickup camera 5 is connected to the image recognition device 7, and the image recognition device 7 is controlled by exchanging signals with the computer 8.

A specific method of position detection in the first embodiment of the present invention will be described below.

First, as shown in FIG. 2, the measurement range S of the portion where the electronic component 3 is placed on the XY table 2 by the field-of-view size frame F of the image pickup camera 5 in the X and Y2 directions is calculated according to the following formula. The fields of view size F are divided so that they overlap, and the positions are: X (k) = x + (DX1) * (i)-(measurement range X direction) Y (k) = y + (DY1) * (j)- (Measurement range Y direction) However, (x, y) is the target center position of the measurement target DX1 = (size in the X direction of the visual field size frame F of the imaging camera) − (overlap length X direction of the visual field) DY1 = (of the imaging camera) Size of field of view size frame F in Y direction)-(overlap length of field of view in Y direction) k = 1, 2, ..., N1 i = 1, 2, ..., M1 j = 1, 2 , ... m2 m1 = (measurement range X direction) / DX1 m2 = (measurement range Y direction) ) / DY1 and is registered in the computer 8 in advance.

XY table 2 according to the position registered next
Are moved to relatively move the imaging camera 5 to each division position of the measurement range S in a predetermined order.

At the time of imaging at each division position, a circumscribing quadrangle A for a partial pattern 3a of the electronic component 3 placed on the imaged screen as shown in FIG. 3 is calculated by image processing in the image recognition device 7. , Is input and stored in the computer 8 as partial circumscribed quadrangle information.

When the image pickup and the image processing are completed by the number of the respective positions divided in the measurement range S, NO. The circumscribed quadrangle at each imaging position of k is converted into the coordinate system of the entire image processing range.

Image pickup position NO. In the coordinate system of the entire image processing range. k is [Xmin (1, k)] = [Xmin (k)] + (DX1) * (i) [Ymin (1, k)] = [Ymin (k)] + (DY1) * (j) [ Xmax (1, k)] = [Xmax (k)] + (DX1) * (i) [Ymax (1, k)] = [Ymax (k)] + (DY1) * (j) where DX1 = ( X-direction size of field-of-view size frame F of imaging camera- (overlap length X-direction of field of view) DY1 = (size of field-of-view frame F of imaging camera in Y-direction)-(field-of-view overlap length Y direction) k = 1 It is calculated that i, j = 1, 2, ...

Next, at each of the image pickup positions NO. Partial circumscribed quadrangle data Xmin (1,1), Xmin at k
(1,2), Xmin (1,3), ... Xmin
The smallest value (Xmin) of (1, n) is set as the circumscribed quadrangle (Xmin) of the electronic component 3 to be measured, and Y
min (1,1), Ymin (1,2), Ymin
The smallest value (Ymin) of (1,3), ..., Ymin (1, n) is set as the circumscribed rectangle (Ymin) of the electronic component 3 to be measured, and Xmax (1,1), Xma
x (1,2), Xmax (1,3), ... Xmax
The largest value (Xmax) of (1, n) is set as the circumscribed quadrangle (Xmax) of the electronic component 3 to be measured, and Y
max (1,1), Ymax (1,2), Ymax
The smallest value (Ymax) of (1,3), ..., Ymax (1, n) is set as the circumscribed quadrangle (Ymax) of the electronic component 3 that is the measurement target.

As a result, the circumscribed quadrangle B of the entire measurement target pattern 3b as shown in FIG. 4 becomes (Xmin, Ymi
n) − (Xmax, Ymax), and the circumscribing center of the electronic component 3 to be measured is (XC) = [(Xmin) + (Xmax)] / 2 (YC) = [ It is obtained by calculating (Ymin) + (Ymax)] / 2.

FIG. 5 is a flowchart showing this.

6 and 7 show a second embodiment of the present invention, in which the setting of the measurement range S is set according to the size and shape of the electronic component 3, and the measurement set in this way is performed. When the range S is divided by the field-of-view size frame F of the image pickup camera 5, the number of image pickup screens becomes the minimum necessary amount without excess or deficiency, and position detection can be performed efficiently while supporting any size and shape. It can be done properly.

More specifically, the image pickup positions to be divided are: X (k) = x + (DX1) * (i)-(measurement target X-direction size) Y (k) = y + (DY1) * (j) -(Measurement target Y-direction size) However, DX1 = (size of view field size frame F of the image pickup camera in X direction)-(overlap length X direction of view field) DY1 = (size of view field size frame F of image pickup camera in Y direction) -(Overlapping length of visual field in Y direction) k = 1, 2, ..., N2 i = 1, 2, ..., M1 j = 1, 2, ..., M2 m1 = (Measurement target X direction size) / DX1 m2 = (measurement target Y direction size) / DY1 is calculated and registered in the computer 8 in advance.

XY table 2 according to the position registered next
Are moved to relatively move the imaging camera 5 to each division position of the measurement range S in a predetermined order.

When the image pickup and the image processing for the respective divided positions are completed, as in the case of the first embodiment, NO. The circumscribed quadrangle at each image pickup position of k is converted into the coordinate system of the entire image processing range, and the subsequent processing is performed in the same manner as in the first embodiment.

The processing in this case is shown in a flow chart in FIG.

FIGS. 8 and 9 show a third embodiment of the present invention, in which the measurement range S is set corresponding to the parallel displacement and rotational displacement permitted by the electronic component 3. The number of imaging screens when the measurement range S set to is divided by the field-of-view size frame F of the imaging camera 5 is set to the minimum necessary amount without excess or deficiency, and also tolerates parallel deviation and rotation deviation. Position detection can be performed more efficiently and properly.

This will be described in detail. Parallel deviation amount = DX2, DY2 Rotational deviation amount = DQ Allowable range = DK Measurement target X direction size = LX Measurement target Y direction size = LY Imaging camera position X (k), Y (K) (k = 1, 2, ... N3), X (k)> DX2 or X (k)> DK,
Alternatively, the imaging position NO. Which satisfies Y (k)> (LY) * sin (DQ). Since k is deleted and the imaging and processing of this is omitted, the number of times of imaging is further reduced as shown in FIG. 8, and the size, shape, position and orientation of the measuring object such as the electronic component 3 are reduced. Appropriate position detection corresponding to the shift becomes possible.

The flow chart of the processing in this case is as shown in FIG.

FIG. 10 and FIG. 11 show a fourth embodiment of the present invention, in which the imaging position where the corner shape of the measuring object, such as the polygonal electronic component 3 within the set measuring range S, exists. The image of the measurement target is set too large or too small depending on whether the corners of the measurement target pattern are detected in the captured image at this position. Whether there is a size error or not is determined, and if there is a size error, the subsequent measurement is stopped so that the measurement is not performed unnecessarily.

When stopping the measurement, it is practically convenient to display a warning or to discard the size error measurement target.

The processing of this embodiment will be described in detail below.
An imaging position where a polygonal measurement target corner portion within the set measurement range S exists for imaging each divided position of the measurement range S in each of the first to third embodiments and performing image processing. In FIG. Set as shown in 1-18. These positions are the imaging position NO. 1 X (k) = x− (LX) / 2 Y (k) = y + (LY) / 2 Imaging position NO. 2 X (k) = x− (LX) / 2 Y (k) = y− (LY) / 2 Imaging position NO. 3 X (k) = x + (LX) / 2 Y (k) = y− (LY) / 2 Imaging position NO. 4 X (k) = x + (LX) / 2 Y (k) = y + (LY) / 2 Imaging position NO. k X (k) = x + (LX) * (i) Y (k) = y + (LY) * (j) However, the imaging position NO. in the first to third embodiments. Not common with k.

(X, y) is the target center position of the measurement target.

DX1 = (size of view field size frame F of imaging camera in X direction) − (overlap length of view in X direction) DY1 = (size of view size frame F of imaging camera in Y direction) − (overlap length of view) Y direction) k = 1,2, ... n4 i = 1,2, ... m1 j = 1,2, ... m2 m1 = (measurement target X direction size) ) / DX1 m2 = (measurement target Y direction size) / DY1 is calculated and registered in the computer 8.

Then, with respect to these image pickup positions, the image pickup and the image processing are performed in preference to the other image pickup positions, and the size error is discriminated.

If the size error is not discriminated, the image is picked up at another image pick-up position and processed in the same manner as in the first to third embodiments.

The flow chart of the processing in this case is as shown in FIG.

[0041]

According to the present invention, the measurement range for position detection is sequentially divided and imaged by relative movement with the image pickup device, but a partial measurement target on the screen imaged at each image pickup position. Image processing is performed on the pattern to obtain a partial circumscribed quadrangle, and the circumscribed quadrangle of the entire measurement target pattern is calculated and obtained by the information of each photographing position and the information of the partial circumscribed quadrangle corresponding to each photographing position. The circumscribed quadrangle of the entire measurement target is obtained only by storing the imaging position and the partial circumscribed quadrangle information, and the circumscribed center of the measurement target is detected from the circumscribed quadrangle of the entire measurement target, so that the measurement range is divided. Since the frame memory is not required for the number of images to be taken, the processing can be performed by a simple device, and the process for connecting the divided image pickup screens stored in the frame memory is not necessary, so that the control is easy and the security is low. Processing time is shortened order to achieve to.

By setting the measurement range according to the size and shape of the object to be measured, or the amount of deviation such as the allowable position and orientation, it is possible to handle objects of any size and shape. Even if there is a deviation in position or orientation, it can be handled as long as it is within the allowable range, and position detection can be efficiently and properly achieved with just enough number of times of imaging, and processing time is further shortened while ensuring detection reliability. To do.

Imaging at the imaging position where the polygonal corner of the measurement object exists in the measurement range is given priority over the other, and the size error of the measurement object is detected from these image processing information, and there is a size error. By omitting the imaging at the remaining imaging positions, it is possible to avoid useless measurement of the measurement target having a size error, and further speed up the process.

[Brief description of drawings]

FIG. 1 is a side view and a plan view showing an example of a position detection device used in the present invention.

FIG. 2 is an explanatory diagram showing a state of divided image pickup positions in a measurement range according to the first embodiment of the present invention.

FIG. 3 is an explanatory diagram showing a partial measurement target pattern and a partial circumscribing quadrangle on an image pickup screen at divided image pickup positions.

FIG. 4 is an explanatory diagram showing a state of a circumscribed quadrangle of the entire measurement target obtained by calculation from the partial circumscribed quadrangle of FIG. 3;

FIG. 5 is a flowchart showing a specific process of the first embodiment of the present invention.

FIG. 6 is an explanatory diagram showing a state of divided imaging positions when the measurement range is set according to the size and shape of the measurement target in the second embodiment of the present invention.

FIG. 7 is a flowchart showing a specific process of the second embodiment of the present invention.

FIG. 8 is an explanatory diagram showing a state of divided image pickup positions when a measurement range in the third embodiment of the present invention is set corresponding to a parallel displacement amount and a rotational displacement amount allowed by a measurement target.

FIG. 9 is a flowchart showing a specific process of the third embodiment of the present invention.

FIG. 10 shows a divided image pickup position setting state when a divided image pickup position in which a corner portion of a polygonal measurement target exists in a measurement range in the fourth embodiment of the present invention and the image pickup is preferentially performed. FIG.

FIG. 11 is a flowchart showing a specific process of the fourth embodiment of the present invention.

[Explanation of symbols]

 1 Cartesian Robot 3 Electronic Component 5 Imaging Camera 7 Image Recognition Device 8 Computer A Partial Circumscribed Rectangle B B Whole Circumscribed Rectangle S Measurement Range

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Masanori Yasutake 1006 Kadoma, Kadoma City, Osaka Prefecture Matsushita Electric Industrial Co., Ltd.

Claims (3)

[Claims] 1. A position detecting method for detecting a circumscribing center by image-processing a measurement target on a plurality of screens and image-processing the pattern of the measurement target imaged on the plurality of screens. Divided images are sequentially taken by relative movement with the image pickup device, and measured by information of each image pickup position and information of a circumscribed quadrangle obtained by performing image processing on a partial measurement target pattern on the screen imaged at each image pickup position. A position detecting method, characterized in that a circumscribing quadrangle of the entire target pattern is obtained by calculation, and the circumscribing center of the measurement target is detected from the circumscribing quadrangle of the entire measurement target pattern. 2. The position detection method according to claim 1, wherein the measurement range is set according to the size and shape of the measurement target, or the amount of deviation such as the allowable position and orientation. 3. Prioritizing the imaging at an imaging position where a polygonal corner of the measurement target exists in the measurement range,
The position detecting method according to claim 1, wherein a size error of the measurement target is detected from the image processing information, and if there is a size error, the imaging at the remaining imaging positions is omitted.