JPH10261898A - Positioning method for work - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a positioning method for work in which the feature part can be located in a short time in positioning a work while suppressing the location miss. SOLUTION: In a method for locating a feature part, e.g. a location mark, on an image, a large and a small search areas are preset and pattern matching is performed in the small search area S1 at first and then pattern matching is performed in the large search area S2 if the matching rate is lower than a set threshold. According to the method, the feature part can be located in a short time in most case where the positional variation is insignificant and location miss can be suppressed when the variation is significant. When the feature part can not be located even in the large search area S2, the work and a camera are shifted relatively so that a wide range can be searched and thereby the positioning system is prevented from being stopped due to location miss even when the variation is more significant.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of positioning a work such as a substrate by image recognition.

[0002]

2. Description of the Related Art As a method of positioning a work such as a substrate, a characteristic portion such as a recognition mark of the work is captured by a camera and captured as an image, the position of the characteristic portion is recognized on the image, and based on the recognition result. 2. Description of the Related Art A method of positioning a workpiece by using a conventional method is widely known. When performing this recognition, the time required for recognition is reduced by setting a search area in the screen and limiting the image processing range as much as possible, instead of performing image processing on the entire screen area captured by the camera. Is done. The search area is set to a range in which a feature to be recognized exists with a high probability. Generally, if the search area is set to be narrower, recognition can be performed in a shorter time.

[0003]

However, when the search area is narrowed, a feature to be recognized is out of the search area due to the dimensional accuracy of the work itself and a variation in the position at the time of transfer of the work. Sometimes. In this case, since the area outside the search area is not a target of image processing, the characteristic portion is not recognized, and therefore, the work cannot be positioned. Conventionally, in such a case, the positioning device is stopped and a recognition error is reported. Then, every time a recognition error occurs, there is a problem that the operator must take necessary measures by manual operation based on the notification.

Therefore, the present invention can solve the above problems,
It is an object of the present invention to provide a work positioning method that can recognize a characteristic portion at the time of work positioning in a short time and can reduce occurrence of recognition errors.

[0005]

According to a first aspect of the present invention, there is provided a method of positioning a work, comprising the steps of: obtaining a characteristic portion of a work placed on a movable table from image data obtained by imaging with a camera; Is a work positioning method in which a movable table is driven based on a result of the recognition by pattern matching, and a movable table is driven to position the work. In the image data screen, two types of large and small search areas are set, First, a matching ratio in a small search area is obtained, and the highest value of the matching ratio is compared with a preset threshold value. If the maximum value is less than the threshold value, the search area is switched to find the matching rate in the large search area, and the maximum value of this matching rate is set in advance. Compared to threshold, the maximum value is to recognize that they provide the highest value if exceeding the threshold as the position of the feature.

According to a second aspect of the present invention, if the maximum value of the matching ratio in a large search area is equal to or less than a predetermined threshold value and a characteristic portion cannot be detected, the predetermined value is set in advance. The work is relatively moved with respect to the camera according to the movement path, and a search operation for searching for a characteristic portion by performing pattern matching for each imaging position set on the movement path is performed.

[0007]

According to a first aspect of the present invention, a large and small search area is set in advance, and pattern matching is first performed in a small search area. If no feature exists in the small search area, the matching area is expanded to a large search area. After that, pattern matching is performed in this large search area to detect the position of the characteristic portion. For this reason, the position can be recognized in a short time when the variation in the position of the characteristic portion is small, and the occurrence of recognition errors can be reduced by expanding the search area even when the variation is large.

According to the work positioning method of the present invention, when the search area cannot be recognized even if the search area is enlarged, the work and the camera are relatively moved to search a wide range. Even when the variation is even greater, the positioning device hardly stops due to a recognition error.

An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a perspective view of a work positioning device according to an embodiment of the present invention, FIG. 2 is a block diagram of a control system of the work positioning device, and FIG. 3 is an image diagram showing a search area of the work positioning device. FIG. 4 is a flowchart showing the operation of the work positioning device, and FIG. 5 is a partial plan view of the work positioning device.

First, the overall structure of a work positioning device will be described with reference to FIG. In FIG. 1, reference numeral 1 denotes a movable table including an X table 2 and a Y table 3. On the movable table 1, a substrate holding unit 4 is placed. Reference numeral 5 denotes a substrate as a work, which is held on the substrate holding unit 4. On the upper surface of the substrate 5, two recognition marks A,
B is provided. Cameras 6a and 6b for observing the recognition marks A and B are provided above the substrate holder 4. A mounting head 7 is provided above the substrate holding unit 4. The mounting head 7 mounts the chip 9 that has been vacuum-adsorbed to the lower end of the nozzle 8 on the substrate 5.

Next, the configuration of a control system of the work positioning device will be described with reference to FIG. In FIG. 2, cameras 6a and 6b take images of the recognition marks A and B, and send image data to an image storage unit 11 via an A / D conversion unit 10. The image storage unit 11 stores the A / D converted image data.
The reference pattern (reference pattern) storage unit 12
The reference patterns of the recognition marks A and B serving as references for comparison and collation are stored. The comparison pattern extraction unit 13 extracts, from the image data in the image storage unit 11, image data corresponding to the range of the comparison pattern to be compared with the reference pattern. The pattern matching unit 14 matches the reference pattern and the comparison pattern by pattern matching, and obtains a matching rate α.

The pattern position calculating section 15 compares the matching rate α obtained by the pattern matching section 14 with a preset threshold value. Is stored as the position coordinates (Xc, Yc) of the mark, and the main control unit 1
8 is output. If the matching rate α is equal to or smaller than the threshold value, a command to enlarge the search area is issued. The search area setting unit 16 sets the search area of the size designated by the pattern position calculation unit 15 among the two types of search areas set in advance. The pattern extraction position specification unit 17 specifies the range of the comparison pattern to be compared with the reference pattern in the search area by specifying the representative coordinates (X, Y) of the comparison pattern.

Here, the search area will be described with reference to FIG. In FIG. 3, reference numerals 30 denote cameras 6a and 6b.
Is a frame indicating the range of the screen to be imaged, and the mark A (or B) is imaged within this range. A range surrounded by a small chain line frame indicates a small search area S1, and a range surrounded by a large chain line frame indicates a large search area S2. A range surrounded by a solid frame 31 indicates an area to be extracted as a comparison pattern, and one corner point P (X, Y) indicates a coordinate point representing this area. In FIG. 3, the mark A is imaged substantially at the center of the screen, and is within the search area S1. Therefore,
By moving the position where the comparison pattern is extracted within the small search area S1, the point P giving the highest matching rate α can be obtained.

The mark A 'is located outside the search area S1 and inside the search area S2. In this case,
In the small search area S1, the position of the mark A 'cannot be recognized, and it is necessary to switch to the large search area S2. In this case, since the range in which the comparison pattern is extracted is wide, the time required for recognizing the position of the mark A ′ becomes long.

The main controller 18 calculates the amount of movement for positioning the substrate 5 with respect to the working means based on the position coordinates (Xc, Yc) of the mark obtained by the pattern extraction position designator 17. The learning unit 19 stores the position coordinates (Xc, Yc) of the mark detected for each substrate 5 in the main control unit 18.
Learning and storing the tendency of the displacement of the substrate 5,
The learning result is fed back to the main controller 18.

As will be described later, a small search area S1
If the recognition error still occurs even when the search area is expanded to a larger search area S2, the camera and the substrate 5 are relatively moved to search for a mark in a wider range. In this case, the feedback data is used to perform the search operation along the most rational route. The XY table control unit 20 includes the main control unit 1
The driving of the movable table 1 is controlled based on the movement amount for positioning calculated in step 8.

The configuration of the work positioning device is as described above, and its operation will be described below. In FIG. 4, first, the substrate transfer means 9 is driven to transfer the substrate 5 to the substrate holding unit 4 (ST1). Next, the substrate 5 is
a, and 6b, the recognition marks A,
B is in the field of view of the cameras 6a and 6b (ST
2) The images of the marks A and B are taken by the cameras 6a and 6b (ST3). Next, the image data obtained by the imaging is processed by pattern matching, and the distribution of the matching rate α in a small search area (see S1 in FIG. 3) is obtained (ST4).

Next, it is determined whether or not the maximum value of the matching rate α exceeds a threshold value (ST5). here,
If the highest value of the matching rate α does not exceed the threshold value, the process proceeds to a step described later. If the highest value of the matching rate α exceeds the threshold value, the coordinates ( Xc, Yc) is output as the position of the mark (ST6).

Next, a relative movement amount for positioning the substrate 5 is calculated from the coordinate values (Xc, Yc) (ST).
7), the movable table 1 is driven based on the relative movement amount to move the substrate 5, and the positioning is completed (ST8).
Work such as mounting is performed on the thus positioned substrate 5 by the mounting head 7 serving as a working means.

Next, the case where the maximum value of the matching rate α does not exceed the threshold value in ST5 will be described. In this case, first, the search area is expanded (ST9), and the distribution of the matching rate α in the large search area (see S2 in FIG. 3) is obtained (ST10). Next, similarly to ST5, it is determined whether or not the maximum value of the matching rate α exceeds a threshold value (ST11). If it exceeds the threshold value, the process proceeds to ST6. If the matching rate is still equal to or smaller than the threshold value in this step, the search operation described below is further performed.

FIG. 5 shows an example of a path of this search operation. A mark A which is normally detected inside a large search area S2 in an image of the camera 6a is removed from the search area S2 for some reason. Off, at the position indicated by A '. In this case, the substrate 5 is moved relative to the cameras 6a and 6b in the horizontal direction so that the mark A 'is contained in the search area S2. As described above, the route of this movement is set in advance based on the learning data fed back from the learning unit 19 (see FIG. 2). In the example of FIG. 5, it is learned in advance that the tendency of the displacement of the mark A is in the inner direction of the substrate 5. In order to search a range in the inner direction, the search area S2 is indicated by an arrow a and an arrow. It shows an example of sequentially moving in the directions of b and arrow c. Of course, if the mark A is detected by the search, the search operation is stopped there.

Returning to the flow of FIG. 4, the cameras 6a and 6b are moved relative to the substrate 5 as described above (ST
12). Next, images are captured by the cameras 6a and 6b at the first image capturing position on the moving route, and the image data of the recognition marks A and B is captured. Thereafter, the same processing as in the above steps is performed to obtain the distribution of the matching rate α in a large search area (see S2 in FIG. 3) (ST13), and the maximum value of the matching rate α exceeds the threshold value again. It is determined whether or not (ST14). If the highest value of the matching rate α exceeds the threshold value, the coordinates (Xc, Yc) indicating the highest matching rate α are output as the position of the mark (ST15), and the coordinates (Xc ,
Yc) is sent to the learning unit 19 and stored (ST16).
The subsequent steps are the same as in ST7 described above.

If the maximum value of the matching ratio α does not exceed the threshold value in ST14, the process returns to ST12, and the substrate 5 is further moved relative to the cameras 6a and 6b. Thereafter, the search operation up to ST14 is repeated, and if the maximum value of the matching rate α does not exceed the threshold value even after repeating the predetermined number of times, a recognition error is notified and the positioning device stops.

[0024]

According to the work positioning method of the present invention, two large and small search areas are set in advance, and pattern matching is performed in the small search area at first. If the matching ratio is equal to or less than the set threshold, the matching area is increased. Expand to search area. For this reason, the position can be recognized in a short time in the case where the variation in the position of the characteristic portion is small,
Even when the variation is large, occurrence of recognition errors can be reduced. In addition, if recognition is still not possible even when the search area is enlarged, the workpiece and the camera are relatively moved to search a wide range. Therefore, even if the variation is further large, the positioning device hardly stops due to a recognition error.

[Brief description of the drawings]

FIG. 1 is a perspective view of a work positioning device according to an embodiment of the present invention.

FIG. 2 is a block diagram of a control system of the work positioning device according to the embodiment of the present invention;

FIG. 3 is an image diagram showing a search area of the work positioning device according to the embodiment of the present invention;

FIG. 4 is a flowchart showing an operation of the work positioning device according to the embodiment of the present invention;

FIG. 5 is a partial plan view of the work positioning device according to the embodiment of the present invention;

[Explanation of symbols]

 Reference Signs List 1 movable table 4 substrate holding unit 5 substrate 6 camera 7 mounting head (working means) 11 image storage unit 12 reference pattern storage unit 14 pattern matching unit 15 pattern position calculation unit 18 main control unit 20 movable table control unit

Claims (2)

[Claims] 1. A position of a characteristic portion is recognized by pattern matching from image data obtained by imaging a characteristic portion of a work placed on the movable table with a camera, and the movable table is driven based on the recognition result. This is a work positioning method for positioning a work by setting two types of search areas, large and small, in a screen of image data, first obtaining a matching rate in the small search area, and calculating the maximum value of the matching rate. Compare with a preset threshold,
If the highest value exceeds the threshold value, the point giving the highest value is recognized as the position of the feature, and if the highest value is equal to or less than the threshold value, the search area is switched to obtain a matching rate in a large search area. Comparing the maximum value of the matching rate with a preset threshold value, and recognizing a point giving the maximum value as the position of the characteristic portion if the maximum value exceeds the threshold value. Positioning method. 2. If the maximum value of the matching rate in a large search area is less than a predetermined threshold value and no characteristic portion can be detected, the work is sent to the camera according to a predetermined moving path. 2. The work positioning method according to claim 1, further comprising performing a search operation of searching for a characteristic part by performing pattern matching for each imaging position set on the movement path.