JP2000097677A - Image recognizing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image recognizing method in which recognition can be made stably while enhancing recognition accuracy. SOLUTION: In the method for recognizing the inclining direction and positional shift of a rectangular electronic component through image recognition thereof, image of the electronic component is picked up by means of a camera and inclination data of a vector set on the extracted outline thereof is determined (ST3). The inclination data is divided into data groups for respective sides of a polygon constituting the outline and the inclination data is corrected by an amount of included angle for the data groups of a plurality of sides in specified relationship having known included angles thus obtaining identical data groups (ST4). Subsequently, an angle histogram indicative of the inclination distribution of vector is made (ST5) and the direction of an object to be recognized is detected based on the inclination distribution (ST6). According to the method, number of data can be increased even for a small object to be recognized having small number of data of significant fluctuation and good recognition results can be obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image recognition method for recognizing a shape and a direction of an electronic component by taking an image of the electronic component and recognizing the image.

[0002]

2. Description of the Related Art In mounting electronic components on a substrate, it is general to mount the electronic components held on a mounting head on an image by recognizing the image and detecting a positional deviation, correcting the positional deviation, and mounting the electronic component on the substrate. Is being done. In this position shift detection, a contour line of the electronic component is extracted by performing image processing on image data obtained by imaging the electronic component, and a shape of the electronic component, a direction of inclination in a horizontal plane, and the like are determined based on the contour line. Operation processing is performed. Japanese Patent Application Laid-Open No. 6-1132 discloses a method for detecting the shape and direction of an electronic component by image recognition.
The method disclosed in Japanese Patent Publication No. 6 is known. This method detects the direction of the electronic component on the screen by obtaining the distribution of the inclination of the vector set on the contour obtained from the binarized image.

[0003]

However, with the recent miniaturization of electronic components, the size of electronic components to be image-recognized has also been reduced. For image recognition, even small-sized electronic components can be accurately recognized. Is required. However, the above-described binarization processing and the setting of the vector on the contour line obtained by the binarization processing are performed using a pixel constituting an image as a minimum unit. Therefore, the resolution of the tilt angle when obtaining the tilt distribution of the vector depends on the relationship between the pixel size and the relative size of the recognition target. That is, when the electronic component as the recognition target is small, the pixel size is relatively large with respect to the recognition target, and the resolution for detecting the tilt angle is reduced. As described above, in the case of a small electronic component, there is a problem that the resolution of image recognition is low, and the recognition result such as the detection of an inclination angle is not stable.

Accordingly, an object of the present invention is to provide an image recognition method capable of improving recognition accuracy and performing stable recognition.

[0005]

According to the image recognition method of the present invention, an image of a polygonal object to be recognized is taken into a camera, and a contour of the image is extracted. Setting a vector having a fulcrum and an end point, and shifting the two points by a predetermined distance along the contour line to obtain inclination data of the vector with respect to the contour line; The data is divided into data groups for each side of the polygon constituting the line, and for a data group of a plurality of sides having a predetermined relationship whose mutual included angles are known in advance, the inclination data is corrected by the included angle and the same data is corrected. The method further includes a step of obtaining a gradient distribution of the vectors, and a step of detecting a direction of the recognition target based on the gradient distribution.

According to the present invention, the inclination data of all the contour lines is divided into data groups for each side, and the data groups of a plurality of sides having a predetermined relationship are corrected to obtain the same data group. By obtaining the inclination distribution of the vector by using, the number of data can be increased even for a small recognition target having a small number of data and a large variation in the data, and thus a good recognition result can be obtained.

[0007]

Embodiments of the present invention will now be described with reference to the drawings. 1 is a perspective view of an electronic component mounting apparatus according to an embodiment of the present invention, FIG. 2 is a flowchart of the image recognition method, FIG. 3 is an image diagram of the electronic component, and FIG. 4 is an enlarged image of the electronic component. FIG. 5A is a graph showing an angle histogram of the contour, FIG. 5B is an image diagram of the electronic component, and FIG.
7A and 7B are graphs showing an angle histogram of the contour, and FIG. 7 is an image diagram of the electronic component.

Referring first to FIG. 1, an electronic component mounting apparatus in which an image recognition apparatus of the present invention is incorporated will be described. In FIG. 1, the transfer head 3 is moved in the horizontal and vertical directions by a transfer head movement control means 14. A suction nozzle N is provided at the lower end of the transfer head 3.
Is mounted, and a polygonal (rectangular) electronic component 4 as a recognition target is sucked and held by vacuum suction from a suction nozzle N by a suction unit (not shown). The suction nozzle N is rotated in the θ direction about the central axis of the suction nozzle N by a motor M built in the transfer head 3. The suction nozzle N is provided with a disc-shaped reflecting plate 3a coaxially, and reflects illumination light from the light source 2 below. The camera 1 is disposed below the transfer head 3, and captures an image of the electronic component 4 located below the reflection plate 3 a using transmitted illumination light.

A parts feeder 12 is provided on the side of the camera 1. The parts feeder 12 supplies the electronic components 4. On the opposite side of the parts feeder 12, a substrate holding table 11 is provided. On the substrate holding table 11, a substrate 10 having a surface on which a circuit pattern L is formed is held. The electronic component 4 is picked up from the parts feeder 12 by the transfer head 3 and mounted on the substrate 10.

An A / D converter 5 converts an image signal of the camera 1 into a digital signal. The frame memory 6 stores the converted digital image signal. The ROM 8 stores programs for performing various processes. CPU7 is ROM
Various calculations and processes such as image processing are performed in accordance with the program stored in the memory 8. The RAM 9 is provided with a tilt distribution storage area 9a for storing calculation results and storing a later-described tilt angle distribution of a vector. Slope distribution storage area 9
In a, frequency memories SUM (-90) to SUM (90) are set for each angle in the range of -90 degrees to +90 degrees, and the inclination memory of each vector is accumulated in the frequency memory for each angle. Is stored.

The drive circuit 13 drives a motor M for rotating the suction nozzle N. The machine controller 15 controls the driving circuit 13 and the transfer head movement control means 14,
Controlled by PU7. The electronic component 4 is imaged by the camera 1 and its position is recognized by the CPU 7, and the drive circuit 13 and the transfer head movement control means 14 are controlled based on the position recognition result.
Can be mounted on the substrate 10 after correcting the positional deviation.

Next, image recognition of the electronic component 4 will be described with reference to the flow chart of FIG. First, in FIG. 1, an image of the electronic component 4 of the inspection object is captured in the field of view of the camera 1, and A
The image data is digitized by the / D converter 5 and stored in the frame memory 6 (ST1). Thereby, as shown in FIG. 3, an image of the electronic component 4 is captured. The electronic component 4 has a polygonal shape, and appears on the image as a rectangular dark image having vertices at points A, B, C, and D, and a long side AB,
The direction of the CD is inclined by an angle α with respect to the X axis.

Next, the outline of the electronic component 4 is extracted based on the image data (ST2). This contour extraction is a well-known technique in the image processing technical field. Next, a vector having a start point and an end point at two points on the contour line is set from the data of these contour lines, and inclination data of each of these vectors is obtained (ST3). The inclination of this vector will be described with reference to FIG.

In FIG. 4, squares surrounded by a lattice in the image indicate pixels, and the hatched pixels correspond to the contour lines obtained in ST2. First, paying attention to a certain pixel Pi on the contour, the pixel Pi is set as a starting point, and a pixel located at a position shifted by a constant n pixels (for example, n = 5) from the starting point rightward or upward along the contour. A vector Si ending at Pi + n is set. Then, the inclination θi of the vector Si is obtained as an angle with the counterclockwise direction being a positive direction with respect to the X axis.

Next, along the contour, the starting point of the vector Si
The vector S is obtained by shifting the position of the end point by a predetermined distance, that is, by a predetermined number of pixels (here, one pixel).
Let it be i + 1. Then, an angle θi + 1 between the vector Si + 1 and the X axis is obtained. Similarly, by sequentially shifting the start point and the end point of the vector S along the contour line in the same manner, the vector S having substantially the same absolute value is set on the contour line, and the inclination θ of each vector S is obtained. As can be seen from FIG. 4, even if the vectors are on the contour line indicating the same side,
The inclination θ obtained with the pixel as the minimum element is not always the same angle.

Then, by adding the value of the frequency memory corresponding to the vector inclination θ obtained in this manner, an angle histogram showing the distribution of the inclination of all the vectors determined on the contour line is obtained. Can be. FIG.
(A) shows the angle histogram obtained in this way. In the histogram of FIG. 5A, two peaks 21 and 22 appear. Peak 21 is shown in FIG.
A large number of vectors set on the line segment CD shown in (b) appear at a position centered on the angle α, and similarly, the peak 22 indicates a group of vectors set on the line segment BC. , And an angle − (90−α).

As can be seen from FIG. 5, the longer the line segment on the image is, the higher the peak appearing on the histogram is, and each peak is 90 degrees corresponding to the rectangular shape of the electronic component 4. Appears in a shifted position. That is, the inclination of the electronic component 4, ie, the angle in the longitudinal direction, can be obtained by analyzing the numerical data in the frequency memory by the CPU 7 to obtain the angular position corresponding to the peak position on the histogram.

Here, the shape of the angle histogram when the size of the electronic component 4 to be inspected is small will be described with reference to FIG. As shown in FIG.
Is smaller, each peak on the histogram is
The lower and gentler shapes appear as compared with the example shown in FIG. This is because the number of vectors set on each side is small because the size of the electronic component 4 is small, and thus the number of data itself is also small, and the relative ratio between the size of the electronic component and the pixel size is small and the angle is small. This is because the variation in the inclination angle of the vector set on the contour becomes larger because the resolution is reduced. As a result, when the inclination angle α is obtained based on the angle histogram of FIG. 6A, an error in specifying the angle corresponding to the peak position is large, and therefore, the obtained inclination angle α
Contains a large error.

Therefore, the following processing is performed after ST4 in FIG. 2 in order to reduce this error and improve the detection accuracy of the inclination angle α. That is, on the angle histogram, the inclination data is divided into data groups for each side of the polygon constituting the entire contour line, and a data group of a plurality of sides having a predetermined relationship whose mutual included angle is known in advance is defined as a data group. The inclination data is corrected by the angle to make the same data group (ST4). Then, an angle histogram showing a new inclination distribution is created based on the inclination data corrected and regarded as the same data group (ST).
5).

Hereinafter, this process will be described with reference to a specific example. In FIG. 6A, peaks 31 and 32 are data groups of two sides (corresponding to the line segments CD and BC shown in FIG. 5B) constituting the contour of the electronic component 4, respectively. Since 4 is rectangular, the included angle between the two sides is 9
It is known in advance that it is 0 degrees. Therefore, in order to superimpose two peaks into the same data group, 1
The inclination data of the data group showing two peaks may be corrected by 90 degrees and then superimposed.

That is, a range corresponding to a 90-degree width including one peak is extracted from the angle histogram of FIG. 6A, and is overlapped with a portion corresponding to the remaining 90-degree width. In other words, a process of compressing the inclination data obtained in the entire range of 180 degrees into data in the range of 90 degrees is performed. Here, as the range corresponding to the width of 90 degrees,-
An example using a range of 45 degrees to 45 degrees will be described. Note that the 90-degree width is not limited to the example of the range of -45 to 45 degrees, and any method may be used as long as the two peaks on the angle histogram are clearly divided into the respective ranges. Take out the range from -90 degrees to 0 degrees and go from 0 degrees to 9
You may make it overlap in the range of 0 degree.

In order to perform this processing, when the inclination angle θ is calculated in ST3 described above, if θ ≧ 45 degrees, θ−
90 °, or θ + 90 degrees if θ <−45 degrees, is replaced with data as inclination data to be stored in the frequency memory 9a. By performing such processing, FIG.
The inclination data in the range (a) (the range corresponding to θ <−45 degrees) shown in (a) is moved in the +90 degree direction and superimposed on the range (c), and the range (d) (θ ≧ 45) The inclination data of the range (equivalent to degrees) is moved in the −90 degree direction and is superimposed on the inclination data shown in the range (b).

FIG. 6B shows a new angle histogram produced in this way, and the peaks appear more clearly than the angle histogram of FIG. 6A. Then, based on the new angle histogram, the angle α indicating the direction of the side of the contour is specified (ST6). At this time, since the peak position appears more clearly as described above, an error in specifying the angle corresponding to the peak position can be reduced. Here, since the inclination data of each side is summed up, it is not possible to distinguish between the long side and the short side, and the long side direction is specified in the following steps.

That is, based on the inclination angle α thus obtained, as shown in FIG.
An edge search is performed on the image in the direction of 2 = α ± 90 degrees (ST7). Thereby, the position of the electronic component 4 is detected by specifying the longitudinal direction and the shape of the electronic component 4 (ST).
8) Calculate the displacement amount. Then, the electronic component 4 is corrected while correcting the position shift amount by the machine controller 15.
Is mounted on the substrate 10.

As described above, according to the present invention, in image recognition of an electronic component, the number of data is increased by adding together the inclination data of the contour lines of a plurality of sides whose relative angles are known in advance, thereby increasing the number of data. This is intended to compensate for a decrease in the accuracy of detecting the inclination angle of the electronic component caused by the miniaturization.

[0026]

According to the present invention, the inclination data for all the contour lines is divided into data groups for each side, and the inclination data is corrected for the data group for a plurality of sides having a predetermined relationship to be the same data group. Then, since the inclination distribution of the vector is obtained, the number of data can be increased even for a small recognition target having a small number of data and a large variation in the data, and thus a good recognition result can be obtained. .

[Brief description of the drawings]

FIG. 1 is a perspective view of an electronic component mounting apparatus according to an embodiment of the present invention.

FIG. 2 is a flowchart of an image recognition method according to an embodiment of the present invention.

FIG. 3 is an image diagram of an electronic component according to an embodiment of the present invention.

FIG. 4 is an enlarged image view of the electronic component according to the embodiment of the present invention;

5A is a graph showing an angle histogram of a contour line according to an embodiment of the present invention. FIG. 5B is an image diagram of an electronic component according to an embodiment of the present invention.

6A is a graph showing an angle histogram of a contour according to an embodiment of the present invention; FIG. 6B is a graph showing an angle histogram of a contour according to an embodiment of the present invention;

FIG. 7 is an image diagram of an electronic component according to an embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Camera 3 Transfer head 4 Electronic component 7 CPU 9 RAM 9a Tilt distribution storage area

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 2F065 AA01 AA12 AA20 AA35 AA37 AA51 CC25 DD03 FF01 FF04 JJ03 JJ19 JJ26 NN20 PP11 QQ03 QQ23 QQ24 QQ27 QQ29 QQ31 QQ43 5B057 AA01 AA04 DA06 DB02 FA09 DC06

Claims (1)

[Claims] A step of taking an image of a polygonal object to be recognized into a camera and extracting an outline of the image; setting a vector having a fulcrum and an end point at two points on the outline; Shifting the points by a predetermined distance along the contour line to obtain inclination data of the vector with respect to the contour line; and converting the inclination data into data for each side of a polygon constituting the contour line. Divided into groups, for a data group of a plurality of sides having a predetermined relationship whose mutual included angles are known in advance, after correcting the inclination data by the included angle to be the same data group, the inclination distribution of the vector is calculated. An image recognition method, comprising a step of obtaining and a step of detecting a direction of a recognition target based on the inclination distribution.