JP3529640B2 - Image data recognition method and apparatus - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to the manufacture of electronic component mounting boards used in electronic equipment, and more particularly, to an image data recognizing method of an image pickup means for recognizing the shape and installation position of an object to be mounted and its method. It relates to the device.

[0002]

2. Description of the Related Art A conventional image data recognition method will be described with reference to FIG. 8 which shows the principle of a method of thinning out image data in a grid pattern and detecting it. In FIG. 8, 1 is a window frame of a recognition range of image data, 2 is a start point which is a start position of each point 3 detected for recognizing an object 4 from the image data,
4'is another object. The feature of this method is not to detect all the pixels in the image data in the window frame 1, but to detect an arbitrary number of pixels from the starting point 2 for each point 3 of the thinning width n in the X and Y directions. By doing so, as compared with the case of detecting the image data of all the pixels, it is possible to detect the pixels by thinning them out, so that the detection can be performed at high speed, and the image processing can be performed at high speed.

[0003]

However, according to such a conventional image data recognition method, in the recognized pattern, as the size of the object becomes smaller, as in the object 4'shown in FIG. There has been a problem that the thinning width n of each point 3 for detecting data has to be made fine, and as a result, the number of detections increases and the detection time also increases.

The present invention is directed to solving the above-mentioned problems of the prior art, and even if a pattern in which the shape of an object is recognized small is fine, it is detected in correspondence with the target pattern. It is an object of the present invention to provide an image data recognition method and a device therefor, which are performed at high speed by minimizing the number of detections by shifting the position and have high reliability.

[0005]

In order to achieve this object, an image data recognizing method and an image data recognizing apparatus according to the present invention include an image pickup means for taking in a video signal of an object for recognizing a shape and its position. And a recognition means for converting the video signal and thinning out the accumulated image data for detection processing, and setting the thinning width and the starting position of the image data according to the shape and the installation direction of the object.

Further, the image data recognizing device sets a window frame for recognizing the image data, and sets a thinning width n of the image data in accordance with the shape and installation direction of the object,
Data detection means for sequentially detecting image data at intervals of a thinning width n in one direction from the start position within the window frame, and an interval of 1 / n thinning width n in the one direction of the start position are added , and are orthogonal to the one direction. by adding the distance between the decimation width n in the other direction, means for determining a new starting position, until the distance between the starting position and a new starting position in one direction the decimation width n, means for determining a new starting position Just repeat processing by the data detector from a new starting position and, the distance between the starting position and a new starting position
When the thinning width is n, one direction is the start position and the other direction is
It is characterized by comprising means for setting a new start position obtained by adding the intervals of the thinning width n, and means for repeating the processing by the means to the outside of the window frame in the other direction.

Further, the means for setting the thinning width n of the image data is provided with one direction depending on the shape and installation direction of the object.
It is characterized in that it is provided with means for making a determination as to which of the other directions the detection processing is to be performed in detail.

In the means for setting the thinning width n of the image data, a window frame for recognizing the image data is divided into a plurality of areas, and each divided area is unidirectional depending on the shape and installation direction of the object. , And a means for making a determination as to which of the other directions the detection processing is to be performed in detail.

According to the image data recognition method of the apparatus configured as described above, a video signal of an object is taken in order to recognize the shape and its position, the video signal is converted, and the accumulated image data is thinned out and detected. In the processing, in order to set the thinning width and the start position of the image data according to the shape and installation direction of the object, the window frame that is recognized for the image data is set Setting the thinning width n of the image data to sequentially detect the image data in the window frame in one direction at intervals of the thinning width n; Add interval
And, by adding the distance between the decimation width n in the other direction orthogonal to the one direction, the process of obtaining a new starting position, to the data detected at intervals of the decimating from a new starting position in one direction width n new Repeatedly until the interval between the different start positions becomes the thinning width n, and new start positions and start positions are obtained.
When the interval is a thinning width n, one direction is the start position,
The other direction is a new start position that is the sum of the intervals of the thinning width n.
Ri conducted further to the window outside the frame is repeated in the other direction, it can be recognized without increasing the number of detections by shifting a detection position at which the shape is thinned even become finer object.

Further, in the step of setting the thinning width n of the image data, the object is determined by finely detecting one direction or the other direction according to the shape and installation direction of the object. The detection process can be selected according to the object.

In the step of setting the thinning width n of the image data, the window frame for recognizing the image data is divided into a plurality of areas, and each divided area is unidirectional depending on the shape and installation direction of the object. By determining which of the other directions the detection processing is to be performed finely, the matching area is divided into the objects in the window frame, and the detection processing matched to the objects in the area can be selected.

[0012]

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a schematic diagram showing a configuration of an image data recognition device according to an embodiment of the present invention. In FIG. 1, 4 is an object to be recognized, 5 is an illuminating means for illuminating the object 4, 6 is a camera that shoots the object 4 and outputs a video signal, and 7 is a digital signal of the video signal from the camera 6. A / D converter for converting to image data, 8
Is an image memory unit for storing image data, 9 is an image data detection unit for detecting image data for recognizing the object 4 from the accumulated image data, 10 is an A / D conversion unit 7, an image memory unit 8 and The control unit controls the image data detection unit 9.

The image data recognition apparatus shown in FIG. 1 includes an image pickup means composed of an illumination means 5 for photographing an object 4 and obtaining a video signal, a camera 6, an A / D conversion section 7, an image memory section 8,
The recognition means composed of the image data detection unit 9 and the control unit 10 detects the target pattern of the target object 4 from the image data obtained by converting the acquired video signal.

(First Embodiment) FIG. 2 is a flow chart showing the operation of the image data recognition method according to the first embodiment of the present invention. Further, FIG. 3 is a diagram for explaining the principle of the image data recognition method according to the first embodiment, and 1 is an object 4
In the range in which the pattern is recognized, 2 is a start point at which the target pattern in the window frame 1 is detected, 3 is each point thinned out to detect the target pattern, 4 is the object, 4 ′ Is another object.

Next, the operation of the first embodiment will be described based on the flowchart of FIG. 2 with reference to FIG. First, in order to detect the pattern of the object in the window frame 1, the starting point 2 and the thinning width n of each point 3 for detecting the pattern in the X direction as one direction and the Y direction as the other direction are set. At this time, the thinning width n is determined from the average width of the patterns of the objects 4, 4'in the window frame 1 in the X and Y directions (S1). Further, a new starting point 2 of each point 3 for comparing the minimum value of the pattern width in the X direction of the object and the minimum value of the pattern width in the Y direction and detecting the direction having the smallest minimum pattern width with the thinning width n As ′, the direction and width (1 / nth of the thinning width obtained earlier) for shifting it are set (S2).

In the detection of image data, for example, when the minimum value of the pattern width in the X direction is smaller than the Y direction, each point 3 detected from the start point 2 is windowed to the column in the X direction at intervals of the thinning width n. The detection process is performed until the frame 1 is exceeded (S
3). Then, in the detection of the image data, a new open
It is confirmed whether or not the m-th row in which the interval between the start positions is the thinning width n is detected (S4). This process S
4, if the first column is detected, the next second column shifts the starting position with the starting point moved by 1 / n in the X direction and n in the Y direction as the second starting point 2 '( S5), processing S3
I do.

In the process S4, when the m-th column is selected, Y
It is confirmed whether or not the point 3 to be detected in the direction (newly obtained start point) is out of the window frame 1 (S6), and when it is not out (No), the start point of the next m column is set in the X direction. In the same position as in the above-mentioned first column, the position is moved to the position of the starting point 2 ″ moved by n in the Y direction (S7), and the process S3 is performed. 1
When it is further deviated (Yes), the detection of the image data is ended.

As described above, by detecting the position of the point 3 to be detected by shifting the starting point 2 in the direction having the minimum pattern width of the object 4 ', the number of points 3 to be detected can be increased without increasing the number of points. The object 4'can be recognized as shown in FIG.

(Example) An example of the first embodiment will be described. In the description of the first embodiment, the direction and width for shifting the starting point 2 set by the process S2 shown in FIG. 2 are executed in the X direction (the minimum value of the pattern width in the X direction is smaller than the Y direction). However, here, the case of the Y direction will be described with reference to the diagram for explaining the principle of the image data recognition method shown in FIG. Also,
This is replaced with what is described in parentheses in and after step S3 of the flowchart shown in FIG.

This embodiment is the same up to the process S2 in the flowchart of FIG. 2, but in the process S3 for detecting image data, each point 3 detected from the start point 2 is Y with an interval of a thinning width n. The detection processing is performed until the line in the direction exceeds the window frame 1 (S3). A new development in image data detection
It is confirmed whether or not the m-th row in which the interval between the start positions is the thinning width n is detected (S4). In this processing S4, if the first line is detected, the second line moves to the second point by moving the start point 1 / n in the Y direction and n in the X direction.
The start position is shifted as the start point 2'of (S5), and the process S
Do 3.

In the process S4, when the m-th line, X
It is confirmed whether or not the point 3 detected in the direction is out of the window frame 1 (S6), and when it is not out (No), the start point of the next m-th line is the same as the above-mentioned first line in the Y direction. Position, shift to the position of the starting point 2 ″ moved n in the Y direction (S
7), processing S3 is performed. Further, in the process S6, when the point 3 detected in the X direction is out of the window frame 1 (Ye
s), detection of image data is ended.

As a more specific example, the operation when the thinning width n = 2 will be described with reference to the diagram for explaining the image data recognition method shown in FIG. Detection is performed in a row in the X direction from the first start point 2 in the window frame 1 at intervals of a thinning width n = 2 until the window frame 1 is exceeded. In the second row, 1 in the X direction
/ N (1/2), the second point 3 moved n (2) in the Y direction
The detection is performed in the same manner as described above as the starting point 2'of. New development
Since the thinning width n = 2, the starting point at which the interval between the start position and the starting position is the thinning width n is the starting point 2 ″ at the same position as the first starting point 2. As shown in FIG. , it is possible to detect position base ratio when performing full detection, to detect at high speed.

(Second Embodiment) FIG. 6 is a flow chart showing the operation of the image data recognition method according to the second embodiment of the present invention. FIG. 7 is a diagram for explaining the principle of the image data recognition method according to the second embodiment. here,
The same constituent elements as those in FIG. 3 showing the first embodiment are designated by the same reference numerals. In FIG. 7, 1 is a window frame, 2
Is a start point, 3 is a point, and 11, 12, 13, and 14 are divided areas obtained by dividing the window frame 1.

Next, the operation of the second embodiment will be described based on the flowchart of FIG. 6 with reference to FIG. First, in order to detect the pattern of the target object in the window frame 1, the pattern is divided into several regions based on the average width and minimum value of the pattern of the target object and the arrangement thereof. In the second embodiment, four areas 11,
It is divided into 12, 13, and 14 (S11). Divided 1
In one area, for example, the thinning width n in the area 11 is set to the starting point 2 and each point 3 for pattern detection in the X and Y directions (S12). Further, the minimum values of the pattern widths in the X and Y directions of the object in the area 11 are compared, and the smaller one is
As the position of the new start point of each point 3 detected by the thinning width n, the direction and width for shifting it are set (S13).

Here, in the detection of image data, when the minimum value of the pattern width in the X direction is smaller than that in the Y direction, each point 3 detected from the start point 2 is separated from the row in the X direction at intervals of the thinning width n. The detection process is performed until the area 11 is exceeded (S1).
4). Then, in the detection of the image data, a new open
It is confirmed whether or not the m-th row in which the interval between the start positions is the thinning width n is detected (S15). In this processing S15, if the first column is detected, the next second column has a start position which is 1 / n in the X direction and n moved in the Y direction as the second start point 2'and the start position. Shifting (S16) and processing S14 are performed.

In the process S15, in the m-th column,
It is confirmed whether or not the point to be detected in the Y direction is out of the area 11 (S17), and when it is not out (No), the next m
The starting point of the column is shifted to the position of the starting point 2 ″ moved n times in the Y direction at the same position as the first column (initial position) in the X direction (S18), and the process S14 is performed.

In step S17, when the point 3 to be detected in the Y direction (newly obtained starting point) is out of the area 11 (Yes), is there an unprocessed area to be detected in the window frame 1? It is confirmed whether or not (S19). When there is an unprocessed area in the processing S19 (Yes), the processing S12 is executed to detect the next area (S20). When there is no unprocessed area in the process S19 (No), the detection of the image data ends.

Also in the second embodiment, the first embodiment
In the same way as in the embodiment of X,
It is possible to perform the detection in which the direction and the Y direction are interchanged, and the description thereof will be duplicated and will not be repeated here.

From the above, for each divided area in the window frame 1 to be detected, the direction in which the starting point 2 is finely shifted can be set in accordance with the pattern of the object in that area, and the number of points to be detected can be increased. The precise processing can be performed without any need, and the thinning width n that matches the pattern of the object can be set individually, and the accurate image data can be detected at a higher speed.

[0030]

As described above, according to the present invention,
When detecting thinned image data, even if the shape of the pattern of the target object becomes fine, the number of detections is minimized.Therefore, by finely shifting the position of the starting point for detection according to the target object, High-speed and highly reliable detection processing can be performed according to the fineness of the pattern. Further, by dividing the window frame to be detected into a plurality of regions, it is possible to select a direction in which the thinning width or the starting point is finely shifted in consideration of the arrangement of the target object for each region,
The present invention has an effect of enabling more efficient and highly reliable detection processing of image data.

[Brief description of drawings]

FIG. 1 is a schematic diagram showing a configuration of an image data recognition device according to an embodiment of the present invention.

FIG. 2 is a flowchart showing the operation of the image data recognition method according to the first embodiment of the present invention.

FIG. 3 is a diagram illustrating the principle of an image data recognition method according to the first embodiment of the present invention.

FIG. 4 is a diagram illustrating the principle of the image data recognition method in the example of the first embodiment of the present invention.

FIG. 5 is a diagram illustrating a specific image data recognition method according to the first embodiment of the present invention.

FIG. 6 is a flowchart showing the operation of the image data recognition method according to the second embodiment of the present invention.

FIG. 7 is a diagram illustrating the principle of the image data recognition method according to the second embodiment of the present invention.

FIG. 8 is a diagram illustrating the principle of a conventional image data recognition method.

[Explanation of symbols]

1 window frame 2,2 ', 2 "starting point 3 points 4,4 'object 5 Lighting means 6 cameras 7 A / D converter 8 Image memory section 9 Image data detector 10 Control unit

Front page continuation (58) Fields surveyed (Int.Cl. ⁷ , DB name) G06T 1/00 300 G06T 3/40 H05K 13/08

Claims (6)

(57) [Claims] 1. In a process of capturing a video signal of an object for recognizing a shape or its position, converting the video signal, and thinning out and storing accumulated image data, the processing is performed depending on the shape or installation direction of the object. An image data recognition method for setting a thinning width and a start position of the image data by setting a window frame for recognizing the image data,
Decimation of the image data according to the shape and installation direction of the elephant
One of the step of setting the width n and the start position in the window frame
Image data is sequentially detected at intervals of the thinning width n.
Data detection process and thinning out in one direction of the starting position
The interval of 1 / nth of the width n is added to make it orthogonal to the one direction.
The new start position is added by adding the interval of the thinning width n in the direction.
Position, and a new start position in the one direction and
Until the interval with the start position becomes the thinning width n, the new
From the step of obtaining the starting position and the new starting position,
Repeat the data detection process to
When the distance from the start position is the thinning width n
At the start position, the new direction is obtained by adding the interval of the thinning width n in the other direction.
The step of setting the shelf start position and the step in the other direction
A method of recognizing image data, which comprises the step of repeating the process to the outside . 2. A process for setting a thinning-out width n of the previous SL image data, whereas according to the shape and installation direction of the object
Judges whether to perform fine detection processing in one direction or another direction
Image data recognizing method according to claim 1, wherein the performing. 3. A window frame for recognizing image data in a plurality of areas in the step of setting a thinning width n of the image data.
2. The image data recognition according to claim 1 , wherein the image data recognition is performed by dividing each of the divided areas into one direction and the other direction and finely performing detection processing according to the shape and installation direction of the object. Method. 4. An object for recognizing a shape and its position
Image capturing means for capturing the video signal of
In other words, a recognizer that thins out the accumulated image data and performs detection processing.
And a step, depending on the shape and installation direction of the object
Image data to set thinning width and start position of image data
A recognition device, which sets a recognition window frame for the image data,
A means for setting the thinning width n of the image data according to the shape and installation direction of the elephant, and one of the means from the start position in the window frame.
Image data is sequentially detected at intervals of the thinning width n.
Data detection means and the thinning out in one direction of the start position
The other of which the interval of width 1 / n is added and which is orthogonal to the one direction
To the new start position.
And a new starting position in the one direction and the opening
The new opening is performed until the distance from the start position becomes the thinning width n.
A means for determining the starting position and data detection from the new starting position
The processing by the output means is repeated until the new start position and the previous position are reached.
When the distance from the start position is the thinning width n, one direction is
At the start position, the interval of the thinning width n is added in the other direction.
The means for setting a new start position and the processing by the means
Images data recognition device you characterized in that a means for repeating the direction to the window outside the frame. 5. A thinning width n of the image data is set.
Depending on the shape and installation direction of the object, one direction, the other
A method to determine which direction to perform fine detection processing
The image data recognition device according to claim 4 , further comprising a step . 6. A means for setting a thinning-out width n of the previous SL image data, the image data recognizing window frame into a plurality of areas divided
Then, according to the shape and installation direction of the object for each divided area
Whether to detect finely in one direction or in the other direction
The image data recognition device according to claim 4, further comprising means for making a determination .