JP2000105827A - Position shift detecting device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a position shift detecting device which can detect a position shift quantity fast with high precision. SOLUTION: This is a position shift detecting device which detects the quantity of the position shift of an object image to be inspected from a reference image for image inspection obtained by picking up a web by combining horizontal scanning by a line sensor camera and vertical scanning by feeding the web, and has a parallel matching means which calculates the degree of matching by matching in parallel and a position shift quantity decision means 10 which decides the quantity of the position shift giving the largest degree of matching among all calculated degrees of matching.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an inspection apparatus for comparing a reference image and an image to be inspected in order to correct a positional deviation between the reference image and the image to be inspected and then perform the comparison. It belongs to the technical field of detecting misalignment.
In particular, the present invention relates to a position shift detection device capable of detecting a position shift amount with high accuracy and high speed.

[0002]

2. Description of the Related Art For example, Japanese Patent Publication No.
The image data read from the pattern of the running printed matter at the time when the normal printed matter is printed is stored in the image memory as reference image data, and the reference image data read from the image memory is read from the printed matter to be inspected during printing. A technique relating to an inspection apparatus of a type in which the quality of a printed matter is determined by comparing the read inspection image data with a pixel unit is described.

As described above, when the reference image data and the inspection target image data are compared on a pixel-by-pixel basis, if the accuracy of defect detection is increased, even if the positional deviation between the two data is slight, the positional deviation is erroneously detected as defective. In order to prevent erroneous detection of a position shift as a defect, it is necessary to lower the accuracy of the defect detection. Therefore, if a certain amount of displacement cannot be avoided, inspection performance sufficient to satisfy the requirements cannot be obtained. Therefore, for example, Japanese Patent Application Laid-Open No. 7-249122 describes an inspection apparatus configured to correct a positional shift between reference image data and inspection target image data, and then compare and inspect the two data. I have.

[0004]

The inspection apparatus can detect and correct a position shift. However, a lot of calculation time is required to detect the displacement. In particular,
A huge amount of calculation time is required to detect a position shift with high accuracy. For example, when a template having 100 pixels in each of the vertical and horizontal directions is moved by 5 pixels in each of the vertical and horizontal directions to search for a moving position where the maximum matching degree is obtained, 100
× 100 × 11 × 11 times of subtraction and accumulation are required.
When the amount of calculation is large as described above, it is difficult to detect the displacement each time the inspection target image is captured, and it is possible to apply the detection data of one displacement to a plurality of imagings. Is being done. For this reason, there is a problem that there is a difference between the position shift detection data and the actual position shift amount, and it is not possible to cope with a rapidly changing position shift.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a position shift detecting device capable of detecting a position shift amount with high accuracy and high speed.

[0006]

The above objects are achieved by the present invention described below. In other words, the misalignment detection device according to the first aspect of the present invention includes a reference image for image inspection obtained by imaging the web by combining main scanning by a line sensor camera and sub-scanning by web transport. What is claimed is: 1. A misalignment detection device for detecting an amount of misalignment between a target image and a parallel matching operation unit that performs a plurality of matching operations in parallel to calculate a degree of coincidence. And a position shift amount determining means for determining a position shift amount that gives the maximum degree of coincidence.

According to the present invention, a plurality of matching operations are performed in parallel by the parallel matching operation means to calculate the coincidence, and the maximum coincidence among all the coincidences calculated by the displacement amount judging means. Is determined. That is, since a plurality of matching operations are performed in parallel, a high-precision operation can be performed within the same time, and the operation can be completed at high speed within the same time. Therefore, there is provided a position shift detecting device capable of detecting the position shift amount with high accuracy and high speed.

According to a second aspect of the present invention, there is provided a displacement detecting apparatus according to the first aspect, wherein the parallel matching calculating means is configured to divide a predetermined search range into the plurality of search ranges. The plurality of matching operations in the partial search range are performed in parallel to calculate the degree of coincidence. According to the present invention, the predetermined search range is divided into a plurality of partial search ranges, and a matching operation is performed on each of them in parallel. Therefore, the amount of displacement can be detected at high speed.

[0009] In a displacement detecting apparatus according to a third aspect of the present invention, in the displacement detecting apparatus according to the second aspect, the partial search range includes pixels in even rows and even columns, odd rows and pixels in the search range. The four partial search ranges are obtained by extracting pixels in even columns, pixels in even rows and odd columns, and pixels in odd rows and odd columns. According to the present invention, four partial search ranges are generated from one search range, so that it is possible to detect a position shift at approximately four times the speed.

According to a fourth aspect of the present invention, there is provided a position shift detecting apparatus according to any one of the first to third aspects, wherein the reference image is provided in a pixel row direction and / or a pixel column direction. Image interpolation means for generating a quasi-reference image moved within a range of less than one pixel by applying an interpolation method, wherein the parallel matching calculation means uses the criterion image and the quasi-reference image to enter a predetermined search range. On the other hand, the matching operation is performed in parallel to calculate the degree of coincidence.
According to the present invention, a quasi-reference image in which the reference image is moved by less than one pixel in the pixel row direction and / or the pixel column direction by the image interpolation unit is generated by applying the interpolation method,
A matching operation is performed in parallel for each of the reference image and the moved quasi-reference image. Therefore, the amount of displacement can be detected with high accuracy in a range of less than one pixel.

According to a fourth aspect of the present invention, there is provided a displacement detecting apparatus according to the third aspect, wherein the image interpolating means sets the reference image as a reference image (0, 0). The quasi-reference image (0.5, 0) shifted half a pixel from the image (0, 0) in the pixel row direction, and the quasi-reference image (0) shifted a half pixel from the reference image (0, 0) in the pixel column direction. , 0.5), and a quasi-reference image (0.5, 0.5) shifted by half a pixel from the reference image (0, 0) in the direction of pixel columns and pixel rows.
0.5), and the parallel matching calculation means generates the reference image (0, 0) and the quasi-reference image (0.5,
0), the quasi-reference image (0, 0.5), and the quasi-reference image (0.5, 0.5). According to the present invention, the quasi-reference image shifted by 0.5 pixel in the pixel row direction and / or the pixel column direction from the reference image by the image interpolation means is generated, and the total of four images is generated by the parallel matching calculation means. Are subjected to a matching operation in parallel. Therefore, a high-accuracy positional displacement of 0.5 pixel unit can be detected at high speed.

[0012]

Next, the present invention will be described with reference to embodiments. FIG. 1 shows a configuration of an image inspection apparatus to which the position shift detecting device of the present invention is applied. Although there is no particular limitation on the applicable range of the position shift detecting device of the present invention, here,
A description will be given assuming an example in which the present invention is applied to an image inspection apparatus for printed matter or the like as a particularly preferred embodiment. In FIG. 1, 1
Is an image input unit, 2 is a reference image memory, 3 is an inspection image memory, 4 is a template position setting unit, 5 is an image interpolation unit, 6, 7, 8, and 9 are matching units, 10 is a displacement amount calculating unit, 11 Is a comparative inspection means.

The image input means 1 captures an image of a printed web (printed material) to be inspected and generates a digitized image (image data). Image input means 1
For example, a transfer device for transferring a web, a transfer detector for detecting a transfer distance and a transfer speed of the web, and a specific position in one cycle of an image printed on the web, for example, a start position of a picture and a mark position are detected. A timing detector, a line sensor camera for imaging a linear region in a direction perpendicular to the web transfer direction, and an analog-to-digital converter for converting an image signal (analog signal) of the line sensor camera into digital data. ), Etc.

The image input means 1 captures an image of the web by combining the main scanning by the line sensor camera and the sub-scanning by transporting the web. From the output signals of the timing detector and the transport detector, a series of signals indicating predetermined imaging positions obtained by equally dividing one cycle of the image of the web transported by the transport device is generated. An imaging signal output from the line sensor camera by the main scanning is converted into digital data by an A / D converter, and the digital data is sequentially stored in a storage device in synchronization with a series of signals indicating the imaging position. Some line sensor cameras synchronize main scanning with a signal indicating the imaging position, and others do not. However, at least digital data is always stored in synchronization. In this manner, one cycle of an image (image data) is stored in the storage device.

The reference image memory 2 is a storage device for storing image data obtained by the image input means 1 picking up an image (of a non-defective printed matter) serving as a reference for the web. The inspection image memory 3 is a storage device that stores image data obtained by the image input unit 1 capturing an image to be inspected on the web.

The template position setting means 4 sets a specific area in the image data as the position of the template. The template is a set of pixel data of the specific area. The image data may be either reference image data or inspection target image data. If it is a template of reference image data, a matching operation (described later) for changing the position of the template in the search range of the inspection target image data and obtaining an evaluation value of the degree of coincidence is performed. If it is a template of the image data to be inspected, a matching operation is performed to change the position of the template in the search range of the reference image data and obtain an evaluation value of the degree of coincidence. In the example shown in FIG. 1, the position of the template is set based on the reference image data, and the template itself is a template cut out from the inspection target image data.

FIG. 2 shows an example of the image data and the template. 2, 21 is an image data shown by the image, T ₀₀ through T ₁₁ was set to the image data template (rectangular portion indicated by hatching). Although the example shown in FIG. 2 is four templates T ₀₀ through T ₁₁ are provided, the present invention can be applied independently to each template, at least one template may be set. The template position setting means 4 sets an area suitable for the matching operation from the image data as a template. The area suitable for the matching calculation is an area where the change in the evaluation value of the matching degree becomes large when the relative position between the template and the image data is moved in the direction of the pixel row and / or the direction of the pixel column due to the positional shift. That is.

The image interpolating means 5 generates a reference image obtained by moving the reference image in a range of less than one pixel in a direction of a pixel row and / or a direction of a pixel column by applying an interpolation method. A well-known interpolation method can be applied as the interpolation method. For example,
Lagrange interpolation, difference interpolation, difference quotient interpolation,
Hermite interpolation, spline interpolation, etc. The reference image moved in a range of less than one pixel is used for detecting a position shift in a range of less than one pixel. The moving amount (distance) of the reference image having the highest matching degree is regarded as the positional deviation amount (distance).

For example, the image interpolation means 5 converts the reference image stored in the reference image memory 2 into a reference image (0, 0).
The quasi-reference image (0.5,0) shifted by a half pixel in the pixel row direction from the reference image (0,0), and the quasi-reference image shifted by a half pixel in the pixel column direction from the reference image (0,0) The reference image (0, 0.5), the quasi-reference image (0.5, 0.5) shifted by a half pixel in the pixel row direction from the reference image (0, 0)
Generate When the moved amount is a half pixel, as a simple interpolation method, the pixel value of the reference image shifted by a half pixel from the pixel value of the nearest pixel 2 or the pixel value of the nearest pixel 4 by simple averaging Can be requested.

These quasi-reference images generated by the image interpolation means 5 do not need to be in the same image range as the reference image for the purpose of detecting only the positional deviation. The same image range as the search range in the matching calculation is sufficient. However, in the image inspection apparatus, misregistration correction is performed, and a comparison operation between the inspection target image and the corresponding pixel is performed. Therefore, the quasi-reference image generated by the image interpolation means 5 has the same image range as the reference image, and is stored in the reference memory 2 in the same manner as the reference image.

The matching means 6 to 9 perform matching operations in parallel. That is, it is a part that constitutes the parallel matching calculation means. The parallel matching calculation means calculates a degree of coincidence by performing a plurality of matching calculations in each of the partial search ranges obtained by dividing the predetermined search range into a plurality by the matching means 6 to 9 in parallel. The parallel matching calculation means calculates the degree of coincidence by performing the matching calculation in parallel with respect to a predetermined search range using the reference image and the quasi-reference image generated by the image interpolation means by the matching means 6 to 9.

The parallel matching calculation means performs a first case of performing a plurality of matching calculations in each partial search range using the reference image and a plurality of matching calculations in a predetermined search range using the reference image and the quasi-reference image. There is a second case to be performed. In the first case, the displacement can be detected at high speed, and in the second case, the displacement can be detected with high accuracy. For example, the partial search range is obtained by extracting pixels in even rows and even columns, pixels in odd rows and even columns, pixels in even rows and odd columns, and pixels in odd rows and odd columns in the search range. Range.

As the evaluation value of the degree of coincidence in the matching operation, a correlation coefficient, a sum of absolute values of differences of corresponding pixels, a sum of squares of differences of corresponding pixels, and the like can be used. An evaluation value with a high degree of coincidence is the maximum value (relative
maximum), but the relative sum of the absolute value of the difference and the square of the difference between the corresponding pixels is a relative minimum.

The displacement amount judging means 10 judges a displacement amount which gives the maximum matching degree among all the matching degrees obtained by the matching calculation. That is, the degree of coincidence at each position obtained by changing the position of the template in the search range is compared, and the amount of positional deviation is determined by the positional deviation amount determining means 10 from the position of the template that gives the maximum degree of coincidence.

The comparison / inspection means 11 compares the reference image and / or the quasi-reference image stored in the reference image memory with the image to be inspected by inputting the positional deviation amount output from the positional deviation The correction is performed while the amount is corrected for each pixel, and the quality of the inspection target image is determined. Whether the reference image or the quasi-reference image is used for the pass / fail determination is, of course, given by the amount of positional deviation of the pixel of the inspection target image which is the target of the comparative inspection, and the maximum coincidence with the reference image. It depends on whether the received image has been given the highest degree of coincidence with the quasi-reference image.

Next, the operation of the image inspection apparatus to which the position shift detecting device of the present invention is applied in the above configuration will be described. As described above, an image inspection device for a printed material or the like is assumed. When printing is performed in a web printing machine to produce a non-defective print, an image of the non-defective print is input to the image input unit 1.
Is imaged. The image input means 1 outputs reference image data, and the reference image memory 2 stores the reference image data.

The template position setting means 4 sets an area suitable for the matching operation as the position of the template based on the reference image data. This position is a position using a captured image (which may be a reference image or an inspection target image) as a coordinate system. It is not a coordinate system based on the picture of the printed matter to be imaged. Therefore, the pattern included (pasted) at the position of the template in the inspection target image is, of course, the position of the reference image if there is a positional shift of the printed matter or the positional shift of the imaging position after capturing the reference image. Has a different pattern.

The image interpolating means 5 differs between the case where the above-described matching operation in the first case is performed and the case where the above-described matching operation in the first case is performed. When performing the above-described matching operation in the first case, the image interpolation means 5 determines the search range in the matching operation for the set template region from the reference image data stored in the reference image memory 2. The reference image data of the search range that is the reference image data of the added area is generated. Then, the image interpolation means 5 sets the reference image data itself in the search range in each of the matching means 6 to 9.

When performing the above-described matching operation in the second case, the image interpolating means 5 includes the reference image memory 2
To generate quasi-reference image data by performing interpolation based on the reference image data stored in the. The reference memory 2 stores the quasi-reference image data in the same manner as the reference image data.
The image interpolation unit 5 cuts out the reference image data and the quasi-reference image data of the search range from the reference image data and the quasi-reference image data, and sets them in the matching units 6 to 9.

First, the first case will be described. The image input unit 1 captures an image of an inspection target printed on a printing machine and outputs inspection target image data. The inspection image memory 3 stores the inspection target image data.
The matching means 6 to 9 cut out data of the position of the template from the inspection target image data based on the position of the template set as described above, and set the data as a template. Using this template, the matching units 6 to 9 set the predetermined search range to the matching units 6 to 9
In each of the partial search ranges obtained by dividing into four, the position of the template is changed to match with the reference image data of the search range (the reference image data of the region obtained by adding the search range to the template region). The calculation is performed in parallel to calculate the degree of coincidence.

Assuming that the position in the predetermined search range is expressed by the top and bottom left and right directions, the partial search range is defined by the right and left directions with respect to the center position of the predetermined search range (usually coincident with the center position of the template position). There can be four search ranges: leftward, leftward, and rightward. The partial search range includes four search ranges obtained by extracting pixels in even rows and even columns, pixels in odd rows and even columns, pixels in even rows and odd columns, and pixels in odd rows and odd columns in the search range. It can be.

The position shift amount judging means 10 inputs the position of the template and the degree of coincidence (evaluation numerical value thereof) output by the matching means 6 to 9 performing the matching operation. The displacement amount determining means 10 determines the position of the template that gives the highest degree of coincidence, and calculates the displacement amount from that position. The comparison inspection means 11 corrects the positional deviation amount between the reference image stored in the reference image memory 2 and the inspection target image stored in the inspection image memory 3 by inputting the positional deviation amount output by the positional deviation amount determining means 10. Are compared in a pixel unit, and the quality of the inspection target image is determined.

Next, the second case will be described. The image input unit 1 captures an image of an inspection target printed on a printing machine and outputs inspection target image data. The inspection image memory 3 stores the inspection target image data.
The matching means 6 to 9 cut out data of the position of the template from the inspection target image data based on the position of the template set as described above, and set the data as a template. Using this template, the matching units 6 to 9 change the position of the template and perform matching operations with the reference image data or the quasi-reference image data in the search range set in each of the matching units 6 to 9 in parallel. Then, the degree of coincidence is calculated. For example, the matching means 6 is the aforementioned reference image (0, 0), the matching means 7 is the aforementioned quasi-reference image (0.5, 0), and the matching means 6 is the aforementioned quasi-reference image (0, 0.5). ) And the matching means 7 perform a matching operation with the quasi-reference image (0.5, 0.5).

The position shift amount judging means 10 inputs the position of the template and the degree of coincidence (evaluation numerical value thereof) output by the matching means 6 to 9 performing the matching operation. The displacement amount determining means 10 determines the position of the template that gives the highest degree of coincidence, and calculates the displacement amount from that position. The comparison inspection unit 11 compares the reference image or the quasi-reference image stored in the reference image memory 2 with the inspection target image stored in the inspection image memory 3 by inputting the displacement amount output by the displacement amount determination unit 10 and The correction is performed while correcting the shift amount on a pixel-by-pixel basis, and the quality of the inspection target image is determined.

[0035]

As described above, according to the present invention, high accuracy,
A position shift detecting device capable of detecting a position shift amount at high speed is provided.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of an image inspection apparatus to which a position shift detection device of the present invention is applied.

FIG. 2 is a diagram illustrating an example of image data and a template.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Image input means 2 Reference image memory 3 Inspection target image memory 4 Template position setting means 5 Image interpolation means 6, 7, 8, 9 Matching means 10 Position shift amount judging means 11 Comparative inspection means 21 Image data

Claims (5)

[Claims] 1. A method for detecting a positional shift between a reference image for image inspection and an image to be inspected obtained by imaging the web by combining main scanning by a line sensor camera and sub-scanning by transporting the web. A parallel matching operation means for performing a plurality of matching operations in parallel to calculate a degree of coincidence, and a position deviation amount providing a maximum degree of coincidence among all the calculated degrees of coincidence. A positional deviation detecting device, comprising: a positional deviation amount determining means for determining. 2. A position shift detecting apparatus according to claim 1, wherein said parallel matching calculation means performs the plurality of matching calculations in each of the partial search ranges obtained by dividing a predetermined search range into the plurality. A misalignment detecting device for calculating a degree of coincidence. 3. The position shift detecting device according to claim 2, wherein the partial search range includes an even row and even column pixel, an odd row and even column pixel, an even row and odd column pixel in the search range. An apparatus for detecting a positional shift, comprising four partial search ranges obtained by extracting pixels in odd rows and odd columns. 4. A position shift detecting apparatus according to claim 1, wherein said reference image is displayed in a direction of a pixel row and / or in a direction of a pixel row.
Alternatively, the image processing apparatus further includes image interpolation means for generating a quasi-reference image moved within a range of less than one pixel in the direction of the pixel row by applying an interpolation method, and the parallel matching calculation means uses the reference image and the quasi-reference image. A misalignment detection device, wherein a matching operation is performed in parallel on a predetermined search range to calculate a degree of coincidence. 5. The apparatus according to claim 4, wherein said image interpolating means sets said reference image as a reference image (0,0) and shifts the reference image (0,0) in a direction of a pixel row from the reference image (0,0). The quasi-reference image (0.5,0) shifted by a pixel, the quasi-reference image (0,0.5) shifted by half a pixel from the reference image (0,0) in the direction of a pixel row, and the criterion image (0,0) ), A quasi-reference image (0.5,
0.5), and the parallel matching calculation means generates the reference image (0, 0) and the quasi-reference image (0.5,
0), the quasi-reference image (0, 0.5) and the quasi-reference image (0.5, 0.5) are subjected to a matching operation in parallel with each other. apparatus.