KR100883544B1 - Image tester capable of high speed position correction of image and method thereof - Google Patents

Abstract

Disclosed are an image inspector and an image inspection method capable of fast position correction of an image. The image inspection method may include converting an input image to gray scale, dividing the converted input image into line units, calculating a histogram value for each line, and generating a first histogram based on the calculated histogram value for each line; Dividing the first histogram by a first interval unit, calculating an average of the histograms per line for each of the divided first intervals, and generating a second histogram based on a calculation result; Dividing the second histogram by a second interval unit, detecting a change amount of the second histogram value for each divided second interval, and generating a third histogram based on a detection result; And comparing the third histogram with a previously stored reference histogram and detecting whether there is a position error of the input image based on a result of the comparison. There is an effect that can be corrected.

Lens, focus

Description

Image tester capable of high speed position correction of image and method

The present invention relates to an output image inspection technique, and more particularly, to an image inspector and an image inspection method capable of fast position correction of an image.

Computer vision is a field of artificial intelligence, and the purpose of computer vision is to understand and analyze scenes, features, etc. in an image.

The computer vision is mainly used to recognize a specific object in an image, to analyze and register a recognized specific object, to track a specific object in a continuous image, or to map a specific scene to a three-dimensional model. Techniques in various fields are used, such as pattern recognition, image processing, graph theory, or cognitive psychology.

In particular, machine vision, one of the computer vision fields, is an artificial intelligence field in which a computer automatically receives, senses, and interprets an image of a work process site to improve a process.

The computer uses template matching when recognizing a specific object in an image input from the machine vision, and the model matching is a method of determining where a model (that is, a specific object) to be found in the image is located. to be.

For example, the computer is used in an industrial system for producing continuous objects at high speed, such as in the printing industry, papermaking, or film production, and compares the original image and the copied image output from the industrial system through the model matching and compares the result with the comparison result. On the basis of this, a quality inspection of the copied image may be performed.

The computer may measure how well the model matches the copied image by comparing the original image including the position or shape of the model desired by the model through the model matching, and correct the position of the model by reflecting the measured result. Alternatively, shape correction may be performed.

However, the commonly used model matching performs computation by processing the image two-dimensionally when comparing the original image and the copied image, and finds the optimal position by moving the pixels of the copied image little by little. This can be time consuming.

Accordingly, a technical problem to be achieved by the present invention is to provide an image inspector and an image inspection method capable of high-speed position correction of an image by histogramting a two-dimensional image in one dimension during image inspection.

In accordance with another aspect of the present invention, an image inspection method includes converting an input image to gray scale, dividing the converted input image by line, and calculating a histogram value for each line and a first histogram based on the calculated histogram value for each line. Generating a; Dividing the first histogram by a first interval unit, calculating an average of the histograms per line for each of the divided first intervals, and generating a second histogram based on a calculation result; Dividing the second histogram by a second interval unit, detecting a change amount of the second histogram value for each divided second interval, and generating a third histogram based on a detection result; And comparing the third histogram with a pre-stored reference histogram and detecting whether there is a position error of the input image based on a comparison result.

The line unit may be based on at least one of the vertical direction and the horizontal direction.

The histogram value for each line may be a total sum or an average of pixels constituting the line unit.

The image inspection method may further include generating a position correction signal based on the comparison result, wherein the position correction signal is transmitted to a printing apparatus and the printing apparatus configures the input image in response to the position correction signal. You can correct the position of the subject.

The second histogram value for each second interval may include a sum of histogram values constituting any one of the sections of the second histogram divided by the second interval unit and another section adjacent to the one interval. It may be generated based on the difference value of the sum of the histogram values constituting the.

In order to accomplish the above technical problem, an image inspector converts an input image to gray scale, calculates a histogram value for each line by dividing the converted input image by line units, and generates a first histogram based on the calculated histogram value for each line. A gray scale conversion unit to generate; A blurring unit for dividing the first histogram into units of a first interval, calculating an average of the histograms per line for each of the divided first intervals, and generating a second histogram based on a calculation result; An edge detector for dividing the second histogram by a second interval unit, detecting an amount of change in the second histogram value for each divided second interval, and generating a third histogram based on a detection result; And a comparison determiner configured to compare the third histogram with a previously stored reference histogram, and detect whether a position error of the input image is detected based on a comparison result.

The line unit may be based on at least one of the vertical direction and the horizontal direction.

The histogram value for each line may be a total sum or an average of pixels constituting the line unit.

The image inspector may further include a correction unit generating a position correction signal based on the comparison result, wherein the position correction signal is transmitted to a printing apparatus and the printing apparatus configures the input image in response to the position correction signal. You can correct the position of the subject.

The change amount of the second histogram for each second interval is a sum of histogram values constituting any one of the sections of the second histogram divided by the second interval unit and the other one adjacent to the one interval. It may be generated based on the difference value of the sum of the histogram values constituting the interval.

Machine vision system for achieving the technical problem may include the image inspector.

A camera used for quality inspection of a product for achieving the technical problem may include the image inspector.

As described above, the image inspector and the image inspection method capable of high-speed position correction of an image according to the present invention have the effect of performing a high-speed position correction of an image by histogramting a two-dimensional image in one dimension during image inspection.

In addition, an image inspector and an image inspection method capable of high-speed position correction of an image according to an embodiment of the present invention are used in a machine vision system (for example, a system for producing continuous objects at high speed such as printing industry, paper, or film production). Therefore, there is an effect of detecting and correcting errors at high speed during quality inspection of the finished product.

In order to fully understand the present invention, the operational advantages of the present invention, and the objects achieved by the practice of the present invention, reference should be made to the accompanying drawings, which illustrate preferred embodiments of the present invention. Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. Like reference numerals in the drawings denote like elements.

1 is a block diagram of an image inspector according to an exemplary embodiment of the present invention, FIG. 2 is a machine vision system including the image inspector of FIG. 1, and FIG. 3 shows the first to fourth histograms by the image inspector of FIG. 1. It is a figure for explaining the process which is produced | generated.

4 shows first and second histograms generated by the image inspector of FIG. 1, FIG. 5 shows third and fourth histograms generated by the image inspector of FIG. 1, and FIG. 6 shows the histograms of FIG. FIG. 5 is a diagram for describing a process of generating fifth and sixth histograms. FIG. 7 is a fifth and a sixth histogram generated by the image inspector of FIG. 1.

1 to 7, a camera 5 or machine vision system (3, eg, a system for producing continuous objects at high speed, such as in the printing industry, paper, or film production) used for quality inspection. The image inspector 10, which may be implemented in the, may include a gray scale converter 12, a blurring unit 14, an edge detector 16, a comparison determiner 18, and a corrector 19.

Each of the image inspector 10 and the gray scale converter 12, the blurring unit 14, the edge detector 16, the comparison determiner 18, and the compensator 19 configuring the image inspector 10, respectively. May be implemented in hardware (H / W) and / or software (S / W).

The gray scale converter 12 is an input image 22, for example, an image of a product photographed by a camera 5 used for quality inspection or an image printed by a printing apparatus 20 in the machine vision system 3. ) to receive and to convert the received input image (22) to the gray scale (gray scale), the transformed the input image line-by-line (VL1 to VLn, HL1 to HLk) separated by a line-by-line histogram values (P ₁ to P _k , Q ₁ to Q _n ) are calculated and at least one histogram 24 and 26 is generated based on the calculated line-by-line histogram value.

For example, the gray scale converter 12 converts the input image 22 of the RGB format into a gray scale obtained by averaging RGB values, and converts pixels of each line unit of the input image 22 (eg, the "HL1" line). In this case, each of the sums of a ₁₁ to a _1n ) may be calculated and at least one histogram 24 and 26 may be generated based on the calculation result.

The line unit may be based on at least one of the vertical directions VL1 to VLn or the horizontal directions HL1 to HLk, and includes the histogram values P ₁ to P _k and Q ₁ to Q _{n for each line.} ) May be the total sum of the pixels constituting the line units VL1 to VLn and HL1 to HLk.

For example, the gray scale converter 12 converts an input image composed of a plurality of pixels to gray scale, and adds the sum of pixels of each of the horizontal lines HL1 to HLk of the input image 22 line by line. Calculate and generate a first histogram 24 based on the calculation result, calculate a sum of pixels of each of the vertical lines VL1 to VLn of the input image line by line, and generate a second histogram based on the calculation result. (26) can be generated.

In more detail, when the pixel value converted from the gray scale is “0” to “255”, the gray scale converter 12 may include horizontal lines HL1 to HLk constituting the input image 22. ) Calculates the sum of the pixels a ₁₁ to a _kn constituting each line by line, generates a histogram bar corresponding to the calculation result, and generates a first histogram 24 based on the generated histogram bars. Can be.

In addition, the gray scale converter 12 calculates the sum of pixels a ₁₁ to a _kn constituting each of the vertical lines VL1 to VLn constituting the input image 22 for each line. Each histogram bar corresponding to the calculation result may be generated, and a second histogram 26 may be generated based on the generated histogram bars.

The histogram values P ₁ to P _k and Q ₁ to Q _n for each line may be the total sum of pixels per line, but are not limited thereto.

The histogram values P ₁ to P _k and Q ₁ to Q _{n for} each line according to another exemplary embodiment of the present invention may be sums reflecting averages or weights of pixels per line. Here, the average value may be an arithmetic mean, geometric mean, harmonic mean, weighted average, and the like.

The blurring unit 14 divides the at least one histogram 24 and 26 generated by the gray scale converter 12 into at least one interval, and averages the histogram value of each line by the divided at least one interval. Calculate and generate at least one histogram 28 and 30 based on the result of the calculation.

For example, the blurring unit 14 divides the first histogram 24 into first interval units, calculates an average of the histogram values for each line for each divided first interval, and based on the calculation result, the third histogram 28. Can be generated.

For example, when the total interval of the first histogram 24 is "256", the blurring unit 14 divides the first histogram 24 into a first interval (for example, "2") and divides the first. Calculate the average of the histogram values for each line constituting the first interval (eg, "2") for each interval and calculates the result PM ₁ To PM _l ) may be respectively generated, and a third histogram 28 may be generated based on the generated histogram bars.

In addition, the blurring unit 14 divides the second histogram 26 in units of a second interval, calculates an average of histogram values for each line for each divided second interval, and based on a calculation result, the fourth histogram 30. Can be generated.

For example, when the total interval of the second histogram 26 is "256", the blurring unit 14 divides the second histogram 26 into a second interval (for example, 2), and divides each of the divided second intervals. The average of the histogram values for each line constituting the second interval (eg, “2”) is calculated and the calculation result is QM _1. To QM _j ), respectively, may be generated and a fourth histogram 30 may be generated based on the generated histogram bars.

The first interval unit and the second interval unit may be set equal to or different from each other.

The third histogram 28 and the fourth histogram 30 generated by the blurring unit 14 are blurring images compared to the first histogram 24 and the second histogram 26. .

Accordingly, since the noise included in the input image 22 (eg, an image error generated in the input image 22 itself) may be removed, the image inspector 10 may determine a subject included in the input image 22. For example, the position of the text TEXT of FIG. 2 can be accurately corrected.

The edge detector 16 divides the at least one histogram 28 and 30 generated by the blurring unit 14 into at least one interval and changes the amount of change in the histogram value for each of the divided at least one interval H ₁ to H _l. Or V ₁ to V _j ) and generate at least one histogram 32 and 34 based on the detection result.

For example, the edge detector 16 divides the third histogram 28 into third interval units (for example, "1"), detects a change amount of the divided histogram value for each third interval, and detects the detection results (H ₁ to ₁₎ . The fifth histogram 32 may be generated based on H ₁ ).

)을 상기 히스토그램 막대로서 생성할 수 있다. More specifically, the edge detector 16 generates the histogram bar for any one section of the third histogram 28 corresponding to the first interval, when the third interval of the third histogram 28 is generated. The sum of the histogram values constituting the one section among the sections divided into units and the difference value of the sum of the histogram values constituting the other section adjacent to the one section (eg, H ₁ , H ₁₎ =

) Can be generated as the histogram bar.

In this case, the first interval unit and the third interval unit may be the same.

In addition, the edge detector 16 divides the fourth histogram 30 into fourth intervals (for example, “1”), detects an amount of change in the histogram value for each divided fourth interval, and detects the detection results (V ₁ to V). _j ) may generate the sixth histogram 34.

)을 상기 히스토그램 막대로서 생성할 수 있다. 이때 상기 제2 간격과 상기 제4 간격은 같을 수 있다.More specifically, the edge detector 16 generates the histogram bar for any one section of the fourth histogram 30 corresponding to the second interval, when the fourth interval of the fourth histogram 30 is generated. The sum of the histogram values constituting the one section V ₁ among the sections divided into units and the difference value of the sum of the histogram values constituting the other section V ₂ adjacent to the one section (for example, , V ₁ , V ₁ =

) Can be generated as the histogram bar. In this case, the second interval and the fourth interval may be the same.

The comparison determiner 18 may compare the histograms 32 and 34 output from the edge detector 16 with the reference histogram, and detect whether the input image 22 has a position error based on the comparison result.

The reference histogram is a histogram that is a reference for detecting a position error of the input image 22 as a histogram of the original image. The reference histogram is a histogram of an image including a position of a desired subject TEXT.

For example, the reference histogram may include a seventh histogram (not shown) and an eighth histogram (not shown). In this case, the comparison determiner 18 may compare the fifth histogram 32 with the original image. A seventh histogram (not shown) may be compared and it may be detected whether the fifth histogram 32 and the seventh histogram (not shown) match.

In addition, the comparison determiner 18 compares the eighth histogram (not shown) of the original image compared with the sixth histogram 34, and the sixth histogram 34 and the eighth histogram (not shown) are compared. It can be detected whether or not a match.

The correction unit 19 may generate a position correction signal based on the comparison result of the comparison determination unit 18.

The position correction signal causes the printing apparatus 20 to shift the position of a subject (TEXT, that is, an object to be photographed of the input image 22) constituting the input image 22, so that the input image 22 is moved. This signal is made to match the original video.

In addition, when the image inspector 10 is implemented in the printing apparatus 20, the position correction signal may be transmitted to a subject transfer unit (not shown), and the subject transfer unit (not shown) may respond to the position correction signal. The input image 22 may be matched with the original image by moving the position of the subject.

That is, the image inspector 10 according to an embodiment of the present invention separates the two-dimensional input image 22 into at least one axis in one dimension and compares it with the original image to determine whether the input image 22 has a high positional error. And position compensation.

In addition, the image inspector 10 according to an embodiment of the present invention is used in a machine vision system such as silicon manufacturing in the semiconductor industry, printing industry publishing, or steelmaking in the steel industry, at high speed during quality inspection of finished products. It has the effect of detecting and correcting errors.

For example, the image inspector 10 according to an exemplary embodiment of the present invention may increase the speed of an existing inspection apparatus when implemented in an inspection apparatus that operates at a high speed (eg, the printing industry using an inspection technique using a camera). It can maximize the utilization of various industries that produce continuous objects at high speed, such as paper, paper, and film production.

8 is a flowchart of an image inspection method according to an exemplary embodiment. 1 to 8, the gray scale converter 12 converts the input image 22 to gray scale, and converts the converted input image in line units (eg, VL1). Or HL1) to calculate a histogram value (eg, P ₁ to P _k , Q ₁ to Q _n ) for each line (eg, VL1 to VLn, HL1 to HLk), and calculate the calculated histogram value for each line (P ₁ to P). First histograms 24 and 26 are generated based on _k and Q ₁ to Q _n (S10).

The blurring unit 14 divides the first histogram into units of a first interval, calculates an average of the histograms per line for each divided first interval, and generates second histograms 28 and 30 based on the calculation result. (S12).

The edge detector 16 divides the second histograms 28 and 30 in units of a second interval and changes the amount of change in the second histogram value H ₁ to H _l or V ₁ to V _{j for} each divided second interval. Is detected and third histograms 32 and 34 are generated based on the detection result (S14).

The third histograms 32 and 34 are compared with the previously stored reference histograms, and based on the comparison result, whether the input image 22 is in position error is detected (S16).

The invention can also be embodied as computer readable code on a computer readable recording medium. The computer-readable recording medium includes all kinds of recording devices in which data that can be read by a computer system is stored.

Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage device, and the like. The program code for performing the image inspection method according to the present invention may include a carrier wave ( For example, transmission via the Internet).

The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion. And functional programs, codes and code segments for implementing the present invention can be easily inferred by programmers in the art to which the present invention belongs.

Although the present invention has been described with reference to one embodiment shown in the drawings, this is merely exemplary, and those skilled in the art will understand that various modifications and equivalent other embodiments are possible therefrom. Therefore, the true technical protection scope of the present invention will be defined by the technical spirit of the appended claims.

BRIEF DESCRIPTION OF THE DRAWINGS In order to better understand the drawings cited in the detailed description of the invention, a brief description of each drawing is provided.

1 is a block diagram of an image inspector according to an exemplary embodiment of the present invention.

FIG. 2 is a machine vision system including the image inspector of FIG. 1.

FIG. 3 is a diagram for describing a process of generating first to fourth histograms by the image inspector of FIG. 1.

4 is first and second histograms generated by the image inspector of FIG. 1.

5 is a third and fourth histograms generated by the image inspector of FIG. 1.

FIG. 6 is a diagram for describing a process of generating fifth and sixth histograms by the image inspector of FIG. 1.

FIG. 7 is a fifth and a sixth histogram generated by the image inspector of FIG. 1.

8 is a flowchart of an image inspection method according to an exemplary embodiment.

Claims (13)

Converting the input image to gray scale, dividing the converted input image by line unit, calculating a histogram value for each line, and generating a first histogram based on the calculated histogram value for each line; Dividing the first histogram by a first interval unit, calculating an average of the histograms per line for each of the divided first intervals, and generating a second histogram based on a calculation result; Dividing the second histogram by a second interval unit, detecting a change amount of the second histogram value for each divided second interval, and generating a third histogram based on a detection result; And Comparing the third histogram with a pre-stored reference histogram and detecting whether there is a position error of the input image based on a comparison result. The method of claim 1, wherein the line unit, An imaging inspection method based on at least one of the vertical direction and the horizontal direction. According to claim 1, wherein the histogram value for each line, And a total sum or an average value of the pixels constituting the line unit. The method of claim 1, wherein the image inspection method comprises: Generating a position correction signal based on the comparison result, wherein the position correction signal is transmitted to a printing apparatus, and the printing apparatus corrects the position of a subject constituting the input image in response to the position correction signal. Imaging test method. The method of claim 1, wherein the amount of change in the second histogram value for each second interval is Based on the sum of the histogram values constituting any one of the sections of the second histogram divided by the second interval unit and the difference value of the sum of the histogram values constituting the other section adjacent to the one section Image inspection method generated by. A recording medium readable by a computer program which records a computer program for performing the method of claim 1. A gray scale converter configured to convert the input image to gray scale, calculate the line histogram value by dividing the converted input image by line units, and generate a first histogram based on the calculated line histogram value; A blurring unit for dividing the first histogram into units of a first interval, calculating an average of the histograms per line for each of the divided first intervals, and generating a second histogram based on a calculation result; An edge detector for dividing the second histogram by a second interval unit, detecting an amount of change in the second histogram value for each divided second interval, and generating a third histogram based on a detection result; And And a comparison determiner configured to compare the third histogram with a pre-stored reference histogram and detect a position error of the input image based on a comparison result. The method of claim 7, wherein the line unit, An image inspector based on at least one of the vertical direction and the horizontal direction. The image inspector of claim 7, wherein the histogram value for each line is a total sum or an average value of pixels constituting the line unit. The method of claim 7, wherein the image inspector, A correction unit configured to generate a position correction signal based on the comparison result, wherein the position correction signal is transmitted to a printing apparatus, and the printing apparatus corrects the position of a subject constituting the input image in response to the position correction signal. Video checker. The method of claim 7, wherein the amount of change in the second histogram value for each second interval, Based on the sum of the histogram values constituting any one of the sections of the second histogram divided by the second interval unit and the difference value of the sum of the histogram values constituting the other section adjacent to the one section Generated by the image inspector. A machine vision system comprising the image inspector of claim 7. A camera used for quality inspection of a product comprising said image inspector of claim 7.

Images