JP2003281538A - Image processor, image inspecting method and printing failure detecting method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide more detailed inspection in inspection of a printed state of an arced image for a missing spot or a thin spot or the like. <P>SOLUTION: A missing spot ratio of the arced image is calculated by conducting polar coordinate transformation on the arced image and determining a number of missing spots from a pattern of a maximum projection value (linearity) determined from a plurality of candidates. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus, an image inspection method, and a print defect detection method for inspecting a printing state such as a missing or faint arc image of a stamp.

[0002]

2. Description of the Related Art Conventionally, in the case of inspecting a printing state such as a lack or a blur of an arc image such as a stamp, the printing state in an object image area is inspected from an average value of a density histogram in the area obtained by extracting the area. Alternatively, the stamp image of which the area has been extracted is divided into N * M, individual histograms are analyzed, and it is checked whether or not the printing state is appropriate.

Further, when the circular arc pattern is known, reference patterns of circular arcs are created, raster scanning is performed on them, the target image with the highest matching is detected, and the detected circular arc pattern is analyzed in the whole or divided histogram as described above. Then, the print state is inspected.

The arc detection by pattern matching is good if the arc is the same as the reference pattern, but if it deviates from the reference pattern, detection becomes difficult. On the other hand, the histogram analysis of the print inspection has a drawback in that it is possible to inspect the entire or local density, but it is difficult to perform more detailed inspection.

[0005]

DISCLOSURE OF THE INVENTION The present invention is intended for inspecting a printing state of a circular arc image such as chipping or blurring.
An object of the present invention is to provide an image processing apparatus, an image inspection method, and a print defect detection method that can perform more detailed inspection.

[0006]

In an image processing apparatus of the present invention for inspecting a loss of an arc image, an image pickup means for picking up the arc image and a picked up image by this image pickup means are subjected to a differential process to be binarized. First processing means for obtaining an image, second processing means for setting a virtual frame of an arc in the binarized image obtained by the first processing means to obtain a plurality of polar coordinate conversion patterns, and second processing means. From the linearity evaluation of a plurality of polar coordinate conversion patterns obtained by the processing means, a selection means for uniquely selecting a pattern for the chipping detection processing, and detecting the number of chips from the pattern uniquely selected by this selection means,
And a third processing means for obtaining a chipping ratio.

In the image processing apparatus of the present invention for inspecting a loss of an arc image, an image pickup means for picking up the arc image and a first image obtained by differentiating the image picked up by the image pickup means to obtain a binarized image. The processing means, the second processing means for setting a virtual frame of an arc in the binarized image obtained by the first processing means to obtain a plurality of polar coordinate conversion patterns, and the second processing means A selection means for uniquely selecting a pattern for a chip detection process from linearity evaluations of a plurality of polar coordinate conversion patterns; a chip position detection device for detecting a chip position from the pattern uniquely selected by the selection means; A polar coordinate conversion means for polar coordinate conversion of an image picked up by the image pickup means by a virtual frame of an arc which is uniquely detected, and a chipped area detected by the chipped position detection means are used to extract the chipped area. And only region extraction means, the extraction area by the lack region extracting means and a histogram calculation, and a evaluation means for evaluating a printing failure of chipping area of the arc image in the plurality of thresholds.

[0008]

DETAILED DESCRIPTION OF THE INVENTION An image processing apparatus according to an embodiment of the present invention will be described below with reference to the drawings. The image processing device
This is for inspecting a printing state such as a lack of an arc image of a printing stamp or a blur, and as shown in FIG.
A / D conversion unit 2, multi-valued image memory 3, differentiation processing units 4, 2
Quantization unit 5, region extraction unit 6, virtual frame setting unit 7, first polar coordinate conversion unit 8, maximum projection value detection unit 9, chipping processing pattern selection unit 10, chipping detection processing unit 11, chipping ratio calculation unit 1
2, a second polar coordinate conversion unit 13, a missing region extraction unit 14, a histogram calculation unit 15, and a threshold value evaluation unit 16. The image pickup unit 1 picks up an image of the paper P on which the print stamp is printed, outputs a two-dimensional analog image signal of the print stamp, and is configured by a two-dimensional camera.

The A / D conversion unit 2 converts the two-dimensional analog image signal of the print stamp from the image pickup unit 1 into a multivalued digital image signal (input multivalued image) as shown in FIG. Is.

The multi-valued image memory 3 stores a multi-valued digital image signal (input multi-valued image) from the A / D converter 2. The differential processing unit 4 converts the multivalued digital image signal from the multivalued image memory 3 into a multivalued differential image (differential multivalued image) as shown in FIG.

The binarization unit 5 converts the multivalued differential image from the differential processing unit 4 into a binary image signal (differential binary image) as shown in FIG. The area extraction unit 6
The differential binary image from the binarization unit 5 is labeled to detect the circumscribing rectangle of the largest size as shown in FIG.

The virtual frame setting unit 7 compares the vertical or horizontal size of the circumscribing rectangle detected by the region extraction unit 6 with a known circular diameter size, and selects the closest circular diameter size as the reference size. A square circumscribing rectangle is created using this reference size, and the created virtual frame is set in a binary differential image as shown in FIG. In FIG. 6, a frame position is uniquely assigned in the horizontal direction because the reference size is in the horizontal direction, but in the vertical direction it is assumed that the upper side is the starting point and the lower side is the starting point (the thick broken line). The case frame (thin broken line) is set.

The first polar coordinate conversion unit 8 includes a virtual frame setting unit 7
The polar coordinate conversion is performed for the virtual frame created in the binary differential image. For example, FIG. 7 shows an example in which the polar coordinate conversion range 1 is an example surrounded by a circle obtained from a square circumscribing rectangle with the upper side of FIG. 6 as a reference. FIG. 8 shows an example in which the polar coordinate conversion range 2 is an example surrounded by a circle obtained from a square circumscribing rectangle with the lower side of FIG. 6 as a reference.

In FIG. 7, polar coordinate conversion is performed from the outer broken line to the inner broken line to obtain a polar coordinate conversion result as shown in FIG. Similarly, in FIG. 8 as well, polar coordinate conversion is performed from the outer broken line to the inner broken line, and the polar coordinate conversion result as shown in FIG. 10 is obtained. The broken lines on the inside of FIGS. 7 and 8 are created in consideration of the line width of the circle. The resolution of the polar coordinate conversion pattern from the circle to the line is 360 degrees at maximum, where 1 degree is one pixel.
9 and 10, the horizontal axis is 0 to 360 degrees,
The vertical axis represents the radial distance.

The maximum projection value detection unit 9 integrates the black projection values in the horizontal direction in the pattern converted from the orthogonal system to the polar coordinate system from the first polar coordinate conversion unit 8 to obtain the largest projection value. Is to detect. The maximum projection value as shown in FIG. 11 is output for the polar coordinate conversion pattern of FIG. 9, and the maximum projection value as shown in FIG. 12 is output for the polar coordinate conversion pattern of FIG.

The chipping processing pattern selection unit 10 compares a plurality of maximum projection values from the maximum projection value detection unit 9 and selects a pattern having a large projection value. For example, comparing the maximum projection value in FIG. 12 of the maximum projection value in FIG. 11,
A pattern having a large maximum projection value in FIG. 11 is selected.

The chipping detection processing unit 11 uses the polar coordinate conversion pattern of the maximum projection value from the chipping processing pattern selection unit 10 to determine the number of chips, which is one from 0 to 360 degrees, as shown in FIG. It is something to detect. That is, the number of consecutive "0" s in the data for 360 degrees of the Cartesian coordinates at the radial position of the maximum projection value is the number of defects.

The deficiency ratio calculation unit 12 includes a deficiency detection processing unit 1.
The result (result output 1) of calculating the chip ratio from the number of chips detected by 1 is output.

FIG. 13 shows the pattern of FIG. In FIG. 13, scanning is performed from the upper level to the lower level, and a threshold value represented by a broken line (for example, in FIG.
The number of scanning lines that exceeds the inner circumference of FIG. 5) is counted and taken as the number of defects. Then, the chip ratio is obtained by dividing the number of chips by 360. You may make it judge whether this multiplication rate is in an allowable range, and output this judgment result.

The differential processing by the differential processing section 4 is shown in FIG.
As shown in FIG. 4, the vertical and horizontal operations of Sobel are scanned and ORed (by taking the logical sum) to obtain a differential multi-valued image. The polar coordinate conversion processing by the first polar coordinate conversion unit 8 converts black pixels and white pixels in a polar coordinate conversion range enclosed by a circle obtained from a reference square circumscribed rectangle into rectangular coordinates.

For example, an example in which the radial position from the center of the circle is r and the angle is θ will be described with reference to the flowchart shown in FIG. First, 0 degree (ST1, 2) at the radial position r = 1 is checked for black or white (ST
3) In the case of black, "1" is recorded in the Cartesian coordinates corresponding to the radial position r and the angle θ (ST4). After this, the angle θ
Is changed (ST5), the above steps 3 and 4 are repeated, and when one round is examined (ST6), the radial position is changed (ST7) and the process returns to step 2. When the outer circumference of the circle is completed (ST8), the process is completed.

The second polar coordinate converter 1 shown in FIG.
3 is a polar coordinate pattern (multivalued image) of the multivalued image itself using the rectangular area (FIG. 6) in which the maximum projection value of the multivalued digital image signal of the multivalued image memory 3 is obtained, as shown in FIG. Image conversion pattern). As shown in FIG. 16, the chipped area extracting unit 14 defines a surface from the position information of the chipping detected by the chipping detection processing unit 11 and the scanning start point to the threshold (a chipped area extraction portion, a shaded portion).
Is to be extracted as. Histogram calculator 15
Is to calculate a histogram for the missing area extracted by the missing area extracting unit 14, as shown in FIG.

The threshold evaluation unit 16 is a histogram calculation unit 15
As shown in FIG. 17, the calculation result (histogram) of
The threshold value 1 (Th1) and the threshold value 2 (Th2) are evaluated, and the result (result output 2) is output. For example, it is determined whether the threshold value 1 is printed but the density is low, the threshold value 2 is not printed, and the like, and the cause of the print failure is mechanical one-sided ejection or defective ink output. Etc.

Next, the operation of the above configuration will be described. That is, the two-dimensional analog image signal of the print stamp input from the image pickup unit 1 is converted into a multivalued digital image signal (FIG. 2) in the A / D conversion unit 2 and stored in the multivalued image memory 3. The multi-valued digital image signal sent from the multi-valued image memory 3 is converted into a multi-valued differential image in the differential processing section 4, and is converted into a binary image signal (FIG. 4) in the next binarization section 5. It

Next, the area extraction unit 6 performs a labeling process on the differential binary image of FIG. 4 to obtain the circumscribed rectangle of the largest size as shown in FIG. Next, the virtual frame setting unit 7 compares the detected vertical or horizontal size of the circumscribed rectangle with the known circular diameter size, and selects the one having the closest circular diameter size as the reference size. A square circumscribing rectangle (virtual frame) is created using this reference size. Then, the virtual frame created in the binary differential image is set, and polar coordinate conversion is performed. In FIG. 6, since the horizontal direction is the reference size, the frame position is uniquely assigned, but in the vertical direction, assuming that there is a chip, the upper side is the starting point (thick broken line) and the lower side is the starting point. The case frame (thin broken line) is set. FIG. 7 shows an example in which the polar coordinate conversion range 1 is an example surrounded by a circle obtained from a square circumscribing rectangle with the upper side of FIG. 6 as a reference. FIG. 8 shows an example in which the polar coordinate conversion range 2 is an example surrounded by a circle obtained from a square circumscribing rectangle based on the lower side of FIG. Figure 7
Is obtained by performing polar coordinate conversion from the outer broken line to the inner broken line to obtain FIG. Similarly, in FIG. 8 as well, polar coordinate conversion is performed from the outer broken line to the inner broken line to obtain FIG. 7,
The inner broken line in FIG. 8 is created in consideration of the line width of the circle. The resolution of the polar coordinate conversion pattern from the circle to the line is 360 degrees at maximum, where 1 degree is one pixel.

For the pattern converted from the orthogonal system to the polar coordinate system by the first polar coordinate conversion unit 8, the projection values of black are integrated in the horizontal direction, and the maximum projection value detection unit 9 detects the projection value having the largest value. To do. 9 and 11 are patterns obtained by converting FIGS. 7 and 8 into a polar coordinate system. FIG. 9 is the conversion pattern of FIG. 7, and FIG.
0 is the conversion pattern of FIG. The maximum projection value detection unit 9 compares the maximum projection value in FIG. 9 with the maximum projection value in FIG.
A pattern having a large projection value is selected by the chipping processing pattern selection unit 10. That is, FIG. 9 is selected in FIGS. 9 and 11.

Next, the defect detection processing unit 11 detects the number of defects using the polar coordinate conversion pattern, and the defect ratio calculation unit 1
Calculate the chipping rate in 2 and output the result. FIG. 13 is the pattern of FIG. In FIG. 13, scanning is performed from the higher order to the lower order, and the number of scanning lines that exceeds a threshold value (for example, the inner circumference in FIGS. 7 and 8) represented by a broken line is counted and set as the number of defects. Then, the chip ratio is obtained by dividing the number of chips by 360.

Next, the second polar coordinate conversion unit 13 uses the rectangular area (FIG. 7) in which the maximum projection value is calculated for the multi-valued digital image signal of the multi-valued image memory 3 as the polar coordinate pattern of the multi-valued image itself. Convert to. The missing area extraction unit 14 extracts the missing position information obtained by the missing detection processing unit 11 and the area from the scanning start point to the threshold value as a surface. FIG. 16 shows the detected area. The shaded area is the area. Then, in the histogram calculation unit 15, as shown in FIG.
Histogram processing. Next, the histogram is set to the threshold 1 (T
h1) and the result of evaluation with the threshold value 2 (Th2) are output.
For example, it is determined whether the threshold value 1 is printed but the density is low, the threshold value 2 is not printed, and the like, and the cause of the print failure is mechanical one-sided ejection or defective ink output. Etc.

As described above, a more accurate defect rate can be calculated by polar-transforming the arc image and determining the number of defects from the pattern of the maximum projection value (linearity) obtained from a plurality of candidates. Further, the arc image is subjected to polar coordinate conversion, the missing area is detected from the pattern of the maximum projection value (linearity) obtained from a plurality of candidates, and the histogram is evaluated, so that the factor of the printing failure can be immediately fed back to the device.

[0030]

As described above in detail, according to the present invention, an image processing apparatus, an image inspection method, and a printing defect that can perform more detailed inspection in an inspection of a printing state such as a lack of an arc image or a blur. A detection method can be provided.

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration of an image processing apparatus for explaining an embodiment of the invention.

FIG. 2 is a diagram for explaining an input multi-valued image.

FIG. 3 is a diagram for explaining a differential multi-valued image.

FIG. 4 is a diagram for explaining a differential binary image.

FIG. 5 is a diagram for explaining detection of a circumscribing rectangle.

FIG. 6 is a diagram for explaining setting of a virtual frame.

FIG. 7 is a diagram for explaining a polar coordinate conversion range.

FIG. 8 is a diagram for explaining a polar coordinate conversion range.

9 is a diagram for explaining the polar coordinate conversion result for FIG. 7. FIG.

10 is a diagram for explaining a polar coordinate conversion result with respect to FIG.

FIG. 11 is a diagram for explaining projection values for FIG. 9.

FIG. 12 is a diagram for explaining projection values for FIG. 10.

FIG. 13 is a diagram for explaining the detection result of the number of chips each of which is 0 to 360 degrees.

FIG. 14 is a diagram for explaining differential processing by a differential processing unit.

FIG. 15 is a flowchart for explaining polar coordinate conversion processing.

FIG. 16 is a diagram for explaining a multi-valued image conversion pattern and a missing area extraction portion.

FIG. 17 is a diagram for explaining a histogram for a lack area and a threshold value for evaluating the histogram.

[Explanation of symbols]

1 ... Imaging unit, 2 ... A / D conversion unit, 3 ... Multi-valued image memory,
4 ... Differentiation processing unit, 5 ... Binarization unit, 6 ... Region extraction unit, 7 ...
Virtual frame setting unit, 8 ... First polar coordinate conversion unit, 9 ... Maximum projection value detection unit, 10 ... Missing processing pattern selection unit, 11 ... Missing detection processing unit, 12 ... Missing ratio calculation unit, 13 ... Second Polar coordinate conversion unit, 14 ... Missing area extraction unit, 15 ... Histogram calculation unit, 16 ... Threshold evaluation unit.

Claims (6)

[Claims] 1. An image processing apparatus for inspecting a loss of an arc image, comprising: an image pickup means for picking up the arc image; and a first processing means for differentiating the image picked up by the image pickup means to obtain a binarized image. Secondly, a virtual frame of an arc is set in the binarized image obtained by the first processing means to obtain a plurality of polar coordinate conversion patterns.
Processing means, a selection means for uniquely selecting a pattern for the defect detection processing from the linearity evaluation of a plurality of polar coordinate conversion patterns obtained by the second processing means, and a selection means uniquely selected by this selection means. An image processing apparatus comprising: a third processing unit that detects the number of deficiencies from a pattern and obtains a deficiency ratio. 2. An image processing apparatus for detecting a printing failure in a missing area of an arc image, wherein: an image pickup means for picking up the arc image; and a first image obtained by differentiating the picked up image by the image pickup means to obtain a binarized image. A processing means and a second processing for setting a virtual frame of an arc in the binarized image obtained by the first processing means to obtain a plurality of polar coordinate conversion patterns.
Processing means, a selection means for uniquely selecting a pattern for the defect detection processing from the linearity evaluation of a plurality of polar coordinate conversion patterns obtained by the second processing means, and a selection means uniquely selected by this selection means. A chipping position detecting unit for detecting a chipping position from a pattern, a polar coordinate converting unit for polar coordinate conversion of an image picked up by the image pickup unit with a virtual frame of an arc uniquely detected, and a chipping position detected by the chipping position detecting unit are used. Then, a missing area extracting means for extracting the missing area, and an evaluating means for calculating a histogram of the extracted area by the missing area extracting means and evaluating defective printing of the missing area of the arc image with a plurality of thresholds are provided. An image processing device characterized by: 3. An image pickup means for picking up an arc image, a first converting means for converting the picked-up image by the image pickup means into an input multi-valued image, and a differential multi-valued image for the input multi-valued image by the first converting means. Second conversion means for converting into an image, third conversion means for converting the differential multi-valued image by the second conversion means into a differential binary image, and labeling of the differential binary image by the third conversion means A circumscribing rectangle detecting means for processing and detecting the circumscribing rectangle of the largest size; a virtual frame setting means for setting a virtual frame of a plurality of arcs based on the circumscribing rectangle detected by the circumscribing rectangle detecting means; Polar coordinate conversion means for converting the differential binary image by the conversion means into a polar coordinate conversion pattern from a circle to a line based on the virtual frame by the virtual frame setting means, and a plurality of polar coordinates obtained by the polar coordinate conversion means. A selection means for uniquely selecting a pattern for the defect detection process from the linearity evaluation of the replacement pattern, a detection means for detecting the number of defects from the pattern uniquely selected by this selection means, and a defect detected by this detection means. An image processing apparatus comprising: a calculating unit that calculates a defect ratio based on a number. 4. An image pickup means for picking up an arc image, a first converting means for converting an image picked up by the image pickup means into an input multi-valued image, and a differential multi-valued image for the input multi-valued image by the first converting means. Second conversion means for converting into an image, third conversion means for converting the differential multi-valued image by the second conversion means into a differential binary image, and labeling of the differential binary image by the third conversion means A circumscribing rectangle detecting means for processing and detecting the circumscribing rectangle of the largest size; a virtual frame setting means for setting a virtual frame of a plurality of arcs based on the circumscribing rectangle detected by the circumscribing rectangle detecting means; And a first polar coordinate conversion means for converting the differential binary image by the conversion means into a polar coordinate conversion pattern from a circle to a line based on the virtual frame by the virtual frame setting means, and the first polar coordinate conversion means. Was Selecting means for uniquely selecting a pattern for the chipping detection processing from linearity evaluation of a number of polar coordinate conversion patterns, a chipping position detecting means for detecting a chipping position from the pattern uniquely selected by the selecting means, and the above-mentioned imaging Second polar coordinate conversion means for performing polar coordinate conversion on the image picked up by the means using a virtual frame of an arc that is uniquely detected, and chipped area extraction means for extracting the chipped area using the chipped position detected by the chipped position detection means And an evaluation unit that calculates a histogram of the extracted region by the defective region extraction unit and evaluates the defective printing of the defective region of the arc image with a plurality of thresholds. 5. An image inspection method for inspecting a loss of an arc image, wherein an arc image is captured, the captured image is differentially processed to be converted into a binary image, and a virtual image of a plurality of arcs is added to the binary image. A frame is set, converted into a polar coordinate conversion pattern based on the virtual frames of the plurality of arcs, and a pattern for chipping detection processing is uniquely selected from the linearity evaluation of the plurality of polar coordinate conversion patterns. An image inspection method characterized in that the number of defects is detected from the formed pattern and the ratio of the defects is obtained. 6. A printing defect detection method for detecting a printing defect in a missing region of an arc image, wherein an arc image is captured, the captured image is subjected to differential processing to be converted into a binarized image, and this binarized image is obtained. A virtual frame of a plurality of circular arcs is set, converted into a polar coordinate conversion pattern based on the virtual frames of a plurality of circular arcs, and a pattern for chipping detection processing is uniquely selected from the linearity evaluation of the multiple polar coordinate conversion patterns. Detecting a chipping position from this uniquely selected pattern, performing polar coordinate conversion on the captured image with a virtual frame of a uniquely detected arc, and extracting the chipping region using the detected chipping position, A method of detecting a printing defect, wherein a histogram calculation is performed on the extraction region, and the printing defect in the missing region of the arc image is evaluated by a plurality of thresholds.