JP2882273B2 - Black pattern extraction method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a black pattern extraction method for automatically extracting a black-printed pattern from a background by superimposing it on an object on which a single-color pattern / character other than black is printed as a background.

[0002]

2. Description of the Related Art Conventionally, printing inspection / recognition of a black pattern printed on a background printed in a single color other than black is performed by dropping out a background color by using a color filter together with a black-and-white camera. A fixed or automatic uniform or local binarization process based only on the shading of the luminance signal is performed. Even in the method using a color camera, RGB
Or simple switching of Y single color, or
After performing monochrome conversion by table conversion from RGB, the same processing as that of the monochrome camera is performed.

[0003]

In the conventional method using dropout by a color filter, a single color (blue, red, etc.) is used as a background on a base (white, gray, brown, etc.) like a cardboard carton for food. , Green, yellow, etc.) for various workpieces with printed color filters,
Lighting switching and fine adjustment of the optical system are required, and much labor is required. A method of detecting the background printing color of the object with a color sensor and selecting an appropriate color filter is also conceivable. However, there is a problem that a long switching time is required in a mixed line.

In the method using a color camera,
In the simple selection of RGB, sufficient background removal cannot be performed for background printing colors other than red / blue / green. When converting to a single color by table conversion, it is effective if the target work does not change, but a large number of conversion tables according to the type of work are required in the mixed flow line. Since the number of accesses and the area are large, it is not suitable for software processing by a CPU equipped with a cache function, and hardware needs to be provided to obtain a sufficient execution speed.

Accordingly, it is an object of the present invention to extract a black print pattern efficiently by software, automatically removing a two-color background with a small switching load even in a mixed flow line.

[0006]

In order to solve the above-mentioned problems, a black pattern extraction method according to the present invention creates a density histogram in a designated area for each of multiple color single color images captured by an image input means. Calculating the average of each created density histogram, calculating the slice value based on the density histogram of the image in which the calculated average is the maximum, binarizing this image using the slice value, Extract the background + black area within the specified area,
For the extracted area, create density histograms of the other original colors except for the first processed color, calculate the average of the generated density histograms, and calculate the density of the image in which the calculated average is the minimum. A slice value is calculated based on the histogram, the image is binarized using the slice value, and a black portion in the designated region is extracted.

[0007]

Next, a first embodiment of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of the first embodiment. The image input unit 1 has a background in which a monochromatic pattern / character other than black is printed as a background, and a black pattern is superimposed thereon. The image in the designated area of the printed object is
The image is captured as a GBY quaternary color single color image.

A density histogram calculation unit 2 calculates a density histogram for each of RGBY colors in a designated area in the RGBY space for the color image captured by the image input unit 1.

The variance calculator 3 calculates the variance of the density histogram of each color calculated by the density histogram calculator 2.

The image selecting section 4 selects an image of a color in which the variance of the density histogram of each color calculated by the variance calculating section 3 is the maximum among the RGBY colors.

[0012] The slice value calculating section 5 is provided with the image selecting section 4.
Calculates a slice value from the density histogram of the image selected in.

The background + black part extraction unit 6 extracts a background + black part in a designated area from a corresponding density histogram based on the slice value calculated by the slice value calculation unit 5.

The density histogram calculator 7 calculates the background +
A density histogram of three colors not selected by the image selection unit 4 is calculated for the background + black portion extracted by the black portion extraction unit 6.

The variance calculator 8 calculates the variance of each density histogram calculated by the density histogram calculator 7.

The image selector 9 selects an image having the minimum value among the variances calculated by the variance calculator 8.

The slice value calculator 10 calculates a slice value of the density histogram of the image selected by the image selector 9.

The black part extraction unit 11 extracts a black part in the designated area from the corresponding density histogram based on the slice value calculated by the slice value calculation unit 10.

The background + black portion extraction unit 12 calculates the slice value based on the slice value calculated by the slice value calculation unit 10.
The background + black portion in the specified area is extracted from the corresponding density histogram.

The slice value calculation unit 13 calculates a slice value from the background + black portion density histogram extracted by the background + black portion extraction unit 12.

The black part extracting unit 14 extracts a black part in a designated area from the density histogram extracted by the background + black part extracting unit 12 based on the slice value calculated by the slice value calculating unit 13.

Next, the operation of the first embodiment will be described with reference to FIG.
This will be described with reference to FIGS. 2, 3, 4, and 5. FIG.
FIGS. 3, 4 and 5 show a density histogram and its slice values at each processing stage.

First, in the image input unit 1, an image of a designated area of an object in which a single color pattern / character other than black is printed as a background as a background, and a black pattern is printed over the background, is an RGBY quaternary. The quaternary color single-color image captured by the image input unit 1 is captured by the density histogram calculation unit 2 as RGB.
In the Y space, a density histogram for each of the RGBY colors in the designated area is calculated.

Next, the variance calculator 3 calculates the variance of the density histogram of each color calculated by the density histogram calculator 2, and the image selector 4 calculates
From among the RGBY colors, an image of a color having the maximum variance value of the density histogram of each color calculated by the variance calculation unit 3 is selected.

Next, a slice value calculation section 5 calculates a slice value L from the density histogram of the image selected by the image selection section 4. Here, the slice value L is determined by the inter-group variance maximization method (basic of image recognition [I] P42).
47, see Ohm Co.). By using this method together with the P tile method (see Image Recognition [I] P39, Ohm Co.), a more appropriate slice value is calculated.

Next, in the background + black portion extraction section 6, the density histogram in which the slice value L is calculated is binarized using the slice value L calculated in the slice value calculation section 5 as a threshold value, and the density histogram in the designated area is obtained. The background + black portion S1 is extracted.

Next, in the density histogram calculation unit 7, the background + black portion S extracted by the background + black portion extraction unit 6 is used.
The remaining 3 not selected by the image selection unit 4 for 1
A color density histogram is calculated, and a variance calculator 8 calculates a variance value of the density histogram of each color calculated by the density histogram calculator 7.

Next, the image selecting unit 9 selects an image of a color in which the variance of the density histogram of each color calculated by the variance calculating unit 8 is the minimum, and selects the slice value calculating unit 1.
At 0, the slice value M of the density histogram of the image selected by the image selection unit 9 is calculated.

Next, in the black portion extraction section 11, the density histogram in which the slice value M is calculated is binarized using the slice value M calculated in the slice value calculation section 10 as a threshold value, and the black portion S2 in the designated area is obtained. Is extracted.

When the image selected by the image selection unit 9 is binarized using the slice value M calculated by the slice value calculation unit 10 as a threshold value, if the separation state is poor, the slice value M The background + black part image S3 in the designated area extracted by the binarization according to is extracted by the background + black part extraction unit 13, and the slice value O is calculated by the slice value calculation unit 13 for the background + black part image S3. Next, the black part extracting unit 14 binarizes the background + black part image using the slice value O as a threshold value to extract a black part in the designated area.

In this embodiment, first, the image selecting section 4
, The image of the color having the largest variance of the density histogram is selected from among the RGBY, and then the image selecting unit 9 performs the processing in the order of selecting the image of the color having the smallest variance. , Depending on the characteristics of the object,
First, the image selecting unit 4 selects an image of a color with the smallest variance value, and then the image selecting unit 9 selects an image of a color with the largest variance value. .

Next, a second embodiment of the present invention will be described in detail with reference to the drawings.

FIG. 6 is a block diagram showing the configuration of the second embodiment. Since the configuration of the second embodiment is the same as that of the first embodiment shown in FIG. Description is omitted.

The average calculator 15 calculates the average of the density histograms for each of RGBY calculated by the density histogram calculator 2.

The average calculator 16 calculates the average of the density histograms of the three colors calculated by the density histogram calculator 7.

Next, the operation of the second embodiment is almost the same as that of the first embodiment, but the difference is that the RG calculated by the density histogram calculation unit 2 is different.
Average calculation unit 1 for each density histogram of B
The average value is calculated in step 4, and the density histogram indicating the maximum value is selected in the image selecting section 4. The processing for the selected density histogram is the same as that of the first embodiment. Next, for each density histogram calculated by the density histogram calculation unit 7, the average value is calculated by the average calculation unit 15. The density histogram in which the average value indicates the minimum value is selected in the image selection unit 9. The same processing as in the first embodiment is performed for the processing for the selected density histogram, and a black pattern is finally extracted.

In the above embodiment, the processing is performed in such a manner that an image showing the maximum value is selected first for the average value of each density histogram, and then an image showing the minimum value is selected. However, according to the characteristics of the target object, an image having the minimum average value may be selected first, and then an image having the maximum value may be selected.

Next, a third embodiment will be described with reference to FIG.

Since most of the structure is the same as that of the first and second embodiments, the description of the overlapping portions will be omitted.

The setting unit 16 detects the characteristics of the target object when the color image of the target object is taken in from the image input unit 1 and outputs the characteristics to the average / variance calculation units 17 and 18 and the image selection units 4 and 9. I do.

The average / variance calculation unit 17 is provided with the setting unit 16
An average value or a variance value is calculated for each of the density histograms of RGB calculated by the density histogram calculation unit 2 based on the characteristics of the object output from.

The image selecting section 4 selects an image having the maximum average or variance of each of the RGB output from the average / variance calculating section 17 based on the characteristics of the object output from the setting section 16. Select or select the smallest image.

The average / variance calculating section 18 calculates an average value if the average value of each density histogram is calculated by the average variance calculating section 17, and calculates a variance value if a variance value is calculated. I do.

The image selecting section 9 includes the average / variance calculating section 1
8, if an image having the maximum average or variance value of each of RGB is selected in the image selection unit 4, an image having the minimum average value or variance value is selected; If the image selecting section 4 selects an image having the minimum average or variance value, the image having the maximum average or variance value is selected.

Here, the setting unit 16 has been described to change the processing settings based on the characteristics of the target when the color image is captured by the image input unit 1. If the characteristics are known, the processing may be set in advance by the external input means.

In the first, second and third embodiments, when the objects of the same package are conveyed to some extent continuously, the applied dispersion calculated by the first object or By diverting the average maximum or minimum image and each slice value to the second and subsequent images,
Further, processing efficiency is improved.

[0047]

As described above, the method for extracting a black pattern according to the present invention can be applied to a package in which a product brand or the like is printed in various monochromatic colors other than black in a cardboard carton or the like.
Even if many kinds of objects are mixed in a pattern with production date / lot / destination printed in black,
It is possible to automatically and accurately extract a black pattern and inspect and recognize printed characters in a short time.

Further, since a conversion table is unnecessary and memory access is binarized in a narrow range, it is very efficient. In particular, by packaging the background + black part image calculated first, the subsequent processing can be executed on the cache memory, so that the execution speed can be greatly improved.

[Brief description of the drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention.

FIG. 2 is a density histogram illustrating the operation of the present invention.

FIG. 3 is a density histogram illustrating the operation of the present invention.

FIG. 4 is a density histogram illustrating the operation of the present invention.

FIG. 5 is a density histogram illustrating the operation of the present invention.

FIG. 6 is a block diagram showing the configuration of a second embodiment of the present invention.

FIG. 7 is a block diagram showing the configuration of a third embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Image input part 2 Density histogram calculation part 3 Variance calculation part 4 Image selection part 5 Slice value calculation part 6 Background + black part extraction part 7 Density histogram calculation part 8 Variation calculation part 9 Image selection part 10 Slice value calculation part 11 Black part extraction part 12 Slice value calculation unit 13 Black part extraction unit 14, 15 Average calculation unit 16 Setting unit 17, 18 Average / variance calculation unit

Claims (5)

(57) [Claims] A first step of creating a density histogram in a designated area for each of the multi-color monochrome images captured by the image input means; and a second step of calculating an average of the density histograms created in the first step. Step, calculate a slice value based on the density histogram of the image in which each average calculated in the second step is the maximum,
A third step of binarizing the image using the slice values and extracting a background + black area in the designated area, excluding the first processed color from the area extracted in the third step A fourth step of creating density histograms of the other original colors, a fifth step of calculating an average of the density histograms created in the fourth step, and an image in which each average calculated in the fifth step is minimized Calculate the slice value based on the density histogram of
A sixth step of binarizing the image using the slice values and extracting a black portion in the designated region. 2. In the third step, a background + black area in a designated area is extracted based on a density histogram of an image in which each average calculated in the second step is minimum. The black step according to claim 1, wherein in the sixth step, an area of a black portion in the designated area is extracted based on a density histogram of the image in which each average calculated in the fifth step is maximum. Pattern extraction method. 3. In the second step, the variance of each density histogram created in the first step is calculated. In the third step, the variance of each image calculated in the second step is maximized. Based on the density histogram, a background + black area in the designated area is extracted. In the fifth step, the variance of each density histogram created in the fourth step is calculated. In the sixth step, 2. The black pattern extraction method according to claim 1, wherein a black portion in the designated region is extracted based on a density histogram of the image in which each variance calculated in the fifth step is minimum. 4. In the third step, based on a density histogram of an image in which each variance calculated in the second step is minimum, an area of a background + black portion in a specified area is extracted. The black step according to claim 3, wherein in the sixth step, a black area in the designated area is extracted based on the density histogram of the image having the maximum variance calculated in the fifth step. Pattern extraction method. 5. In the second step, an average or a variance of each density histogram created in the first step is calculated. In the third step, the average or the variance calculated in the second step is maximum or minimum. Based on the density histogram of the minimum image, a background + black area in the designated area is extracted. In the fifth step, the average or variance of each density histogram created in the fourth step is calculated. In the sixth step, in the black pattern extraction method for extracting an area of a black portion in the designated area based on a density histogram of an image having the maximum or minimum average or variance calculated in the fifth step, At the stage where the image of the object is captured by the input means, the printing characteristics on the object are detected, and based on the detection result, , Black pattern extracting method according to claim 1, characterized in that for setting whether the use of the maximum image or minimum image mean or variance at each step.