JP3262326B2 - Image processing apparatus and image processing method - Google Patents

Description

The present invention relates to an image processing apparatus and an image processing method.

[Conventional technology]

In recent years, as the spread of color original copying apparatuses and the improvement of image quality have progressed, there is an increasing possibility that counterfeit crimes such as bills and securities will occur. In response to this, a specific pattern is extracted from the output of the line sensor for reading the original, pattern matching is performed with predetermined template data, the presence or absence of a bill, etc. is detected, and if detected, the copying operation is interrupted. Then a new method has been devised.

[Problems to be solved by the invention]

However, in general, the position and angle of a document placed on a platen of a copying machine are completely unknown and are mixed with other documents. Therefore, it is extremely difficult to extract a desired specific pattern therefrom. Have difficulty.

Even if the pattern is successfully extracted, the target pattern differs depending on the type of banknote (amount, issuing country), and it is necessary to perform pattern matching in consideration of the possibility of each banknote. It will be something.

Furthermore, the data amount of the template for pattern matching also increases as the number of objects increases, and the required memory amount becomes enormous.

Therefore, an object of the present invention is to provide an image processing apparatus and an image processing method capable of quickly determining an input color image signal as a plurality of specific images.

[Means for Solving the Problems] In order to solve the above problems, an image processing apparatus of the present invention includes:
Input means for inputting a color image signal, storage means for storing a plurality of types of color information representing respective characteristics of a plurality of regions of a specific document for each of a plurality of types of specific documents, and color input by the input means at one time First counting means for counting the frequency of occurrence of the plurality of pieces of color information for each type of the specific document based on the image signal and the color information stored in the storage means; and A second counting unit that counts a frequency of occurrence of the plurality of pieces of color information as a whole, for each type of the specific document, based on the color information stored in the storage unit; Calculating means for calculating the degree of dispersion of occurrence of the plurality of color information based on the individual occurrence frequency of the plurality of color information counted by the first counting means; and counting by the second counting means. Depending on whether the occurrence frequency and the degree of dispersion calculated by the calculation means have exceeded respective predetermined values, the input color image signal is any one of the plurality of types of specific documents, or Determining means for determining whether or not there is not.

〔Example〕

In the following embodiments of the present invention, paying attention to the colors constituting the bills, etc., by detecting the hue in a limited window in the original image and its distribution, the presence or absence of the bills, etc. is detected, Further, when it is determined that a bill or the like is present, an image in the window is output in a mode different from a normal mode, thereby preventing a counterfeit crime.

Hereinafter, embodiments 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 of the present invention. A document 102 placed on a document table glass 101 is illuminated by a document illumination halogen lamp 103 and is imaged on a CCD line sensor 105 by a rod lens array. 105
R, G, and B color separation filters are applied in a dot-sequential manner, and output a color separation signal of a document at a pixel density of about 400 dpi (dot per inch). Reference numerals 103 to 105 scan in the direction of the arrow and sequentially read and output the original image line by line. The 106 S / H circuit samples and holds the CCD output for each pixel, and the 107 A / H circuit
The D conversion circuit converts an analog signal to a digital signal,
The shading correction circuit 08 corrects output unevenness due to sensitivity variation between CCD pixels. Output from 108
The R, G, and B signals are converted to complementary colors C (cyan), M (magenta), and Y (yellow) by a logarithmic conversion circuit 109, and a minimum value K (black) of C, M, and Y is converted by a black extraction circuit 110. Is extracted. The masking circuit 111 and the UCR circuit 112 perform a well-known color correction process, generate C, M, Y, and K plane-sequential signals, and output the signals to the selector 118.

On the other hand, the R, G, and B outputs of 108 are averaged by N × N pixels (N = about 8 to 32) by an averaging circuit of 113 and passed through a look-up table 114 which will be described later. The data of the entire manuscript is saved. At this time, at the same time as averaging, sub-sampling is performed at M pixel intervals (M = about 8 to 32 pixels) to reduce the data amount. The data stored by the CPU 115 is also processed by the CPU 116 as will be described later, and the presence or absence of a bill or the like is determined. The determination is made by dividing the document into a plurality of areas. When it is determined that there is a bill, the CPU sends an address of an image corresponding to the area to the gate signal generation circuit 117. 117 sends a gate signal to the selector 118, 118 outputs the image signals C, M, Y, and K from the area 112 where no bills or the like exist, and a dummy signal (fixed value) Signal). The output of 118 is converted again into an analog signal by the D / A converter 119, compared with the output of the triangular wave generation circuit 120 synchronized with the image clock, and pulse width modulates the semiconductor laser 123 through the laser driver 122. The output of the semiconductor laser is collimated by an appropriate optical system, and is scanned on the photosensitive drum 125 by a polygon mirror 124 rotating at a high speed to write a latent image. The latent image on the surface 125 undergoes development, transfer, and fixing steps (not shown), and a visible image is formed on paper in the order of C, M, Y, and K planes, and the copying operation is completed.

In the above operation, the process of taking in the data into the frame buffer and the determination process may be performed at the time of pre-scanning, and a gate signal as a determination result may be generated at the time of the main scan, or the Y, M, C, and BK plane sequential scanning may be performed. Alternatively, the data may be fetched the first time, the determination process may be performed during the second and third scans, and the gate signal may be generated only on the final surface (the fourth time).

Next, determination processing in the LUT 114 and the CPU 116 will be described. Now, consider a case where three types of bills A, B, and C are to be determined. First, a standard sample is prepared for A, B, and C, and the sample is sampled at 32 points as shown in FIG. 2 to obtain R, G, and B values corresponding to each point. Those corresponding to _{A R a1 ~R a32, G a1} ~G a32, B a1 ~B a32, those corresponding to B to _{R b1 ~R b32, G b1 ~G} b32, B b1 ~B b32, C The corresponding ones are R _{c1 to} R _c32 , G _{c1 to} G _c32 , and B _{c1 to} B _c32 . Next, α (= RG) and β (= GB) are obtained for each color difference signal. The distribution of α and β is, for example, as shown in FIG. 3, and there are 32 point groups for each of A, B and C. For the R, G, and B values of the original image, the LUT 114 has a code Cg indicating which of the groups A, B, and C is closest and a code indicating which of 1 to 32 is the closest among the closest groups. Cn is output (for example, Cg = 2 (B group, Cn = 6 is output in the example of FIG. 3))
The details are as shown in FIG.

The code Cg is 1 when the code is close to A, 2 when the code is close to B, and 3 when the code is close to C. 401,402 are R, G, B
Subtractor 403 for calculating α (= R−G) and β (= GB) from .alpha., 403 indicates which of the groups A, B and C the α and β are closest to in FIG. Codes Cg (outputs 1, 2, and 3 corresponding to A, B, and C) and codes Cn (1 to 32) that indicate which of 32 points of A, B, and C are closest to each other Is a LUT that outputs However, when the shortest distance is equal to or greater than a certain value (that is, neither A nor B or C), Cg = 0, Cn =
0 is output. In addition, since the brightness information is lost in the classification using only the color difference signals α and β, the LUT 404 and the window comparator 405 for taking this into account are taken into account.
Is used. A second LUT 404 stores a standard value of G corresponding to Cg and Cn. For example, when Cg = 2 and Cn = 6, a G value _Gb6 for the standard sample is output. This G _b6
And G of the document image are input to the window comparator 405 to determine whether or not the following condition is satisfied.

G _b6 −d ≦ G ≦ G _b6 + d (1) where d is a predetermined constant. The determination result is output to the selector 406, and when (1) holds, Cg,
Output Cn as it is, and if (1) does not hold, Cg, C
Resets n to 0 and outputs. With the above processing, the original image is converted into code information Cg and Cn, and the frame buffer 1
15 will be stored.

When Cg and Cn for the entire surface of the document are written in the frame buffer, the CPU 116 reads out this information and performs a process for determining the presence or absence of a bill. Next, this determination processing will be described.

First, the contents of the frame buffer are divided into a plurality of areas as shown in FIG. Since the frame buffer is obtained by subsampling the entire original image every M pixels and converting it into Cg and Cn codes, each area is obtained by dividing the original into areas of a certain size window.

Next, a histogram of Cg and Cn is created in each area. First, a histogram of Cg is created, for example, as shown in FIG. 6 (1)-(a) or (2)-(a).
Here, Cg = 0 is excluded because it is not close to any of A, B, and C. Next, when a histogram of Cn is made for each of Cg = 1, 2, and 3, FIG. 6 (1)-(b), (c), (d),
(2)-(b), (c), and (d). FIG. 6A shows how many pixels having a color close to any of A, B, and C are present in the target area.
(C) and (d) indicate which of colors 1 to 32 the group A, B, and C depend on. Therefore, when banknotes are present in the target area, the histogram of Cg is concentrated on one of 1, 2, and 3, whereas the histogram of Cn is uniformly distributed on 1 to 32 (1 to 32 is the second Since it corresponds to the color of each place of the bill as shown in the figure, if the bill exists, it should be included evenly). Sixth
In the example shown in the figure, (1) Area 1 is concentrated on Cg = 2, but the histogram of Cn when Cg = 2 is concentrated on one point, and only a lot of similar colors happen to be included. it is conceivable that. On the other hand, in (2) area 2, since Cg = 3 is concentrated and the histogram of Cn corresponding to Cg = 3 is also uniformly distributed from 1 to 32, it can be determined that there is a bill.

Therefore, the following parameters may be calculated from the Cg and Cn histograms to make the determination.

It is assumed that N _n (1) to N _n (32) are rearranged in descending order of the count number, and then the parameters are calculated.

Ng (1), Ng (2), Ng (3) are the degrees of inclusion of colors close to groups A, B, C, and σ (1), σ (2), σ (3) are groups A, B, C Represents the degree of variation from 1 to 32.

From the above, when the following condition is satisfied, it is determined that a bill exists.

Here, th ₁ and th ₂ are predetermined constants.

If a bill exists, a gate signal is generated as described above, and the image in the corresponding area is erased.

FIG. 7 shows the result of the processing performed in this manner.
(A) shows a case where bills are placed on 701 and 702 in a document image. (B) is an output copy image. 70
At 3,704,705,706, the determination is that there is a bill, and a gate signal is generated so that the image in the area is painted black (hatched portion).

[Other embodiments]

FIG. 8 is a block diagram of the second embodiment. The description of the same parts as those in FIG. 1 is omitted. Here, the output of the UCR circuit 112 is sent to one of the smoothing circuits 801 and the other to the line buffer 802, and both outputs are sent to the line buffer 118.
The selector is switched by the gate signal of 117 which is the determination result. In this case, a blurred image that has undergone the smoothing process is output in the area where it is determined that there is a bill, and the purpose of duplicating the specific image cannot be achieved.

FIG. 9 shows another configuration example of the LUT 114. Where R, G, B
LUT to input is 901,902,903 and 3 corresponding to Cg = 1,2,3
Cn are output independently, and three frame buffers 904, 905 and 906 are also provided independently. CPU116 is 904,905,906
Are sequentially referred to, the same determination processing as in the above-described embodiment is performed independently, and it is determined whether or not the determination that there is a banknote is made at any of Cg = 1, 2, and 3. Here, if the input of R, G, B is set to about 4 to 5 bits, the LUTs 901 to 903 need only have a small capacity.

According to the above embodiment of the present invention, counterfeit crime can be prevented with a relatively simple configuration. In addition, since the determination is performed only based on the color of the document, a stable determination result can be obtained irrespective of the position and angle at which the bill or the like is placed. Further, even if a plurality of determination targets are set or the targets are changed, it is possible to easily cope with them only by changing the look-up table.

In the above embodiment, R, G, and B are used as input signals, but input signals such as Y, M, and C inputs may be used. Also,
The color difference signals RG, GB are used to determine the color. For example, (Y, I, Q), (L ^* , a ^* , b ^* ), (L, U, V),
It may be converted to a signal such as (H, L, S). Further, the code information may be created based on a three-dimensional distribution including not only the color but also the brightness. If it is determined that the document is a specific document, the process is not limited to filling and smoothing, but may be a process different from a process for a general document, such as stopping a copying operation.

[Effects of the Invention] As described above, according to the present invention, an input unit for inputting a color image signal and a plurality of types of color information representing characteristics of a plurality of regions of a specific document for each of a plurality of types of specific documents are provided. The individual generation of the plurality of pieces of color information for each type of the specific document based on the stored storage means, and the color image signal input at one time by the input means and the color information stored in the storage means. A first counting unit that counts the frequency, and a frequency of occurrence of the plurality of pieces of color information as a whole for each type of the specific document based on the color image signal and the color information stored in the storage unit. A second counting means for counting, and a degree of dispersion of occurrence of the plurality of color information based on an individual occurrence frequency of the plurality of color information counted by the first counting means for each type of the specific document Calculate The input color image signal is calculated based on whether the occurrence frequency counted by the second counting means and the degree of dispersion calculated by the calculating means each exceed a predetermined value. A determination unit that determines whether or not any of the plurality of types of specific originals, a determination using a color image signal having a large data amount and a determination of the plurality of types of specific originals. Despite the use of two factors that require a long determination time, it is possible to accurately determine a plurality of types of specific documents by inputting a single color image signal.

[Brief description of the drawings]

FIG. 1 is a diagram showing an example of the configuration of the present invention, FIG. 2 is a diagram for explaining sampling of a target color, FIG. 3 is a diagram showing an example of a distribution of the target color, and FIG. 4 is a color determination lookup table. 5, FIG. 5 is a diagram for explaining a determination area, FIG. 6 is a diagram showing an example of a histogram, FIG. 7 is a diagram showing an example of a processing result, and FIG. 8 is a block diagram showing another embodiment. FIG. 9 is a diagram showing another embodiment of the color judgment look-up table. 113 Averaging circuit 114 LUT 115 Frame buffer 116 CPU

Continuation of front page (56) References JP-A-1-302962 (JP, A) JP-A-2-150982 (JP, A) JP-A-1-300285 (JP, A) JP-A-1-316783 (JP) JP-A-61-80961 (JP, A) JP-A-56-27493 (JP, A)

Claims (3)

(57) [Claims] An input unit for inputting a color image signal; a storage unit for storing a plurality of types of specific documents, each storing a plurality of pieces of color information representing respective characteristics of a plurality of regions of the specific document; First counting means for counting the frequency of occurrence of the plurality of pieces of color information individually for each type of the specific document based on the color image signal input at a time and the color information stored in the storage means, A second counting unit that counts the frequency of occurrence of the plurality of pieces of color information as a whole for each type of the specific document based on the color image signal and the color information stored in the storage unit; Calculating means for calculating the degree of dispersion of occurrence of the plurality of color information based on the individual occurrence frequency of the plurality of color information counted by the first counting means for each type of document; Counting hands The input color image signal is one of the plurality of types of specific originals, depending on whether or not the occurrence frequency counted by the above and the degree of dispersion calculated by the calculation unit each exceed a predetermined value. An image processing apparatus comprising: a determination unit configured to determine whether or not any of the above is true. 2. The image processing apparatus according to claim 1, wherein said storage means has a plurality of look-up tables respectively corresponding to specific documents. 3. A step of inputting a color image signal; and a step of inputting the color image signal at one time in the input step and a plurality of types of color information representing characteristics of a plurality of regions of the specific document with respect to each of a plurality of types of specific documents. For each type of the specific document,
A first counting step of counting the frequency of occurrence of each of the plurality of color information; and, for each type of the specific document, the plurality of the plurality of color information based on the color image signal and the color information stored in the storage unit. A second counting step of counting the frequency of occurrence of the color information as a whole, and for each type of the specific document, based on the individual frequency of occurrence of the plurality of pieces of color information counted in the first counting step. A calculating step of calculating the degree of dispersion of the occurrence of a plurality of color information; and determining whether the occurrence frequency counted in the second counting step and the degree of dispersion calculated in the calculating step each exceed a predetermined value. A determination step of determining whether the input color image signal is any one of the plurality of types of specific originals or none of them.