JP2757868B2 - Image information binarization processing circuit - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Summary] For input multi-valued image information in which gray-scale images and black-and-white two-level images are mixed, image information for judging respective image areas and performing optimal binarization processing The object of the present invention is to realize means for accurately distinguishing an input image as a gray-scale image or a two-level image with a simple configuration by using a predetermined dither threshold for each pixel. A pseudo-halftone processing circuit for performing dither processing using a value pattern and converting it to pseudo-halftone binary image information; and inputting simple binary image information and pseudo-halftone binary image information, and indicating one of them And a selector circuit for selecting each pixel of the input multi-valued image information as to whether the pixel belongs to a two-level image area or a gray-scale image area. Select value image information And an image area determination circuit for instructing to perform the operation.

[Industrial applications]

The present invention relates to a binarization processing circuit for image information that determines the respective image areas and performs an optimal binarization process on input multivalued image information in which a grayscale image and a monochrome two-level image are mixed. .

Identify images that require light and shade representation, such as photographs, and images that emphasize resolution, such as characters and figures.
Both provide a technique for performing optimal binarization processing so as to obtain high-quality binarized image information.

[Conventional technology]

In a general image processing system, when an image is input by an image scanner or the like, the original image is scanned by a CCD, converted into multi-value (multi-gradation) image information, and further subjected to a simple slicing process using a threshold. And binarized image information.

However, the original image may contain various types of images, such as photographic images expressed in gray levels (halftones), and images of characters and line drawings expressed in two levels such as black and white. When the binarization process using a simple slice is performed using the threshold value of
Original quality may be significantly reduced.

Therefore, the image expressed in gray level (hereinafter referred to as gray image) and the image expressed in two levels such as black and white (hereinafter referred to as two level image) are distinguished and binarized by applying the optimum threshold value to each. Processing methods have been developed.

In this case, when an original image is read, it is necessary to determine for each pixel whether the original image is in a grayscale image area or a two-level image area. Conventionally, from a plurality of pixels around an arbitrary pixel of interest on an image, a. A method for examining the density histogram, b. A method for calculating the difference between the maximum density and the minimum density, c. A method for calculating the average value and variance of the density D. Calculate the run length, and determine whether the pixel of interest belongs to an image area of two levels or shades based on the value of the run length.

In this last method d, x is set as the reading position and D
The binarization processing is performed using the conversion function g (D (x)) which is the density level of (x). This conversion function g (D (x))
Is a function such that g (D (x)) = 1: D (x) ≧ α g (D (x)) = 0: D (x) <α by the threshold value α.

Based on the image information binarized using this conversion function,
The run length around the pixel of interest is calculated.
When the calculated run length is larger than a certain threshold value β, it is determined that the target pixel belongs to the grayscale image area.

In this case, 2 such as a document mainly composed of characters
Level images generally have very short run lengths,
In the photographic image, the long run length is used as the basis for the judgment.

 Next, a specific example will be described.

FIG. 6 is a graph showing the relationship between the
The arrows indicate the scanning direction of the CCD. Each arrow of,... Indicates the main scanning (line scanning) electrically performed in the CCD, and the vertical scanning indicated by the sequential main scanning,. The scanning in the direction indicates a sub-scan performed by the mechanical movement of the CCD in the downward direction or the mechanical movement of the original in the upward direction.

When the run length is counted, in the main scanning direction, it can be easily obtained from the pixel information continuously obtained in the main scanning of the CCD.
Since it is not possible to obtain the run length continuously, prepare a line memory or input buffer corresponding to the maximum value for which the run length is to be obtained, and temporarily store the information of each pixel of a plurality of continuous lines in those lines. Is obtained by reading out the same pixel in the vertical direction.

FIG. 7 shows an example of image information, and shows a density signal waveform when a document image in which characters, line drawings, photographs, and the like are mixed is read at 16 gradations in the main scanning direction (x). In the illustrated example, the center portion is a two-level image area such as a character or a line drawing, and the left and right portions are a grayscale image area such as a photograph.

FIG. 8 shows binary image information resulting from simple slice binarization of the image information of 16 gradations shown in FIG. 7 with a threshold value of density level 4.

In this case, the above-described conversion function g (D (x)) is given as follows (α = 4).

g (D (x)) = 1: D (x) ≧ 4 g (D (x)) = g (D (x)) = 0: D (x) <4 where g (D (x)) = 1 as a black pixel and g (D
(X)) = 0 pixels are defined as white pixels. Therefore, the eighth
In the figure, or represents a continuation of black pixels, that is, a black run length. The number of pixels for each black run length or pixel is 14,4,3,20.

Therefore, in the case of the example of FIG. 8, by setting “5” as the threshold value β for separating the gray image area and the two-level image area, the black run length in FIG. It can be determined that the black run length belongs to the two-level image area.

FIG. 9 shows an example of a conventional binarization processing circuit having a function of distinguishing between a grayscale image area and a two-level image area by such a method, where 1 is a binary converter, and 2 is a binary converter. Demultiplexer, 3
-1 to 3-n are n input buffers (n is the maximum run length to be obtained), 4 is a multiplexer, 5 is an n-stage shift register, 6 is a reference pattern generator, 7 is a comparator, 8 is a two-stage shift register, and 9 is a logical product circuit.

The binary converter 1 operates according to the scanning method shown in FIG.
The multivalued image information output from the CCD as shown in FIG. 7 is converted into binary image information as shown in FIG. 8 using the threshold value α.

The binary image information output from the binary converter 1 is input to the input buffer 3-1 for each line by the demultiplexer 2.
To 3-n cyclically. That is, the input buffers 3-1 to 3-n each have a pixel capacity for one line, and the demultiplexer 2 selects the next empty input buffer when the writing of the line data to one input buffer is completed. Then, image information for six lines is continuously written.

On the other hand, the multiplexer 4 selects the input buffer in which the writing has been completed by a certain method, and transfers the content to the shift register 5.

When determining the run length in the main scanning direction, the multiplexer 4 determines whether the target pixel exists in the input buffer 3-1.
To 3-n, and the information of n pixels in the horizontal direction is successively read out to the shift register 5. When the run length in the sub-scanning direction is to be obtained, the target pixel position ( For each x), the multiplexer 4 scans the information of n pixels corresponding to the vertical direction of each of the input buffers 3-1 to 3-n, and performs continuous reading to the shift register 5.

The reference pattern generator 6 is composed of a ROM and generates reference patterns for various runs for distinguishing between a grayscale image and a two-level image. That is, a run reference pattern is generated for all possible combinations of n pixel values.

Each time the image information is written into the shift register 5, the comparator 7 compares the content of the image information with each reference pattern, and compares the result of comparison with the run-length reference pattern larger than the threshold β with the input of the shift register 8. Write to the column. Here, the comparison result when the run length of the image information is larger than the reference pattern is “1”.

This run length detection operation is continuously performed for each pixel of interest in the main scanning direction and the sub-scanning direction.

The shift register 8 has a two-stage configuration, performs a shift operation when a comparison result is input to an input stage, and holds a comparison result in the main scanning direction and a comparison result in the sub-scanning direction in each stage.

The AND circuit 9 outputs a judgment result "1" indicating that the target pixel belongs to the grayscale image area when the contents of each stage of the shift register 8 are both "1", and in other cases, A determination result “0” indicating that the pixel belongs to the two-level image area is output.

[Problems to be solved by the invention]

Conventionally, when binarizing halftone image information in which a grayscale image and a two-level image are mixed in real time, various methods have been used to determine to which image a pixel to be processed belongs. However, all of these methods have drawbacks such as an inability to obtain an accurate determination result, an inadequate real-time processing due to a complicated operation, and a large-capacity memory.

In particular, in the circuit using the run length, a large number of input buffers (or line memories) are required to determine the run length in the sub-scanning direction.

Originally, even if it is a two-level image of a character and a line drawing, the line width of which is larger than a certain level has a large run length, so that there is a disadvantage that the image is erroneously determined as a grayscale image.

SUMMARY OF THE INVENTION It is an object of the present invention to realize means for accurately determining whether an input image is a grayscale image or a two-level image with a simple configuration.

[Means for solving the problem]

According to the present invention, in a binarization processing circuit, a grayscale image area and a two-level image area in an input image are input to an image area determination circuit in the binarization processing circuit in order to accurately determine each pixel. The run-length is obtained by slicing the multi-valued image information with an appropriate threshold value, and those with a small run-length value are assumed to correspond to the line width of characters, line drawings, etc. Judgment is made, and instead of finding the run length in the sub-scanning direction, edge detection is performed in the main scanning direction. If there is an edge portion where the density level changes rapidly, it is regarded as a part of a character / line image. Even if the run length value is large, a function is provided to determine that the pixels at the edge part also belong to the two-level image area.
Pixels determined to belong to the level image are subjected to a binarization process using a slice having a single threshold value.

FIG. 1 is a diagram showing the basic configuration of the present invention. Reference numeral 10 denotes a delay circuit which converts input multi-valued image information into
It has a timing adjustment function that delays by the time required for image area determination.

Numeral 11 denotes a simple slice processing circuit, which is a two-level image.
It has a threshold value that determines a slice level suitable for binarization, and converts input multi-valued image information delayed to simple binary image information.

Reference numeral 12 denotes a pseudo halftone processing circuit which performs dither processing on delayed input halftone image information using a dither threshold pattern to convert it into pseudo halftone binary image information.

Reference numeral 13 denotes a selector circuit, which selects one of simple binary image information and pseudo halftone binary image information on a pixel basis.

Reference numeral 14 denotes an image area determination circuit which determines whether the input multi-valued image information belongs to a two-level image area or a grayscale image area on a pixel-by-pixel basis. Simple binary image information or pseudo halftone 2
Select value image information.

Reference numeral 15 denotes a simple slice processing circuit similar to the simple slice processing circuit 11. The threshold value for determining the slice level is, for example, a value slightly smaller than the threshold value for binarizing a two-level image. It is not limited to.

Reference numeral 16 denotes a small run-length detection circuit, which slices the input multi-valued image information at an appropriate level.
The run length of the black run is determined, and if the target pixel is included in a run of a run length equal to or less than the predetermined threshold value r, the pixel is regarded as a line segment or a part of a point, and the pixel is regarded as a two-level image area. Is determined to belong to

Reference numeral 17 denotes an edge detection circuit for examining a change in density level around a pixel of interest in delayed input multi-valued image information.
If there is a certain level of density level change or more, it is regarded as an edge portion, and it is determined that the target pixel belongs to the two-level image area.

(Operation)

FIG. 2 is a signal waveform diagram for explaining the operation of the binarization processing circuit of the present invention shown in FIG. 1, and FIG.
FIG. 1B shows original multi-valued image information represented by density level waveforms obtained by reading original pixels in the main scanning direction, and FIG. 1B shows a binary image output from the simple slice processing circuit 15 in FIG. The information is represented by a binary waveform.

In FIG. 2A, waveforms I and II are character image portions of two levels of black and white, and waveform III is a portion of a grayscale image. Further, the threshold value is a threshold value used for a simple slice in the simple slice processing circuit 11 of FIG. 1, and the threshold value is a threshold value used for a simple slice in the simple slice processing circuit 15.

The simple slicing processing circuit 11 in FIG. 1 uses the threshold value to simply slice the waveforms I, II, and III in FIG. 2A, respectively, and converts them into simple binary image information. Similarly, the pseudo halftone processing circuit 12 dithers the waveforms I, II, and III using an appropriate dither threshold pattern, and converts the waveforms into pseudo halftone image information. Each of the converted image information is input to the selector circuit 13, respectively.

On the other hand, in the image area determination circuit 14 shown in FIG. 1, the simple slice processing circuit 15 simply slices the waveforms I, II, and III shown in FIG. The detection circuit 16 detects a run having a run length value smaller than the threshold value r. In the case of FIG.
The waveform I corresponds to this, and the image area determination circuit 14 instructs the selector circuit 13 to select simple binary image information as that corresponding to the waveform I.

Further, the edge detection circuit 17 detects an edge portion where the density level sharply changes for each of the waveforms I, II, and III. For the edge detection circuit 17, a digital filter using an operation such as Laplacian, gradient, or Sobel can be used. In the example of FIG. 2A, the leading edge and the trailing edge of the waveform II are edge-detected, and the image area determination circuit 14
It instructs the selector circuit 13 to select simple binary image information.

The image area determination circuit 14 discriminates the image other than the detected small run lengths and edges from the grayscale image, and selects the selector circuit 13.
To select pseudo halftone binary image information.

In this way, the selector circuit 13 outputs the simple binary image information and the pseudo halftone 2 as shown in FIG.
Binary image information is output in a form in which value image information is mixed.

Here, in the waveform II, the leading edge and the trailing edge are each represented as a black level, but the central portion is represented as a shading level according to the original. It can be reproduced as an easy-to-view image.

〔Example〕

FIG. 3 is a configuration diagram of a binarization processing circuit according to one embodiment of the present invention.

In FIG. 3, reference numeral 10 denotes a delay circuit composed of an r-stage shift register, 11 denotes a simple slice processing circuit based on a threshold, 12 denotes a pseudo halftone processing circuit, 13 denotes a selector circuit, 14 denotes an image area determination circuit, 15 and 19 are simple slice processing circuits based on thresholds, 16a is a run-length counter, 16b is a run-length determination circuit, 16c is a JK flip-flop, 16d is an AND circuit, 16e is an inverter, 17a is a line buffer, and 17b is an edge. The detection filter 18 is an OR circuit.

Here, in FIG. 1 and FIG. 3, the elements referred to by the same numbers respectively have corresponding functions. However, 16 corresponds to 16a to 16e, and 17 corresponds to 17a and 17b. The thresholds, r are common to both figures.

FIG. 4 is a signal waveform diagram for explaining the operation of the embodiment of FIG. 3; CLK: input pixel clock A: signal obtained by binarizing an input pixel with a threshold value B: signal of A Among them, a signal for detecting a run where the run length is larger than the threshold value r C: a signal obtained by delaying the signal A by r pixel clocks * D: a run having a run length of r or more from the signals B and C −E: A signal that takes the logical product of signal C and signal D, and that detects a run whose run length is smaller than r F: a signal that detects an edge portion from a signal delayed by r pixel clocks G: a signal This is an image area determination signal based on the logical sum of E and the signal F.

 The operation will be described with reference to FIG. 3 and FIG.

The input multi-level image signal is applied to the simple slice processing circuit 15 and the delay circuit 10.

The signal A binarized by the simple slice processing circuit 15 is counted by a run length counter 16a for the run length of black runs, and the result is input to a run length determination circuit 16b.
A signal B that is compared with the threshold value r and becomes “H” when the run length is larger than r is output to J to the J terminal of the flip-flop 16c. In the illustrated example, the threshold value r has a length corresponding to five pixel clocks.

On the other hand, the input multi-valued image signal delayed by r pixel clocks by the delay circuit 10 is binarized by the simple slice processing circuit 19 to become a signal C, and further inverted by an inverter 16e to be synchronized with the signal B. JK flip-flop
It is added to the K terminal of 16c.

The JK flip-flop 16c turns on when the signal levels of the J and K terminals become “H” and “L”, respectively, and makes the Q terminal “L”. “L”,
Turns off when it goes to "H", and sets the Q terminal to "H" (signal * D).

The signal * D is obtained by extracting only a black run having a run length larger than the threshold value r from the signal C.

The AND circuit 16d removes a black run portion having a run length larger than the threshold value r from the signal C by taking the logical product of the signal C and the signal * D (signal F).

The functions of the simple slice processing circuit 11, the pseudo halftone processing circuit 12, and the selector circuit 13 are the same as those described with reference to FIG.

The line buffer 17a stores the delayed input halftone image signals for the number of lines required for edge detection.

The edge detection filter 17b extracts data of a plurality of pixels determined from the line buffer 17a for each pixel of interest, and detects an edge portion by a filtering process (signal F). The threshold value T is used as a reference value for detecting an edge portion.

The signal E and the signal F are combined by an OR circuit 18 to generate a signal G.
Becomes The signal G indicates only a run or a part (edge) of the run belonging to the two-level image area in the signal C. When the signal G is at “H” level, the signal G of the simple slice processing circuit 11 The output is selected, and the output of the pseudo halftone processing circuit 12 is selected when the output is at the "L" level.

Next, the filter processing of the edge detection filter 17b will be described.

The edge detection filter 17b is a kind of second-order or first-order differential filter. In the case of a 3 × 3 pixel matrix as shown in FIG. 5A, for example, in the Laplacian method, | 2S0− ΣSi / 4 | ≧ T1 (1) (i = 0, 1, 2,..., 8) is calculated, and in the gradient method, | S8−S4 | + | S2-S6 | ≧ T2 (2) T1 and T2 represent threshold values.

In the Laplacian method, a digital filter that passes a second-order differential component of an image is formed, and in the gradient method, a digital filter that passes a first-order differential component is formed.

FIG. 5B is a configuration diagram of an embodiment of an edge detection filter based on the Laplacian method. In the figure, the line buffer A stores the line data of the previous line, and the line buffer B stores the line data of the previous line.

The 演算 Si operation unit inputs the current line data and the line data in each line buffer A, B, respectively, calculates ΣSi, and calculates the difference operation with So in the | 2S0−ΣSi / 4 | operation unit. Do. The result of this operation is compared with a threshold value T1 in a comparator (not shown), and the result of equation (1) is obtained.

Since the Laplacian-based edge detection filter extracts high spatial frequency components, in the case of a document image such as a halftone photograph, many edges are detected in the photograph.
Image quality may be degraded.

In order to solve this problem, it is effective to use an edge detection filter based on sobel operation.

The edge detection filter by the Sobel operation calculates | S1 + 2S8 + S7− (S3 + 2S4 + S5) | + | S1 + 2S2 + S3− (S7 + 2S6 + S5) | ≧ T3 (3). This indicates that a difference operation is performed. Note that T3 is a threshold value. By including this smoothing process, the effect of noise at high frequencies can be reduced.

〔The invention's effect〕

As described above, according to the present invention, input multi-valued image information is accurately separated into a grayscale image area and a two-level image area, and optimal simple slice binarization and pseudo halftone binarization are performed. In particular, characters and line drawings having a large line width are reproduced with sharp edges, and grayscale images such as original photographs are reproduced with image quality not inferior to the conventional method.

Further, the present invention has an advantage that the amount of hardware such as an input buffer is smaller than that of the conventional one, and that it can be easily operated in real time.

[Brief description of the drawings]

FIG. 1 is a diagram showing the principle of the present invention, FIG. 2 is a signal waveform diagram for explaining the operation of the present invention, FIG. 3 is a diagram showing the configuration of a binarization processing circuit according to an embodiment of the present invention, and FIG. 3 is a signal waveform diagram of the circuit of the embodiment of FIG. 3, FIG. 5 is an explanatory diagram of an embodiment of the edge detection filter, FIG. 6 is an explanatory diagram of document scanning by CCD, and FIG. 8 is a density signal waveform diagram of an example of image information to be processed, FIG. 8 is a signal waveform diagram of image information obtained by simply binarizing the density signal waveform of FIG. 7, and FIG. 9 is a configuration of a conventional example of a binarization processing circuit. FIG. In FIG. 1, 10: delay circuit 11, 15: simple slice processing circuit 12: pseudo halftone processing circuit 13: selector circuit 14: image area determination circuit 16: small run-length detection circuit 17: edge detection circuit

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-57-6964 (JP, A) JP-A-58-220563 (JP, A) JP-A-61-4369 (JP, A) JP-A-61-369 43078 (JP, A) (58) Field surveyed (Int. Cl. ⁶ , DB name) H04N 1/40 G06F 15/20 G06F 15/68

Claims (5)

(57) [Claims] 1. A binary system which inputs multilevel image information obtained by reading a two-level image such as a photograph or a two-level image such as a character or a line drawing in multiple gradations and converts it into binary image information in real time. A simple slicing circuit (11) for slicing the input multi-valued image information in pixel units at a predetermined threshold and converting the input multi-valued image information into simple binary image information; A pseudo-halftone processing circuit (12) for performing dither processing using a predetermined dither threshold pattern and converting it into pseudo-halftone binary image information; and inputting simple binary image information and pseudo-halftone binary image information A selector circuit (13) for selecting one of them in units of pixels according to an instruction; and determining whether each pixel of the input multi-valued image information belongs to a two-level image area or a gray-scale image area. And the selector circuit (13) An image area determination circuit (14) for instructing selection of pure binary image information is provided. The image area determination circuit (14) slices the input multi-value image information at an appropriate threshold. A run length is obtained, and when the run length is included in a run smaller than a predetermined value, the corresponding pixel is determined to belong to the two-level image area.
Determination means, having a multi-valued image edge detection unit for detecting an edge portion of the input multi-valued image information, to obtain a change amount from a predetermined number of pixel data in the vicinity of the pixel of interest by a predetermined product-sum operation processing, When the change amount is larger than a predetermined slice value, a second determination unit configured to determine that the pixel belongs to the two-level image area is provided, and the first determination unit and the second determination unit If either is determined to be a binary level image area, the corresponding pixel is determined to be a binary level image, and if neither determination unit is determined to be a binary level image area, the corresponding pixel is determined to be a pseudo halftone image area. Circuit for binarizing image information. 2. The method according to claim 1, wherein the input multi-valued image information is delayed by a time required for processing by the image area determination circuit (14) with respect to the simple slice processing circuit (11) and the pseudo intermediate processing circuit (12). A binarization processing circuit for image information, wherein the binarization processing circuit is provided with a delay by a circuit (10). 3. An image area determination circuit according to claim 1, wherein
4) A binarization processing circuit for image information, which detects an edge portion by Laplacian operation. 4. An image area determination circuit according to claim 1, wherein
4) A binarization processing circuit for image information, wherein an edge portion is detected by a gradient operation. 5. An image area determination circuit according to claim 1, wherein
4) A binarization processing circuit for image information, which detects an edge portion by a Sobel operation.