JPH0831953B2 - Image area identification device - Google Patents

Description

The present invention relates to an image area identification device for inputting image data for each block and recognizing an image area of a pixel of interest in the block.

[Prior Art] Conventionally, as shown in FIG.
The data read by the input sensor 21 such as CCD is A / D converted at 22, and then gradation correction such as γ conversion is performed at 23, and the binarization circuit 24
The method of converting into a dot ON / OFF signal and outputting with the binary printer 25 was adopted.

Representative binarization methods in the binarization circuit 24 include a binarization method with a fixed threshold and a binarization method with a dither method.

[Problems to be Solved by the Invention] With the fixed threshold binarization method, the quality of the reproduced image of the characters and line drawings in the original is good, but the gradation is lost in the photograph and the image quality is There was a problem of lowering. On the other hand, with the dither method, although the gradation of the photograph is good, the characters and line drawing parts are cut off, and moire fringes occur on the halftone dot photograph of a specific screen ruling, which causes deterioration of image quality. There was a problem. Therefore, when characters, line drawings, photographs and halftone pictures coexist in the same document, it is difficult to form all the images satisfactorily.

Therefore, the present invention eliminates the above-mentioned drawbacks of the prior art, and
The present invention proposes an image area discriminating device which faithfully discriminates the image area of image data of a document in which line drawings, photographs and halftone pictures are mixed.

[Means for Solving Problems] An image area identification device of the present invention proposed to achieve the above-mentioned object inputs image data for each block and recognizes an image area of a pixel of interest in this block. A first detecting means for detecting a difference between an average density value of the block and a density value of the pixel of interest; and a second detecting means for detecting a low-frequency edge component in the block.
Detecting means for detecting a high frequency edge component in the block,
And the density difference detected by the first detection means is smaller than a predetermined value, the density difference detected by the first detection means is larger than a predetermined value if the image is a halftone image. If the low-frequency edge component detected by the second detecting means and the high-frequency edge component detected by the third detecting means are both larger than a predetermined value, it is determined that the image is a character image. The density difference detected by the detecting means is larger than a predetermined value, and the low frequency edge component detected by the second detecting means is smaller than a predetermined value. The high frequency edge detected by the third detecting means. If the component is larger than a predetermined value, it is a halftone image,
And an image area identifying means for identifying.

Embodiments Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

FIG. 1 is a block diagram showing an embodiment in which the image area recognition apparatus of the present invention is applied to a printing apparatus which reads an image, processes the image, and then prints the image. According to this printing device,
A document (not shown) is read by the input sensor 1 such as CCD,
The / D converter 2 converts it into a digital signal of, for example, 8 bits. This digital signal represents the gradation of 256 gradations. The gradation correction circuit 3 performs various gradation operations on this signal.

Next, this signal enters three types of binarization circuits, that is, a fixed threshold binarization circuit 4, a dither binarization circuit 5, and a moire suppression binarization circuit 6. The fixed threshold binarization circuit 4 compares the image data with a certain threshold value, and outputs "1" when the image data is larger, and "0" when it is smaller To do. The dither binarization circuit 5 sequentially compares the threshold values indicated by the dither matrix with the image data, and similarly, "1"
Or it outputs a "0" signal. This dither binarization is effective for binarizing a halftone (continuous tone) image. The moire suppression binarization circuit 6 binarizes the original in a method that does not generate moire fringes even when the original is a halftone photograph. As such a moire suppression method, for example, smoothing processing is applied to image data to remove a periodic component in the image data to suppress moire, or a halftone dot number of a document and a dither that hardly causes beats. There is a method of selecting a matrix size, or a method of randomly selecting a dithermatrix component to suppress a beat.

The three binarization circuit outputs are appropriately switched in the switch 11 according to the judgment result of the judgment circuit 10 and are sent to the printer 12 as printer dot ON / OFF signals to form an image.

The operation of the determination circuit 10 and the like will be described. The same image data is input to the block circuit 13 in synchronism with the image data whose gradation has been corrected by the gradation correction circuit 3 being input to the three binarization circuits. The block converting circuit 13 cuts out a pixel block of a predetermined size (for example, a size of 5 × 5 as shown in FIG. 3) from the gradation-corrected image data, and extracts the pixel block from the density detecting circuit 7 and the first block. It is sent to the space filter circuit 8 and the second space filter circuit 9. The density difference detection circuit 7 takes the absolute value of the difference between the average value in the specific pixel block and the target pixel. Generally, when the absolute value of the difference is smaller than a certain threshold value, it can be said that the pixel of interest is in the halftone image area such as a photographic image or the area of the background image. On the contrary, when this absolute value is large, it can be said that it is an image area such as a character portion or a halftone image in which the local density change is large. Therefore, the density difference detection circuit 7 outputs, for example, "0" when it is determined that the target pixel is in the halftone image area, and outputs "1" when it is determined that it is in the character portion and the halftone dot image area. To do so.

Next, the first spatial filter circuit 8 constitutes a spatial filter that performs edge detection at a relatively low frequency (about 2 line pairs lp / mm). This can be realized, for example, by combining first-order differential filters having coefficients as shown in FIGS. 3 (a) and 3 (b).

The second spatial filter circuit constitutes a spatial filter for detecting the edge of the halftone dot frequency. This can be realized by, for example, a quadratic differential filter having a coefficient as shown in FIG. In general, due to the characteristics of the filter, if the kernel size is the same, the first-order derivative filter has a peak of the transmission band characteristic on the lower frequency side than the second-order derivative filter. FIG. 4 (a) shows the characteristics of the two types of spatial filters as a one-dimensional model. The first derivative filter has a characteristic having a peak at about 2 lp / mm, and the second derivative filter has a characteristic having a peak at about 4 lp / mm.

On the other hand, FIG. 4B shows the spatial frequency characteristics of the character portion and the halftone dot photograph portion in the original image. In terms of frequency, the power spectrum of a character decreases from the low-frequency side to the high-frequency side almost continuously, whereas in the halftone dot photograph, the high-frequency component corresponding to the halftone dot frequency is discontinuously periodic. There is a difference in that there is a peak in the power spectrum.

Therefore, in the first spatial filter circuit 8, the transmission band is
Since it is located at 2 lp / mm, the spatial frequency component of the character is detected, but the spatial frequency component of the halftone picture is not detected. On the other hand, the second spatial filter circuit 9 has a transmission band of 4 lp / mm and has a wide half-value width of the peak showing the filter characteristic. Therefore, in this filter, the frequency components of characters are centered around the spatial frequency of the halftone dot photograph. Detect up to.

FIG. 5 is a detailed block diagram of the circuit 10 of FIG. 1 for determining the image area in the image. In the block diagram of FIG. 5, a line memory 50 is used as the block circuit 13. If a 5 × 5 matrix as shown in FIG. 3 is used, the line memory 50 should be prepared for 5 lines. The density difference detection circuit 7 includes an average value calculation circuit 51 and a subtractor 52.
Composed of and. The average value calculation circuit 51 is a line memory 50.
The pixel data for one block from is input, addition and division are performed, and the average value in the block is calculated. Subtractor 52
The density of the pixel of interest (the central pixel in the case of a 5 × 5 matrix) of the line memory 50 is input to one of the inputs. Then, the absolute value of the difference between the average value and the density of the pixel of interest is subtracted.
Taken at 52.

The absolute value of this difference is compared with a specific threshold value T1 in the comparator 57. When the difference between the image data of the pixel of interest and the average value is small, it is determined as a halftone photographic area or a background area of a character image, and "0" is output. On the contrary, when the difference is large, it is determined to be a character area or a halftone dot picture area and "1" is output. The output result A is input to the determination circuit 10.

Further, the image data of the pixel in the block from the line memory 50 is input to the first space filter circuit 8, that is, to the primary differentiating circuits 53 and 54 of FIG. 5, and the image data of FIG.
The primary differential operation as shown in (b) is performed by each differential circuit, and this absolute value is output. The differentiating circuit is a multiplier that multiplies the pixel density and the matrix coefficient corresponding to the pixel,
And an adder that takes the sum of these products. As a result, the circuit 53 detects a vertical low-frequency image edge, and the circuit 54 detects a horizontal low-frequency edge. These outputs are added by the adder 56 and then added by the comparator 58.
Is compared with a specific threshold T2. When the addition result is larger than the threshold value T2, it is determined that there is a low-frequency edge, and "1" is output. When the addition result is small, "0" is output as no edge. The output result B is input to the judging device 10 as in the case of A.

In the second spatial filter circuit 9, the secondary differential circuit 55 inputs the pixel data in the block, performs the secondary differential operation as shown in FIG. 3 (c), and outputs the absolute value. This value is compared in the comparator 59 with a specific threshold T3. When the secondary differential value is larger than the threshold value T3, it is determined that there is a high-frequency edge including a halftone dot, and "1" is output. On the contrary, when it is small, "0" is output as no edge. The result C is input to the determiner 10.

Here, X may be "0" or "1".

As shown in Table 1, the decision unit 10 outputs the above three output results.
A, B, and C determine the photo area and character area. Output A
When is "1" and both B and C outputs are "1", it is judged as a character area. When B is "0" and C is "1", it is determined as a halftone dot area. Other combinations of A, B, and C outputs are possible, but they are ignored because they are very unlikely to occur.

The judging device 10 can be constructed by a simple circuit as shown in FIG. The output A that has passed through the inverter 70 is directly used as the photographic area designating signal, and the outputs of the AND gates 71 and 72 are used as the character area designating signal and the dot area designating signal, respectively. The switch 11 of FIG. 1 is switched by these instruction signals. That is, since the binary image data output from the switch 11 is the image data processed by the binarization most suitable for the image area, the image reproduced by the printer 12 is of high quality.

[Effects of the Invention] As described above, according to the image area recognition device of the present invention, the image data of the original image in which the photograph area, the character area, and the halftone dot photograph area are mixed can be identified according to those areas. it can.

[Brief description of drawings]

FIG. 1 is a block diagram showing an outline of an embodiment in which an image area recognition device of the present invention is applied to a printing device, FIG. 2 is an explanatory view of a conventional binary image forming device, and FIG. 3 (a) is A matrix diagram showing the coefficient values of the primary differential filter for detecting the vertical edge, FIG. 3 (b) is a matrix diagram showing the coefficient values of the primary differential filter for detecting the horizontal edge, and FIG. 3 (c) is A matrix diagram showing coefficient values of a secondary differential filter for detecting a high frequency edge, FIG. 4 (a) is a characteristic diagram showing spatial frequency characteristics of the primary differential filter and the secondary differential filter, and FIG. 4 (b). Is a characteristic diagram showing a spatial frequency characteristic of a typical character region / halftone dot region, FIG. 5 is a detailed block diagram of a circuit for image area determination, and FIG. 6 is a diagram showing a configuration example of a determination circuit. . In the figure, 4 ... Fixed threshold binarization circuit, 5 ... Dither binarization circuit,
6 ... Moire suppression binarization circuit, 7 ... Density difference detection circuit,
8 ... First space filter circuit, 9 ... Second space filter circuit, 10 ... Judgment circuit, 11 ... Switch, 12 ... Printer.

Claims (1)

[Claims] 1. An image area identifying apparatus for inputting image data for each block and recognizing an image area of a pixel of interest in the block, detecting a difference between an average density value of the block and a density value of the pixel of interest. First detecting means, a second detecting means for detecting a low-frequency edge component in the block, a third detecting means for detecting a high-frequency edge component in the block, and the first detecting means. When the density difference detected by the means is smaller than a predetermined value, it is a halftone image, and the density difference detected by the first detection means is larger than the predetermined value and detected by the second detection means. If both the low-frequency edge component and the high-frequency edge component detected by the third detecting unit are larger than a predetermined value, it is a character image, and the dark image detected by the first detecting unit is detected. When the degree difference is larger than a predetermined value and the low frequency edge component detected by the second detecting means is smaller than a predetermined value and the high frequency edge component detected by the third detecting means is larger than a predetermined value, An image area identifying device having an image area identifying means for identifying a halftone image.