JPH0242872A - Halftone image detection device - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a halftone image detection device for detecting a halftone image contained in a scanned paper surface used in a facsimile machine.

Generally, it is known that when a paper surface consisting of non-halftone images such as characters and line drawings and halftone images of photographs is scanned at a certain value of sub-scanning line density, moiré occurs in the halftone strokes. In order to prevent this moiré V, it is necessary to scan the sub-scanning line density at a value that makes moire inconspicuous in a portion including one halftone image.

In order to automatically switch the sub-scanning line density at the halftone dot area during scanning, it is essential to detect the halftone dot area by determining the characteristics of the image.

Conventionally, a method has been proposed in which the sub-scanning line density needs to be changed by marking a mark line or the like on the edge of the paper in advance and reading the mark. Furthermore, in recent years, in order to relieve the trouble of making such marks, a method has been developed and put into practical use that determines the characteristics of the image on the paper by analyzing the image signal and automatically switches the image.

As a method of determining the characteristics of this image, there is a method of measuring the density of points where the image signal changes from white to black or from black to white. This method can determine whether an image consists of thin characters, detailed designs, etc., or large characters, simple designs, etc. However, since the original purpose of this method is to determine the complexity of an image, it cannot necessarily be used as a halftone image detection method.

In general, halftone images are characterized by a high density of changing points because they consist of thin white dots and black dots, but unlike non-halftone areas, a distinctive feature of halftone images is that the white dots and black dots have a regular periodicity. Since the black dots are lined up in parallel, the interval between one black dot and the next black dot is the length of the halftone dot period.

The present inventors focused on the outstanding characteristics of one halftone image, and determined the first position of the pixel signal sequence near the center of one halftone pixel, and the center of that halftone pixel. The first and second pixel signals are respectively obtained from the second position of the pixel signal train, which is near the center of the next halftone dot pixel, and the output of one of them is inverted. The logical product of both is obtained by shifting the pixel signal string one pixel at a time, and the period of the output signal string consisting of "1P" or "O" is based on the halftone dot formation period of the halftone dot pixel. This is to determine whether the period is within a predetermined period range.

With the above configuration, the present invention can detect the halftone dot image and the non-halftone dot image on the surface of the scanning paper, and can prevent the occurrence of moiré when scanning the halftone dot image.

An embodiment of the present invention will be described below with reference to the drawings.

As already mentioned, in the binary image signal obtained when the halftone image area is scanned, the signal indicating the white pixel value and the signal indicating the black pixel value are regularly arranged at the halftone dot period. However, in order to distinguish between the image signals of the non-halftone area and the halftone area, it is sufficient to determine whether the image signal consists of a black and white signal having a halftone period.

Now, if we sample between - halftone dots with 12 pixels,
The halftone dot period is 12 pixels. In order to effectively discriminate between a non-halftone area and a halftone area, the number of pixels required for discrimination may be increased. In other words, the number of pixels between 10 and 20 halftone dots (
It is only necessary to determine whether an image signal (from 120 pixels to 240 pixels) has a halftone dot period0. However, increasing the number of pixels in this way increases the length on the scanning paper surface necessary for determination. This has the drawback that the scanning direction of the scanner must be precisely aligned with the halftone dot array on the paper surface.

In order to eliminate this drawback, as a method for effectively distinguishing non-halftone areas from halftone areas using a small number of pixels, the first and second A pixel signal is obtained, one of the outputs is inverted to obtain the logical product of both, and the period of the signal string consisting of the logical output "1" or "0" is the halftone dot formation period of the halftone dot pixel. The headline is a two-step method of determining whether or not the cycle is within a predetermined periodicity range based on .

FIG. 1 is a block diagram of a transmitter having a function of detecting halftone dot images and switching sub-scanning line density. The area surrounded by the broken line 12 constitutes a halftone dot image (elephant detector).The following explanation will be made assuming that the white of the paper surface is O9 and the black is 1, and the halftone dot period is 12 pixels.

The image signal read by the scanning/reading mechanism 1 is binarized by the binarizer 2, and inputted to an image signal processor 9 that performs processing such as compression, and a logic circuit section 3.

The logic circuit section 3 is located at a first position of the pixel signal train shift register near the center of one halftone pixel, and between the center of that halftone pixel and the center of the next adjacent halftone pixel. The first and second pixel signals are respectively obtained from the second position of the pixel signal column shift register which is near the section, and one of the outputs is inverted and the logical product of both is obtained from the second position of the pixel signal column shift register. The pixel signal string is obtained by shifting it pixel by pixel.

The detailed configuration of the logic circuit section will be described below with reference to FIG.

First, the pixel signal sequence a is input into the shift register 21 one pixel at a time, and among the parallel outputs from the 0 shift register 21, four outputs S1, 82° S3, S4 every 6 pixels are output.
The polarities of the second output S2 and the fourth output S4 are inverted by the signal inverters 22 and 23 and input to the AND gate 24. AND gate 24 matches,
A mismatch signal C is output.

In other words, when an image signal a in which black pixels and white pixels are arranged alternately every 6 pixels is input to the shift register 21, a coincidence signal C is generated.
is output from the AND gate 24.

A logic output can be obtained even with such a simple configuration.

The characteristics of obtaining a logic output from this specific position will be described below.

For example, in a halftone image with rapid changes, halftone dots of different sizes occur on one scanning line, as shown in FIG. 3(a).
When this is partially enlarged, it becomes halftone dots A and B in FIG. 3(b). In this case, it can be seen that all the scanning pixel signal sequences at the same position within the adjacent halftone dots A and B are not necessarily in the same state. That is, the scanning pixel signal sequence a5 of halftone dot A
~a7 (black) and al, an, atz (white) are halftone dot B (7) scanning pixel signal sequences b5 to b7 (black) and b
Corresponding to l, btl, and bt2 (white), they are in the same state. However, scanning pixel signal sequences a2 to a4 of halftone dot A (
white) and a8 to alO (white) are scanning pixel signal sequences b2 to b4 (black) and b8 to b of halftone dot B at the corresponding positions.
It is not in the same state as 10 (black). Therefore, the logic between the white signals of the scanning pixel signal sequences a2 to a4 and a8 to alo of the halftone dot A and the black signals of the scanning pixel signal sequences b2 to b4 and b8 to blO of the halftone dot B at the corresponding positions. Even if the product is obtained, it cannot be recognized as a halftone image.

By the way, in such a halftone image, a6 and b6, which are the centers of each halftone dot, are always black signals, and midpoints aL2 and bl2 between the centers of both halftone dots A and H are always white 0.
becomes. That is, as described above, the binary pixel signal sequence when scanning the halftone image part is the signal (a) indicating the white pixel value.
l2. bt2) and a signal indicating the black pixel value (a6.
bs ) are regularly arranged at a halftone dot period at regular intervals. However, on the other hand, the pixels at that interval (for example, a8 to b4) are not necessarily regular.

Therefore, in order to accurately detect the halftone image portion, the image code string between the logical product input signals obtained to the AND circuit 24 is used as a mask pixel (indicated by The binary pixel signal of register 21 is not used.

Figure 3(C) has the same length of <24 pixels, 2 pixels of 0 that detects the white point, 2 pixels of 1 that detects the black point, and 4ill of mask pixels! This is an example of a simple logical product signal.

In this case, compared to the same signal in Figure 2 or Figure 3 (b), (a) accuracy will be further improved if there is a perfect match at 8 points, and (b) accuracy will be further improved if only one of the adjacent points matches. If it is considered to be a match, the accuracy will be lower, but it is more realistic if reading errors and "stretching" of the paper are taken into account.

Note that in Figure 2 or Figure 3 (b), the same signal is 0.1 for white and black for a total of 4 pixels, and in Figure 3 (C) for 8 pixels, but the circuit If the method is simplified, a halftone image can be detected even with one pixel of 0.1 each for white and black, although the accuracy will be lower.

Returning to the explanation of FIG. 1 again. The output signal C, which is the output of the logic circuit section 3, is input to the period determiner 4.

This period determiner 4 measures the interval between the point of change from mismatch to match and the next point of change from mismatch to match, and determines that the relationship with halftone dot period T is T-α, L≦T+β(α, β natural number). The function of the determiner 4 is to determine whether the output signal C from the logic circuit section 3 has a period within a predetermined range. Even if the image signal in the non-halftone area happens to coincide with the output signal CK, since the non-halftone area has no periodicity, the period determiner 4 detects the non-halftone area by measuring the interval between matching signals. It can be determined that the area is a point.

Note that the values of α and β of the period determiner 4 can be up to a maximum of T/2, but since the density of a normal photograph changes relatively gradually, both α and β may be about Tl2O. Further, by giving a width to the period determined by the period determiner 4 in this way, the detection ability of the halftone dot portion does not depend on slight deviations in the halftone dot period due to elongation, shrinkage, etc. of the paper surface.

Next, the output from the period determiner 4 is measured by a dot counter 5 for each scanning line. That is, - the number of halftone dots in a scanning line is measured. The measured values of the halftone dot counter 5 can further improve the accuracy of halftone dot detection. In other words, depending on this measurement value, the halftone dot area with a narrow range (- in the scanning line,
When it is recognized that the halftone dot area has a low value), it is not necessary to perform the halftone dot processing described below (because halftone dot areas with a narrow range have less moiré). It is often advantageous in post-processing to ignore the

Next, the block switching signal generator 6 outputs a sub-scanning line density switching signal based on the value measured by the halftone dot counter 5, and this switching signal is processed by the mixer 7 and the image signal processor 9. is mixed into the captured image signal and sent to the receiver. Furthermore, when reading one image, it is necessary to set the scanning line density finer for the halftone area, so for the next reading, the sub-scanning line density setting circuit 8 changes the scanning line density to the sub-scanning line density setting circuit 8 for the next reading. The scanning line density is set, and the sub-scanning line is driven at a new sub-scanning line density through a sub-scanning drive circuit 1°.

As explained above, the present invention includes a pixel signal string shift register that holds a binary white/black pixel signal string read in at a scanning line density finer than that of halftone dot pixels, and shifts the pixel signal string one pixel at a time; The first position of the pixel signal string shift register is near the center of one halftone pixel, and the pixel is near the middle between the center of that halftone pixel and the center of the next adjacent halftone pixel. The 11th and 2nd pixel signals are obtained from the second position of the signal string shift register, and one of the outputs is inverted and the logical product of both is obtained, and the pixel signal string of the pixel signal string shift register is converted to one pixel. Whether or not the period of the signal sequence consisting of the logic circuit and the 1'' or 0 pars that the logic circuit outputs is within a predetermined period range based on the halftone dot formation period of the halftone dot pixel. By providing a dot period determination means for determining whether the image is a halftone dot or a halftone dot period, it is possible to distinguish between halftone dots and non-halftone dots from a pixel signal sequence with a small number of pixels. This makes it possible to detect halftone dots without depending strongly on expansion and contraction of the paper surface.Therefore, it is possible to prevent moiré from occurring at the halftone dot portion during scanning.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of a transmitter using the halftone image detection device of the present invention, FIG. FIG. 3 is a conceptual diagram showing the operation of the logic circuit section. 1 Scanning reading mechanism, 2--binarizer, 3--logic circuit unit, 4-period determiner, 5--halftone counter, 6--switching signal generator, 7--mixer, meter ... Sub-scanning line density setting circuit, 9... Image signal processor, 21... Shift register, 22, 23... Signal inverter, 24... AND circuit. Name of agent: Patent attorney Shigetaka Awano and one other person

Claims (1)

[Claims] A pixel signal string shift register that holds a binary white/black pixel signal string read in at a scanning line density finer than that of the halftone dot pixels, and shifts it pixel by pixel; a first position of the pixel signal string shift register that is located near the center of the halftone dot pixel and the center of the next adjacent halftone pixel; a logic circuit that obtains first and second pixel signals from the above, inverts the output of one of them, and obtains the AND of both while shifting the pixel signal string of the pixel signal string shift register pixel by pixel; , a halftone period for determining whether the period of a signal string consisting of “1” or “0” sent out by the logic circuit is within a predetermined period range based on the halftone formation period of the halftone pixels; A halftone image detection device comprising: a determining means.