JPS61146068A - Deciding method for image area - Google Patents

Abstract

PURPOSE:To decide respective image areas easily and securely on real-time basis by deciding on a pattern image part and a variable density part which are both present in the same original according to image information read by a scanner. CONSTITUTION:An image signal read by a readout part 11 is converted into a digital image signal (b), which is supplied to an image area decision part 16 and also applied to a pseudo half-tone processing part 13 and a simple binary- coding processing part 14. The image area decision part 16 performs binary coding with a threshold level lower than the binary coding level of the simple binary coding processing part 14 to generate decision signals of the pattern image part and variable density image part on the basis of the binary-coded data. Those decision signals are applied as a switching signal (f) to a data switching part 15, which switches and sends out the output (d) of the simple binary-coding processing part 14 for the pattern image part and the output (c) of the pseudo halftone processing part 14 for the variable density image part.

Description

[Detailed description of the invention]

The present invention uses a scanner to sample the image of a document pixel by pixel and reads the image information, which is optimally read according to the image characteristics.
The present invention relates to image processing that involves value conversion processing, and particularly relates to an image region determination method for determining image characteristics of image information to be processed. The end of the disaster 1 Generally, when digital image information obtained by sampling a document image pixel by pixel with a scanner and performing multilevel quantization is subjected to binarization processing so that it can be recorded and reproduced on a dot printer, the image In order to improve reproducibility, it is necessary to perform optimal binarization according to the characteristics of the original image. That is, in the case of pattern images such as characters, symbols, figures, etc., simple binarization processing is performed with a fixed threshold value to make the recorded image as clear as possible. In addition, in the case of gradation images such as photographs, in order to particularly improve the reproducibility of intermediate tones, the visual field system integrates the density of pixels in the vicinity of each local area of the recorded image to be perceived by spatially averaging it. It is necessary to perform binarization by spatial filter processing (pseudo halftone processing) using a threshold matrix that utilizes the effect. However, when pattern image areas and grayscale image areas coexist in the same document, it is difficult to uniformly binarize them using either simple binarization processing or pseudo halftone processing. The gradation in the gray image area is lost, resulting in the reproduction of an extremely poor quality gray image.Furthermore, with pseudo halftone processing, the pattern image area is also subjected to spatial filter processing, resulting in the reproduction of a pattern image with blurred outlines. It turns out. Conventionally, therefore, switching between simple binarization processing and pseudo-halftone processing is performed while recognition of the image portion is performed by logical judgment according to a predetermined algorithm according to the image characteristics. However, if such a method is used, the software requires time to process the recognition of the image area, making it impossible to record and reproduce the image in real time in synchronization with the reading of the document image by the scanner. Further, although the processing time can be increased using hardware, there is a problem in that the device becomes complicated. Furthermore, conventionally, marks have been added in advance to the document image to distinguish between pattern image portions and grayscale image portions. Some devices are designed to switch between simple binarization processing and pseudo halftone processing while determining the image area by detecting marks during image reading and scanning, but these devices have poor operability. 1 loss The present invention was made in consideration of the above points, and when a pattern image area and a grayscale image area coexist in the same document, each image area is determined according to the image information read by the scanner. An object of the present invention is to provide an image area determination method that can easily and reliably perform the following in real time. 1. In order to achieve the object, the present invention uses a simple binary method suitable for pattern images of digital image information obtained by multi-level quantization of image information of a document read by sampling pixel by pixel with a scanner. When selectively performing conversion processing and binarization processing that also expresses intermediate tones suitable for grayscale images, a threshold value different from the threshold value for simple binarization processing of the digital image information is used. The image area of the pattern image part and the grayscale image part is determined based on the binarized data, and the edge processing of the determined grayscale image part can be performed optimally. . Hereinafter, one embodiment of the present invention will be described in detail with reference to the accompanying drawings. FIG. 1 shows a block configuration diagram of a digital copying machine in which the present invention is specifically implemented, and a dot-type printer 2 records and reproduces images according to image information of a document read by a scanner 1. It is now possible to The scanner 1 includes an optical system and a CC system that decomposes a document image into pixel units in the main and sub-scanning directions and sequentially reads the images.
A reading section 11 consisting of a D line sensor and a video amplifier;
The read image information a is converted into black and white, and then AD converted for multi-value quantization, and if necessary, shading correction and M
An image processing unit 12 that performs appropriate processing such as TF correction and background removal, and a pseudo intermediate unit that performs optimal binarization processing of the grayscale image portion and pattern image portion in parallel in accordance with the processed digital image information. Adjustment processing section 1
3, a simple binarization processing unit 14, a data switching unit 15 that switches between the respective binarized data c and d and selectively sends recording data e to the printer 2 side, and an image processing unit 1.
According to the digital image information processed by 2.
An image area determining unit 16 that determines whether the digital image information to be subjected to value conversion processing is for a grayscale image portion or a pattern image portion and provides a switching signal f to the data switching unit 15.
It consists of The pseudo-halftone processing section 13 performs binarization processing that also takes into account halftones obtained by spatial filter processing using conventional dithering methods, density pattern methods, submatrix methods, and the like. In the present invention, in particular, the image area determination unit I6 binarizes the digital image information using a threshold different from the threshold for simple binarization processing, and performs the binarization. The pixel to be processed is shifted by several pixels in at least one of main scanning or sub-scanning, and the result of ANDing the shifted data and the binarized data is /71 II. When the pixel is in the gray scale image area and the result is 110#, means is taken to determine that the pixel to be processed is in the pattern image area. FIG. 2 shows a specific example of the circuit configuration of the image area determining section 16 for implementing such an image area determining means. a binarization circuit 161 that binarizes digital image information with a threshold lower than the threshold for simple binarization processing;
A Y-direction shift processing circuit 162 that shifts the binarized data g by 8 pixels in the sub-scanning Y direction and then ANDs the shifted data and the binarized data g;
The binarized data h shifted by 8 pixels in the main scanning direction
an X-direction shift processing circuit 163 that shifts the data by 8 pixels in the direction and ANDs the shift data and the binarized data h; and a shift processing circuit 162 for each of the Y and X directions.
, 163 and a mixer 164 that mixes the processed data j, k according to the state of the simple binary data d when shifted by four pixels each. As shown in FIG. 3, the binarization circuit 161 compares the 6-bit digital image information quantized to, for example, 64 gradations by the digital comparator CMP with a 6-bit threshold value THI accordingly. If A>B, the OR circuit outputs logic tl I II, and if A3B, the OR circuit outputs logic "0". The Y direction shift processing circuit 162, as shown in FIG.
8-bit latch LATCH! , RAM with a capacity of 8 bits x 4, latch groups LATCH2 to LATCH5
and an AND circuit, which include a clock tl for each latch and an output control signal t2. RAM address signal t
3 and read/write control signals

[4] are each given from a controller (not shown). With this configuration, RAM address 0 is X=0
, the address is incremented pixel by pixel in the X direction. Also, the 8 bits of i/o 1 to 1108 in RAM correspond to the Y direction, and when i/o 1 corresponds to Y=yl, 1102 corresponds to yl-1, r t103 corresponds to yl
-2, . . . , 1108 correspond to yl-7, respectively. Now, X=x 1. When binary data g corresponding to position Y=71 is input, RAM address t3 is x
It corresponds to l. Also, as read data from RAM, 1101 is (xl, yl-1), 1102 is (xi, yl-2), +-+ 110s is (xi, yl-1), and 1102 is (xi, yl-2).
8) respectively. The read data is then latched in a 1-bit shifted form and stored in the RAM again.
is written to the same address x1. At this time, the binarized data g is transferred to i/o l, the read data from 1101 is transferred to i/o 2, the read data from 1102 is transferred to i/o 3,..., the read data from 1107 is transferred to 1108. Each is written. At this time 110
The read data from 8 is never rewritten. That is, the operation is a read/shift/write operation in the Y direction, which means that data is delayed by one pixel in the Y direction. This read/shift/write operation is performed from X=0 to X=m (m is the number of pixels in the X direction+the number of delayed pixels). Therefore, the binary data h is shifted by 8 pixels in the Y direction with respect to the input binary data g, and the binary data j4 is the input binary data g.
The value data g is shifted by 4 pixels in the Y direction, and the binary data j is output from the AND circuit.
is obtained by ANDing the shift data for each pixel and the input binary data g. Further, the X-direction shift processing circuit 163 is an 8-bit shift register SR, as shown in FIG. The output data of each bit and the Y direction shift processing circuit 16
2 and an AND circuit that performs AND with the binary data h sent from 2. The shift clock t5 corresponds to each pixel in the X direction, and is given from a controller (not shown) together with the clear signal t6. The binary data h to be input now is X=
When the output data corresponds to xl, 1 is set to xi-1, and 2 is set to xi-2. ..., and 7 correspond to xl-7, respectively. That is, ki is one pixel's worth of delayed data of the binary data h, in other words, it is pixel data that is one pixel before the current binary data h. As shown in FIG. 6, the mixer 1f34 is composed of a combination of an AND circuit, an OR circuit, and an inverter, and is configured to handle binary data j and
An AND signal is applied to the binary data sent from the direction shift processing circuit 163, and an edge signal is applied to indicate that j4 and 4 are ``1'' and each binary data j.k is ``0''. As a condition, a logic judgment signal corresponding to the state of the simple binarized data d is sent to the data switching unit 15 as a switching signal f based on the image area judgment result. The light reflectance when optically scanned, and the image processing unit 12 converts it into black and white,
It also shows the relationship with 64-value quantized digital image information, in which black on both sides of the original corresponds to level 64 and white to level 1, respectively. In the characteristics of the figure, the saturation of white when the light reflectance is around 1 is due to the removal of the background density of the original by background removal, and the saturation of black when the light reflectance is around 0 is above a certain level. This is to increase the number of gradations to be assigned to the intermediate tone level by combining all the densities into level 64. In the device configured as described above, the image area determines whether the current read pixel is in the pattern image area or the grayscale image area based on the digital image information obtained by optically scanning the original image. The operation for determination will be explained below. For example, assume that the characteristics of digital image information in the X direction at a certain sub-scanning position as shown in FIG. 8 have been obtained. The digital image information actually has a 6-bit configuration, but for the sake of clarity, it is represented here in an analog manner, with the vertical axis representing gradation levels from 1 to 64. In reality, the image area is determined from image information in both the X and V directions, but here we will use the
Only directional image information will be considered. In the figure, TH1 indicates the threshold value in the binarization circuit 61, and TH2 indicates the threshold value in the simple binarization processing section 14. Of the digital picture information, I31
．． The characteristic part of B2 is the grayscale image part, and the characteristic part of B3. B4 each indicates a pattern image portion. Also B5. B6 is Hara H
1. B7 indicates uneven density or light dirt on the background, and B7 indicates particularly white areas in photographs. However, when such digital image information is given to the image area determination section 16, the THI
The output data g is binarized using the threshold value , and is supplied as data h to the X direction shift processing circuit 163, where it is shifted by 8 pixels in the X direction. The data shifted by 8 pixels at this time is indicated by 8 in the figure. In the figure, P is a pixel unit, and a deviation 8P between data h and data 8 indicates a delay of 8 pixels. Therefore, as shown in the figure, the X-direction shift processing circuit 163 obtains output data k by ANDing the binary data h and its shift data with 8, and the output data k is used as the switching signal f by the data switching unit 15. will be given to At the same time, the digital image information is converted into simple binarization processing section 14.
The processed data d is binarized using a threshold value of TH2 and sent to the data switching section 15. Note that the output data d from the simple binarization processing unit 14 is converted to the actual binary data in order to match the processing delay in the pseudo halftone processing unit 13.
It is designed to be outputted four pixels later than the digitized data dl. Note that in FIG. 8, it is difficult to represent the pseudo-halftone processing in the same figure due to the content, and even if it is expressed forcibly, it will not make it easier to understand the image area determination operation. The output data C of the pseudo halftone processing section 13 is not particularly shown. Here, for convenience, it is sufficient to consider that the digital image information S delayed by four pixels in the X direction corresponds to the output data C of the pseudo halftone processing section 13. In this way, as is clear from the relationship in FIG. 8, when the switching signal f output from the mixer 164, that is, the output data k of the The digital image information for the pixel currently being processed is in the gray scale image area, and when the low level is "0", the image area is determined to be that the digital image information for the pixel currently being processed is in the pattern image area. Be able to do things accurately. Therefore, when the output data k of the X direction shift processing circuit 163 is given to the data switching section 15 as the switching signal f, the data switching section 15
5, the binary processing data C by the pseudo halftone processing section 13 is selected, and when the output data k is P0'', the data switching section 15 selects the binary processing data C by the simple binarization processing section 14. If d is selected, the digital image information is optimally binarized and the recorded data 8 is sent to the printer 2 side according to the determination of the image area of the grayscale image area or the pattern image area to produce a high quality image. It becomes possible to perform recording and playback of. In the case of Fig. 8, B in the digital image information
The portions 1 and B2 are binarized by pseudo halftone processing. Further, in the B3 portion, the edge portion is simply binarized by a width of four pixels, and the center portion is binarized by pseudo halftone processing. In this case, the halftone output at the center of -83 does not mean that the density is reduced, but that the density corresponds to the density of the original image. This part B3 shows a pattern image part with a thick line width, and since pattern images such as characters in the original document have sufficient density themselves, the quality of the recorded image in the 83 part is not affected by the processing of the present invention. As well as being no longer damaged,
Since the edge portion is subjected to simple binarization processing, a sharp pattern image with no blurred outline can be obtained. Furthermore, in the portion B4 indicating the pattern image portion with a narrow line width, simple binarization processing is performed over the entire width. In the characteristics of the recorded data e, the shaded portion indicates the data portion subjected to pseudo halftone processing. Also, Figure 9 is for explaining the operation of image area determination in the Y direction, and although the digital image information in the Y direction is not particularly shown here, 88 in Figure 1 is in the pattern image part in bold letters. , B9. B13 is in the gray image area, BIO
~B12 may be considered to correspond to the pattern image portions of fine letters, respectively. In this case, as in the case described above, when the output data j from the Y-direction shift processing circuit 162 becomes the high level II 1'g, the digital image information in the pixel currently being processed is in the grayscale image area. , and when it becomes low level rr On, it is determined that the image area is in the digital image information of the pixel currently being processed, but is in the pattern image area,
In the B9 and 813 portions, binarized data C by pseudo halftone processing is selected as recording data e, and 810 to 81
In the second part, simple binarized data d is selected,
In the 88th section, the data d that has been simply binarized is selected at its edges, and the binarized data C that has been subjected to pseudo halftone processing is selected at the center. In the image area determination method according to the present invention, it is possible to effectively determine that a grayscale image area in a document is a grayscale image area even if the grayscale image area has already been subjected to pseudo-halftone processing. In other words, pseudo-halftone processing expresses halftones by the density of binarized dots and the pattern of dot aggregates, and even with such expression, there is no difference between halftones. Therefore, if a method of the present invention is used to determine that the area is a gray image area when the density is higher than a certain level over a certain range, for example, a portion of the document with a high dot density is a gray image area. It becomes possible to determine that. Therefore, it is possible to record and reproduce a good image even in a gray-scale image portion of a document that has already been subjected to pseudo-halftone processing. As explained above, the digital image information is determined by the threshold value TH.
When the data binarized by I is shifted by 8 pixels in both the X and Y directions, each delayed data j, and the data c, which is binarized using only the AND signal of
For example, recording data e corresponding to the grayscale image portions B1 and B2 in FIG.
In the small portions at both ends of the pseudo-intermediate waste-processed data portion indicated by diagonal lines, single iA binarized data is generated even though the original image is a gray-scale image portion. In other words, if there is an edge portion indicated by double hatching in the figure consisting of four pixels at both ends of the grayscale image area, and the corresponding digital image information has a level equal to or higher than the threshold value TH2, simple binarization processing is performed. By logic”
1" is output, and if it has a level lower than the threshold TH2, it is converted to logic "0"' by simple binarization processing.
will be output. In that case, if the density of the grayscale image at the edge part of the original image is originally quite low or high, it will not be a problem, but if the density is between the thresholds THI and TH2, the logic # The output of OII also causes deterioration in the grayscale image portion that is recorded and reproduced. Therefore, in the present invention, especially in the mixer 164 in the image area determination unit 16, at the timing when j4 and 4 are '1', each binary data j and k are '0' as an edge condition. 2
By sending a logic judgment signal according to the state of the digitized data d to the data switching unit 15 as a switching signal f based on the image area judgment result, a simple 2 Valued data is II O,
, the edge portion is processed as a grayscale image area, and when the simple binarized data is tr 1 u, the edge portion is processed as a pattern image area. Specifically, in the mixer 164 of the 1-image area determination unit 16, the binary data Jlk are both 11011 and the 4
The edge condition is when the binary data j4゜k4 are both in the state of II 1 +1, which is at a position advanced by a pixel,
The simple binarized data d at that time can be given to the data switching section 15 as a switching signal f. At this time, if the density at the edge portion is low and its level is lower than the threshold value TH2, d="O", f="I", and the data switching unit 15 selects the pseudo halftone processing data C. selected. Further, when the density at the edge portion is high and its level is higher than the threshold value TH2, d="1", f="O'", and the data switching unit 15 selects the simple binarized data at that time. d is selected. The double-hatched portions of the recorded data e in FIGS. 8 and 9 indicate edge output by pseudo halftone processing. Therefore, J! The edge portions are processed according to the actual shading state of one image, and the quality of the shading image to be recorded and reproduced is improved. In the above embodiment, the digital image information is set to the threshold value T.
The image area is determined based on the delayed data obtained by shifting the binarized data by HI by several pixels (8 pixels) in both the X and Y directions.
It should be noted that the image area may be determined according to delay data in either the X or Y direction, and that each circuit configuration is not limited to the one described above and can be changed as appropriate. Needless to say, for example, the digital image information to be processed is not limited to image information of a document read by a scanner, but may also be digital image information sent from a host computer in a certain system. Furthermore, in the embodiment described above, the image area determination unit 16
Although the threshold value in the binarization circuit 161 and the threshold value TH2 in the simple binarization processing section 14 are fixed, each of the threshold values THI and TH2 may have a hysteresis characteristic. That is, for example, the binarized data g based on the threshold value THI is logical II.
The level of the threshold TH1 is made to change somewhat depending on whether it is 1gg or NOI+. As a result, for example, once it is determined that the area is a grayscale image area, even if a low-density area appears in that area, it is difficult to convert to a binary area, and the determination mode for the grayscale image area can be maintained. become. Further, in the case of the above embodiment, the amount of delay of each digital image information in the X direction and the Y direction is set to 8 pixels, and the pseudo halftone processing section 13. Simple binarization processing unit 1
4 and the image area determination unit 16 so that each processing data is for the same pixel, the timing is set in consideration of each processing delay, and in particular, the processing delay in the pseudo halftone processing unit 13 is taken into consideration. Simply binarize the edge part of the pattern image by 4 pixels,
I am trying to process the center part with pseudo halftone processing,
It is possible to set various delay amounts for the digital image information in the X and Y directions, and one of the multiple delay amounts can be selectively specified depending on the state of the original image. You will be able to take measures easily. As mentioned above, in the image region determination method according to the present invention, a threshold different from the threshold for performing simple binarization processing on digital image information that has been multivalued quantized in units of pixels to be processed is used. The current digital image information is in the pattern image area based on the data that is binarized with the value and then shifted by several pixels in the main scanning direction and/or the sub-scanning direction. It has the excellent advantage that the image area can be accurately determined in real time using simple means.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a digital copying machine in which an image area determination method according to the present invention is specifically implemented, and FIG. 2 is a block diagram showing an example of a circuit configuration for concretely implementing the present invention. FIG. 3 is a block diagram showing an example of the configuration of the binarization circuit in the same embodiment, FIG. 4 is a block diagram showing an example of the configuration of the Y-direction shift processing circuit, and FIG. 5 is a block diagram of the X-direction shift processing circuit. Block diagram showing a configuration example, No. 6
Figure 7 is a block diagram showing an example of the configuration of the mixer, Figure 7 is a characteristic diagram showing the density gradation characteristics of image information with respect to light reflectance when the document surface is optically scanned by a scanner, and Figure 8 is a characteristic diagram showing the density gradation characteristics of image information in the main scanning direction. A time chart showing the timing of each part signal when determining an image area. FIG. 9 is a time chart showing the timing of each part signal during image area determination in the sub-scanning direction. 11... Reading section 12... Image processing section 13... Pseudo halftone processing section 14... Simple binarization processing section 15.
・Data switching unit I6...Image area determination unit 161・
...Binarization circuit 162...Y direction shift processing circuit
163...X direction shift processing circuit 164...Mixer Fig. 1

Claims (1)

[Claims] Simple binarization processing suitable for pattern images of digital image information obtained by multilevel quantization of the image information of the original read by sampling pixel by pixel with a scanner and halftone processing suitable for grayscale images. When selectively performing binarization processing, which expresses is shifted by at least several pixels in either main scanning or sub-scanning, and when the result of ANDing the shifted data and the binarized data is "1", the pixel to be processed is a grayscale image area. means for determining that the pixel to be processed is in the pattern image area when the result is "0", and a simple 2 corresponding to the edge part of the determined grayscale image area.
When the value data is "0", the edge part is processed as a grayscale image area, and when the simple binary data is "1", the edge part is processed as a pattern image area. Image region determination method.