JPS6313577A - Image processing method - 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] Field of the Invention The present invention is directed to a digital image forming apparatus, particularly in an output unit, which uses an area density modulation method to generate pseudo-intermediate images using an output means for low gradation expression such as binary or ternary. The present invention relates to an image processing method for emphasizing the outline of an image when expressing tone.

<Prior Art> Conventionally, this type of technology has been used to enhance the outline of digitally processed images in recent facsimiles and the like for black and white images. This is achieved by performing a conventionally known spatial filtering process using a convolution operation to emphasize the contour. In addition, in the conventional technology, the image resolution was about 8 lines/mm, and since text was mainly processed, there was no problem even if the image was binarized using a fixed threshold after contour enhancement.

Furthermore, when attempting to express halftones in photographs or the like by binary output, pseudo halftones have been expressed by area density modulation such as dithering.

However, in recent years, attempts have been made to faithfully reproduce both text and photographs at a high resolution of 16 lines/mm in digital copying machines that copy manuscripts containing text and photographs. In this case, the aim is to achieve contradictory effects in the text and halftone areas, so as a technique to separate the two image areas,
The image area is determined from the difference between the maximum and minimum pixel density within an image area of a specific size, and image area separation processing such as binarization by selecting a fixed threshold or dither threshold is performed. However, there are still many imperfections in determining the area size, etc., and it also has the disadvantage that it cannot be distinguished in documents with a background color.

Further, in the case of a color image, since the image signal itself is separated into a plurality of color signals, it is difficult to separate the image areas, and a technology that can improve the reproduction of color characters has not yet been developed.

OBJECTIVES> An object of the present invention is to provide an image processing method that matches area modulation with contour enhancement, which cannot be achieved with the conventional techniques as described above.

<Embodiment> (Overview of Apparatus Mechanism) FIG. 1 is a perspective view of a digital color image forming apparatus according to an embodiment of the present invention, and FIG. 2 is a configuration diagram schematically showing FIG. 1. The present invention will be explained in detail based on FIGS. 1 and 2. The document table crow 1 places the document 20 on the plane -L.tm
It's happening. The document surface of the document 20 faces the surface of the document table glass l, and the document 20 is pressed by the thick plate 1a. , G, B) A reading sensor (hereinafter referred to as a CCD unit) 17 consisting of a CCD array consisting of a plurality of reading elements in three columns for three colors, and an exposure lamp 19 are mounted, and main scanning is performed by a main scanning wire 8a. It is coupled and driven by the motor 6a. The sub-scanning table 5a supports one end of the main-scanning wire 8a, and supports the sub-scanning wire 10a.
is connected to and driven by the sub-scanning motor 9a.

A recording paper 21 is placed on a recording table 2, and a copy image is recorded by a recording head (hereinafter referred to as a printer) 4.

Printer 4 prints yellow, magenta, cyan, and black (
A recording element (hereinafter referred to as a BJ head unit) 18 consisting of a multi-ink jet head (hereinafter referred to as a BJ head as a bubble jet head was used in the present invention) for four colors (hereinafter referred to as Y, M, C, BK) is mounted, and a main scanning wire The main scanning motor 6b is coupled to and driven by the main scanning motor 6b. The sub-scanning stand 5b supports one end of the main-scanning wire 8b, and is connected to and driven by the sub-scanning motor 9b by the sub-scanning wire 10b.

When trying to obtain a copy image in the above configuration, the reader 3 is driven by the main scanning motor 6a via the main scanning wire 8a and reciprocates in the main scanning direction. At this time, the exposure lamp I9 is turned on and the reading sensor 17 reads the document 20 from below and outputs image information as an electrical signal. Based on this electrical signal, the printer 4 is driven by the main scanning motor 6b via the main scanning wire 81), and prints on the recording paper 21 while reciprocating. At this time, the main scanning directions of the reading head 3 and the recording head 4 are set in opposite directions to each other in this embodiment. - After the first copying process in the upper scanning direction is completed and the exposure lamp 19 is turned off, the reader 3 and the printer 4 move in the main scanning direction and the curved direction, that is, in the sub scanning direction, to the position where the next main scanning is performed. do. At this time, leader 3
is the sub-scanning wire together with the sub-scanning stand 5a supporting the main-scanning wire 8a] [It is driven by the sub-scanning motor 9a via 1a to move to a predetermined position and stop. Further, the printer 4 has a sub-scanning stand 5b supporting the main-scanning wire 8b.
and the sub-scanning motor 91) via the sub-scanning wire lOb.
is driven by the motor to move to a predetermined position and stop 1/4° (device control operation...preliminary operation) Fig. 3 is a block diagram of the control circuit of the above-mentioned embodiment, and Fig. 4 shows the entire sequence. A timing chart and a flowchart of the program are shown in FIG.

First, an outline of the operation of the apparatus will be explained using FIGS. 4, 5, and 6. Note that the step numbers on the timing chart and flowchart are the same.

Sequence controller 23, image controller 2
4 both have a microcomputer unit in the center,
The sequence control of the apparatus and the timing of image data formation are programmed in each, and both microcomputers communicate data via line 39. To explain the sequence from the time the power is turned on, the sequence controller 23 initializes the copying machine in step 1 according to the flowchart in FIG. 5, and then returns the reader and printer to their home positions for main scanning and sub-scanning in step 2. conduct.

Next, in step 3, a recovery operation of the inkjet head is performed. The head recovery operation is performed using a porous member, etc. in order to forcibly remove ink stuck to the tip of the inkjet nozzle after the device has stopped for a long time, and also to remove a pool of liquid near the tip of the nozzle after the ink ejection operation. This operation is performed by pressing or sliding a material with good water supply properties against the tip of the head. In terms of sequence, the printer main scanning motor 6b is rotated in the backward direction and stopped by the detection output of the recovery system position sensor 22. Next, a drive mechanism such as a solenoid for pressing the porous member against the head is turned on, and the porous member is pressed against the nozzle tip for a predetermined period of time. After finishing printer main scanning motor 7b
is rotated in the forward direction and stopped by the detection output of the printer main scanning home position sensor 12.

Next, in step 4, the head is operated without a cap in order to prevent the viscosity of the ink at the nozzle tip from changing during the pause period before the copying operation of the apparatus. This is accomplished by turning on a drive mechanism such as a solenoid that applies the cap at the home position of the printer. Next, in step 5, the operator waits for input from the operation unit 25, decodes the manually entered data, and sets the copy mode.In step 6, it is determined whether or not it is a copy start command, and if it is not a copy start, step Returning to step 5, if copying is to be started, the process proceeds to step 7, where the cap drive of the head is released in order to start the copying operation. Next, the process advances to step 8 to perform idle ejection processing for the head prior to the copying operation. Ejection processing is a process performed to perform stable recording, and in order to prevent uneven ejection at the start of ejection for image formation caused by changes in the viscosity of the ink remaining in the inkjet nozzle, there is a copy pause time, This is an operation for ejecting and discarding ink within an inkjet nozzle based on programmed conditions such as the internal temperature of the apparatus (a temperature sensor is not shown) and the duration of copying. Next, in step 9, the original exposure lamp 19 is turned on and shading correction processing is performed. Shading correction involves reading a standard white plate that serves as a reference for white data prior to scanning the original, and sampling data for correcting aberrations of the optical system lens and sensitivity variations of each bit of the CCD sensor.

Next, the process proceeds to step 10, and it is determined whether or not it is immediately after the start of the copy start, and if it is immediately after the start, that is, before the start of the first main scan, the process proceeds to step 11, and if it is the second or subsequent time, the process proceeds to step I2. In step 11, a head recovery operation is performed in anticipation of a long period of suspension of the apparatus.

The recovery operation in this case is the same as that described in step 3. Next, the process advances to step 12 to start main scanning. (
Please refer to Figure 6 for each signal) (Device control operation...
- Copying) For main scanning, first, speed data corresponding to the magnification ratio and a rotation start signal in the reader forward direction are sent to the motor driver circuit 26a of the reader via the line 40, and the reader main scanning motor 6a is turned on. Next, after taking a delay time for synchronizing the reader and printer according to the magnification ratio, a rotation start signal in the printer's forward direction is sent to the motor driver circuit 26b of the printer via line 41 to turn on the printer main scanning motor 6b. . Main scanning motor 6a, 6 of reader and printer
The rotation speed of b is detected by the rotation speed detection Kawaguchi-tally encoder 7a.

The pulse (FG multiplied signal) from 7b (hereinafter referred to as an encoder) is compared with the rotation speed reference pulse by the motor driver circuits 26 a and 26 b, and is locked to a predetermined rotation speed by the control, resulting in a constant rotation speed. Further, each encoder pulse is sent to the video data synchronization signal generation circuit 28 and the head data synchronization signal generation circuit 38 via lines 42 and 43.

(Reader side processing) Next, the process advances to step 13 and a copying operation is performed. 7th below
This will be explained with reference to figures -e and 7-b. In the video data synchronization signal generation circuit 28, as shown in FIG. Video line enable signal (hereinafter referred to as V, L) indicating the video line enable signal.

E, ) is created as shown in FIGS. 6-a and 6-b.Furthermore, from the video data start signal inputted from the CCD drive circuit 29, the effective data width of all pixels of the CCD is indicated,
Outputs video data enable signals (V, D, E, ) synchronized with encoder pulses. At the same time, a CCD start signal is synchronized with the encoder pulse to instruct the CCD drive circuit 29 to read images to the CCDs corresponding to the three colors of blue (B), green (G), and red (R) on the CCD unit 17, respectively. It is fed through line 57. The analog video signals for three colors read in the CCI) unit 17 are gain-adjusted so that the sensor sensitivities of each color are equal, and then output as digital values with 8-bit depth through a line 44. At this time, a video data start signal indicating the valid data range of all pixels of the CCD is also output from the CCD drive circuit 29. B,
G.

R3 color digital video data (hereinafter referred to as video data)
is manually input to the reader synchronization circuit 30.

Now, to explain the video synchronization signal generation circuit 58, the video synchronization signal generation circuit 28 receives a signal P HRE G P from the leader register position sensor 15.
Line 45. V, T4. E, Shin-J is line 4
6 and the image controller 24, the values of V, L, 1°C, and signal are respectively inputted through lines 47, which are counted according to the copying magnification. After passing, the time delay until reaching the leading edge of the document, that is, the reading start position, is v, 1. , r>, by clicking the signal]. In addition, a signal video enable signal (
Hereinafter, V, E, signals) are inputted and inputted to the reader synchronization circuit 30 via line 48.

In the reader synchronization circuit 30, as shown in FIG. 6-0,
A positioning operation in the main scanning direction is performed for reading the same portion of the original by CCDs corresponding to each color of G and R.

In other words, if the interval between the CCDs corresponding to each color of B, G, and R is set to Ll, then the image at position S1 of the original is displayed on the CCD corresponding to each color.
If the main scanning speed is V, the input to
/V time difference. Therefore, until the image of point S1 is input to the CCD of R, which is input last in time,
The video data from the and G CCDs are temporarily stored in the buffer memory in the reader synchronization circuit 30, and the three color video data of the image B, G, and R of the S1 point are outputted from the reader synchronization circuit 30. Also, after the V, E and signals are input, that is, the video data of the original is input, the B and G signals are input.
, R outputs a video data area (V, D, A) signal indicating that the three color video data are complete. The 6th
-c The vertical direction in the diagram is the time axis, not the sub-scanning direction.

The video data subjected to color matching processing by the reader synchronization circuit is then inputted to a scaling buffer memory 31 and subjected to scaling processing.

(variable magnification processing) Here, the scaling process will be explained using FIG. 7.

The magnification processing in the main scanning direction is performed by keeping the scanning speed V1 of the printer constant and changing the scanning speed of the reader to Vl/n. (n is the magnification ratio). This is because the upper limit of the driving frequency of the Infusiera I head, which is the image forming means of the printer, is lower than the upper limit of the driving frequency of the CCD.

Therefore, in order to increase the copying speed when copying at the same size, the maximum inkjet drive frequency is used when copying at the same size.

At this time, a variable magnification mode signal is sent from the image controller 24 to the video data synchronization signal generation circuit 28 through line 49 in FIG.

E. The frequency division ratio of the motor encoder pulse of the reader is set so that the signal has a common frequency when the magnification is the same and when the magnification is changed (Figures 7-a and 7-b).

That is, as shown in Figure 7-a, the motor encoder pulse φM
When it is the same size, it is divided into 1/6 as shown in φM1, and the frequency is divided into 1/2.
When scaling down by 2 times, the frequency is divided by 1/12 as shown in φM1/2, when expanding by 2 times, the frequency is divided by 1/3 as shown in φM2, and when it is 3 times the frequency is divided by ]/2. Motor encode pulse φM is 2 times when its frequency is 1/2 of the same frequency.
When it is doubled, it becomes 1/2, and when it is tripled, it becomes 1/3, so φM 1. φM 2 , φM3. The frequencies of φM1/2 are actually the same frequency.

Fig. 7-b shows the reading position on the document, and shows the moving distance of the COD during a certain time t (=V, L, E sections). When reducing the size to 1/2, the distance to be moved is twice that of the original size, and when enlarging it to 2 times, the movement distance is 1/2 compared to the original size.

Further, the scaling process in the sub-scanning direction is performed using the video clock φ(CL
R sent from the reader synchronization circuit 30 in synchronization with K8)
This is done by controlling the address increments of the scaling buffer memory 31 when each pixel of the , G, and B video signals is stored in the scaling buffer memory 31 (FIG. 7-0).

This is achieved by inputting the scaling mode signal from the image controller 24 through the line 50 to the memory control circuit 32 and increasing or decreasing the number of clock pulses of the address counter when writing to the scaling buffer memory 31 in accordance with the scaling ratio. (Figure 7-d).

As a result, data of the same pixel is written to n addresses in the memory 59b of the double buffer memories 59a and 59b in the variable magnification buffer memory 31 in write mode (W) when enlarged by n times, and when reduced by 1/n. One pixel out of n pixels is written to one address, and when the read mode is entered, the address is incremented by the video clock φ-CLK8, and the interpolation and thinning of the pixel data is achieved. It will happen.

In this embodiment, the motor speed on the reading side is changed, but the motor speed on the recording side may also be changed.

Another function of the scaling buffer memory 31 will now be explained using FIG. 7-d. Variable scaling buffer memory 3
The double buffer memories 59a and 59b in 1 switch the address increment clock when writing and reading, but this is because the V, L, E signals are generated from the encoder pulses of the reader main scanning motor 6a. Therefore, if uneven rotation of the motor occurs, the positional information between each main scan over the entire sub-scanning area will be more accurate, but the frequency will be uneven.

In order to synchronize with the V, L, and E signals and to prevent fluctuations in the COD accumulation time, the COD image reading cycle is set to ■.

L, E, shall be 1/2 or less of the minimum value of the signal period, and the CCD
17 shift clock φ-CLK4 is a video clock,
In order to make the frequency more than twice that of φ-CLK8, the address clock when writing equal-sized copies of the double buffer memories 59a and 59b is the shift clock φ-CLK4 of the CCD 17.
When reading, use the reader.

The video clock φ-CLK8, which is a synchronizing signal for pixel data within the printer, is used.

As described above, the scaling buffer memory 31. In the variable magnification mode, the memory control circuit 32 interpolates pixel data in the sub-scanning direction;
In addition to the thinning operation, the COD accumulation time is kept constant, and a pixel reading operation is performed in synchronization with the encoder pulse of the reader main scanning motor 6a.

(Image signal processing) B that has been subjected to the above scaling processing in the scaling buffer memory 31
, G, and R are then sent to the image processing circuit 33, where they are subjected to the processing shown in the blocks of FIG. First, the R, G, and 93-color video data is corrected by the shading correction section 60 based on the data of the standard white board read in step 9. In this embodiment, since the image light is read within a range where the linearity of the CCl) exposure amount E and the optical output voltage V is maintained, the following correction is applied.

However, vs; output after shading correction V, C
Output Vmax from OD; Output Vsmax when reading a white board; Video data that has been corrected by setting output and decanting is input to the next logarithmic conversion unit 61, where it is converted from a light amount value to an ink density value. Complementary color conversion is performed, B, G,
The R video data is converted to y, m, and c density data, respectively. The conversion formula is expressed by the following equation, where the ink density is the D1 standard white plate reflected light amount Ep, and the image light amount is E.

D=-1og-Ep The three color density data after conversion is then processed by the black extraction/UCR section 62.
and is input to the edge extraction section 63. What is black extraction? Y, M
, C3 color density data to calculate the amount of black ink to be applied.・This is because if you try to express black (hereinafter referred to as Bk) using three colors of Y, M, and C ink, it is difficult to express complete black, and the amount of ink applied is large, resulting in "bleeding" and paper problems on the copy paper. This is to prevent excessive expansion. Also, UCR (undercolor removal) is Y, M when black ink is used by black extraction.

This is a method of subtracting the amount of ink of each C color in relation to the amount of black ink, and in this embodiment, the following equation was calculated.

Bk= (min (Y, M, C) a+] a2Y
out= (Y −a3 Bk)a4Mout= (M
-a5 Bk)a6 Cout== (C-87Bk)aB However, El, ~a8 is an arbitrary system edge extraction is a method that extracts edges and lines of the image, and then specifies the amount of edges extracted from the original image data. This is to emphasize the outline of the image by adding them in a relationship. In this example, edges were extracted using a 5×5 convolution mask in the main scanning and sub-scanning directions. In order to remove noise components from the extracted edge amount, we selected an arbitrary threshold so that low-level detection values were not added to the image data. Furthermore, the edge extraction section outputs video data valid signals (hereinafter referred to as V, I, V, signal) indicating an area where edge extraction using a Laplacian mask is possible in the video enabled state.

In other words, when using a 5x5 Laplacian mask, this Jl is the V, E, signal from the 3rd grain onwards after IJ becomes active.

■ Indicates that the double signal is output.

The density data Y0M and C after UCR are input to the masking section 64 and subjected to masking processing. Masking is a matrix calculation process to correct turbidity when ink is superimposed due to unnecessary absorption of ink, and the following calculations are performed.

However, all to a33 are arbitrary numbers.

Next, the masked Y, M, C three colors and Bk density data are input to the output gradation correction circuit 65, and the subsequent two
Correction can be added to flatten the gradation when expressing pseudo-halftones using the dither method used in the value conversion circuit. The correction formula is shown below.

Yout = (a5+ (Y-a52)l a5
3Mout = [as4(M-ass)l a5
6Cout = (a57(C-ass)l a5
9 However, 851 to a59 are arbitrary series numbers.

Next, output gradation corrected density data, Y, M.

C, Bl< and edge fiED are input to the binarization unit 66 and subjected to binarization processing.

In this embodiment, the binarization process uses a systematic dither method to uniformly binarize the image data, and then corrects the pixel of interest using eran data EI).

In other words, when correction is performed based on the truth table shown in FIG. 8-b, an image with blurred edges due to the systematic dithering method becomes a pseudo-halftone expression image with enhanced contours.

As described above, the video signal processed by the image processing circuit 33 and converted into binary signals of four colors Y, M, C, and Bk (hereinafter referred to as density data) for the inkjet head is stored in the memory 34 line 51 of the same amount as the reader/printer. 3 entered through
(Printer side processing) Here, before explaining the operation of the reader/printer synchronization memory 34, the head data synchronization signal generation circuit 37 will be explained.
As shown in Figures -d and 6-c, it is synchronized with the encoder pulse of the printer main scanning motor 61), and the reader main scanning direction position 1iff t+'1 information is sent, and the valid data of the sub-scanning direction resolution l is synchronized with the encoder pulse of the printer main scanning motor 61). Nozzle line enable signals (hereinafter N, L, E, ) indicating the range are generated. The N, L, E signals are sent to the synchronization signal generating circuit 38 through a line 52. A signal from the printer register position sensor 16 is input to the head synchronization signal generation circuit 38 through a line 53, and the time delay from when the BJ head unit 18 passes the registration position until it reaches the copying position is determined by N and L.

E. By counting signals, a signal indicating the copy width in the main scanning direction according to the copy paper size, that is, a nozzle enable signal for each color (hereinafter N, E,) is sent to the reader/printer synchronous memory 34 via line 54. Output to.

The reader/printer synchronization memory 34 buffers the speed difference between the reader main scanning motor 6a and the printer main scanning motor 6b, and synchronizes the density data input from the reader part with the speed of the printer, that is, N, L.

E. Output in synchronization with the signal. When the V, D, and V signals are inputted from the image processing circuit 33, only the valid portion of the video data is sequentially written in synchronization with V, L, and H, and the head synchronization signal generation circuit 38 to N.

When the E signal is input, that is, when the inkjet head is in the copying area, the density data written in the memory is read out sequentially as head data in synchronization with N, L, and E. The data of each recording head read from the reader/printer synchronization memory 34 is outputted to the printer synchronization circuit 35 through a line 55.

In the printer synchronization circuit 35, the data of the four colors Y, M, C, and Ilk, which have been color-separated from the three-point image, are manually input via the line 55 at the same time. The head data is subjected to a process of shifting the position by one inch of the distance in the main scanning direction between the heads corresponding to each color.

In other words, as shown in Figure 6-1. In order for the images to be superimposed at the same point in the main scanning direction of the Infusiera I head and printed at -1, the main scanning speed should be set to V, and each color head should be printed with a time delay of I7□/V. In other words, the rad data for M, C, and Bk colors is temporarily stored in the buffer memory in the printer synchronization circuit 35 up to the Y head position where the image is printed first in the forward direction of main scanning, and then sequentially from the printer synchronization circuit 35. This is achieved by outputting and inputting it to the printer head drive circuit 36.

In FIG. 6-f, the vertical direction is the time axis, not the sub-scanning direction.

Further, the N, E, signals are input to the printer synchronization circuit 35, and the N, E, signals are signals indicating the copy area of the Y head, and from these N, E, signals, the ejection section of each color head is indicated. Head dry enable signal for each color (
After that, it outputs 1(,D,E,signal)°, and the line 56
The signal is input to the printer head drive circuit 36 through the signal line. In the printer head drive circuit 36, N, E, signal, N,
Drive signals for the inkjet head within the printer head unit 18 and herad data corresponding to each color are output to the printer head unit 18 from L, E, signals, H, D, E, signals, and clock φ.

According to the above-described flow, the document is read by the image reader 3 and an image is formed by the printer 4. The image controller then generates V from the reader 3 and printer 4.
, E, signal and N, upon detecting the end of the E signal, determine the end of one line copying of main scanning.Step 1
4) Go to step 15.

(Post-processing) In step 15, the sequence controller 23 first turns off the exposure lamp 19 and turns off the motors to the respective motor driver circuits 26a and 26b of the reader and printer.
input the signal, and then send the speed data in the reverse direction and a rotation start signal to turn on each motor 6a, 6b.
, start reversing, and each main scanning home position 1
Stop at 1.12. At the same time, in step 16, the reader sub-scanning stepping motor 9a (hereinafter referred to as reader sub-scanning motor) is fed a predetermined number of pulses corresponding to the copying magnification in a rotation mode in the sub-scanning forward direction to feed the reader for one sub-scanning. Similarly, the printer sub-scanning stepping motor 9b (hereinafter referred to as printer sub-scanning motor) also performs sub-scanning feeding. Next, the process proceeds to step 17, where the sub-scanning counter is incremented, and in step 18, it is determined whether or not the copying 11 minute sub-scanning counter in the sub-scanning direction is progressing.If the count is not progressing, the process returns to step 8 and the main scanning starts. Repeat this until the sub-scanning counter is counted up.

When the sub-scanning counter increases, the process moves to step 2, where a predetermined number of pulses are sent to each sub-scanning motor of the reader printer in the sub-scanning backward rotation mode to return to the home position. Next, proceed to step 3 and perform a head recovery operation for cleaning the inkjet nozzle head after copying is completed.
Proceed to step 4 and cap the head, then step 5
Waits for the next copy command to be input. The above is an overview of the device operation.

(Spatial Filtering Process) Next, details of the filtering process will be explained using Figures 9-a''-e, Figures 1O and 11.As mentioned above, in this example, a 5x5 convolution mask is used. The mask is shown in Figure 9-a.The number of threads of the mask is expressed by the following relational expression, and the edge amounts E and D of the center pixel E.

can be extracted along with the polarity as a second derivative (Laplacian).

E-(A+B+C+D+F+G+H+I)/8 Generally speaking, if one scanning direction of an image is considered as a function f(x), the shading portion or color changing portion of the image is as shown in Figure 9-6 (A). 1, and in this case, there are four mediums that have the effect of emphasizing the change of 2 x2 in f (x). By using a square mesh Laplacian mask in the -1-1 sub-scanning direction of the image, this effect can be used to enhance the contours in all directions of the image 411. 1 In this embodiment, a convolution operation is used as the spatial filtering process, and contour enhancement by Love Run unmasking is shown. Now, if we try to apply the Love Run Anne described in (1) to the same scanning method as in this embodiment, the problem shown in FIG. 10 will occur. In other words, if we focus on the sub-scanning of [N and N -1-1 lines 1], the application range of the mask is C
CI) If the number of pixels is n, from the 3rd pixel to n-2 pixels 1
Up to 1. This is because when attempting to apply the J mask at discontinuous portions of read pixels, if the pixel of interest (center pixel) is brought to the edge of the image, the number of mask data pixels becomes insufficient.

To solve this problem, the pixel data at the discontinuous part, that is, N
The trailing end data of V, L, and E of the sub-scanning line of the line is stored in the memory, and the V of the sub-scanning of the N+1th line is stored.

It can be used as tip data for L, E, but for example, A
When reading an l-size original at a resolution of 16 lines/mm, 26.3 is calculated from AI longitudinal width x resolution x number of memory lines.
K11yte (as 8bit/pixel) and 1.345
A 6-step counter is required, and the control circuit is also complicated, which poses problems. Therefore, in the present invention, the following method was implemented to solve the two-legged problem. This will be explained using FIG. 11.

Assuming that the number of pixels in the document reading CCT is m, in this example, a 5×5 mask was used, so the pixels m-3 to m at the rear end of the V, L, E, and signal at the beginning of the sub-scanning N line are
When reading the N+1st sub-scanning line, the pixels up to this point are superimposed and read twice. In other words, by superimposing pixels 1 to 4, a Laplacian mask can be applied to all the pixels to be read, and an image with completely enhanced contours can be obtained. Also, at this time, the number n of data to be sent to the inkjet head is m-4, which means that the number of recording elements of the head may be smaller than the number of pixels (CCI). Furthermore, the feed width of the reader and printer at this time in sub-scanning l steps is also m -4. This can be expressed as nm-m-(x-1) where m; number of pixels read by the image sensor n: multi-vI [element r used for recording with the recording head]
−X: Number of meshes in one direction of the mask. ”2 formula is the case when all the reading pixels of the image sensor are used, and when not all are used, 92m-(X-
If 1) holds true, then J2. Also, regarding the main scanning direction,
If 86 in FIG. 11 is the edge of the document, the video data area section starts from S7. Regarding the trailing edge of the document in main scanning, the video data area section ends two places before the edge of the document. However, the reading is performed until the end of the original (edge). FIG. 12-a is a detailed circuit diagram of the edge extraction circuit 63 shown in FIG. 8a. A detailed explanation will be given below using this figure.

The three-color image signals Y, M, and C are stored in the ROM 101a,
.

After the levels of each color are independently converted by b and c, they are added by an adder Add and stored in the RAMs 102a to 102e.

At this time, in order to create the convolution mask shown in FIG.
A to 102e are selected, and the light pulse V port is used to store the selection. Therefore, the line of pixel A in FIG. 9-a is RAM 102 e, and the line of pixel B is RA
M 102 c , the line of pixel C is RAM 10
All image signals for one line are stored in 2a. The count output of the main scanning line counter 103 is decoded by decoders 104a-c, and the image signal gates 1-106a-c are turned on and off to extract sample pixels. That is, the decoder 104a decodes the output of the main scanning line counter 103 as it is, but the decoder 104Jc decodes the output of the main scanning line counter 103 twice by using the shift registers 105a to 105d.
The data that has been shifted four times is manually decoded. As a result of this operation, the image signals of the lines 1, GD, and 0 in the convolution mask of FIG.
They are simultaneously output from 106c. + (17a - c are latches for extracting sample pixels from the lines constituting the convolution mask; in the I 07 a stage, pixels 0, 0.■ are respectively from the left, and in the l 117 C stage, pixel O1 ■, [F] is the pixel at the 107th stage ■
. Therefore, in the process described above, the ROM1
The output of 0Q is the amount calculated by the following formula3, (A+B+C+I)ilc-IP+G+■lI)*a/
b +・■ However, here, a and b are constants for averaging, and in the above formula, a −, −1, and b = 9. The signal averaged by the ROM 109 is compared with the image signal of the pixel E of the convolution mask by the comparator 110, and a polarity signal edge polarity (hereinafter referred to as E
, P.

) is output. In this embodiment, E and P are low level signals when the averaged data is larger than the image signal of pixel E, and are high level signals in other cases.

Next, subtraction is performed by the gates 111 and 112 and the adder 113 using E and P, and an edge data signal (hereinafter referred to as E and D) of an amount expressed by the following formula is obtained.

) is output.

E-(A1B+C+D+E1F+G+H1I)*a/b
-・-■ That is, when E and P are at a low level, the gate 111 is selected, and the image signal of the pixel E is inverted and added by 111b, and when E and P are at a high level, the gate 112 is selected. is selected, and the averaged data is inverted and added by 112b.

In this way, errors in region determination are prevented by always subtracting the smaller E and D from the larger one.

E, D, , E, P obtained in this way are the 12th-
It is output to the binarization circuit shown in figure b.

65 is a ROM that corrects the gradation of the output system; 115 is a comparator that compares the level of the image signal of each color with the dither threshold value and outputs dot data; and 116 stores the dither matrix threshold value. Dither ROM, 117 is dither RO
This is an address generator for generating M addresses, and in this embodiment, two types of address generators 117a and 117b are provided so as to independently dither each color of Y, M, C and Bk.

The aforementioned E, P, and E, D are subjected to spatial filtering by the ROM 1]9, and then compared with the image signal immediately after the output system gradation correction by the comparator 118 to obtain edge dot data3. This edge dot data is logically summed with the dot data output from the comparator 11fi by an OR gate +20. Furthermore, the selector 121 outputs Totsu I data in which the outline of the image is emphasized based on E and P, and is sent to the printer synchronization circuit 35. This operation will be explained using FIG. 13.

Figure 13(b) shows the circle CC of the English letter rAJ in Figure 13(a).
1 Now, if the part in FIG. 13(b) is read by this embodiment and recorded using an 8×8 dither matrix without using E, D, , E, P, , as shown in FIG. 13(0). Here, Fig. 13(b)
Comparing and (c), the dot DI in (C).

D2 is unnecessary, and a dot is required in the D3 part,
This causes the image to become unclear. In this case (c)
is recorded only with the dot data output of the comparator 115, and is further recorded with the output of the comparator 118 and 120a, 120.
By calculating the logical sum in b, it is possible to record as shown in FIG. 13(d). That is, in the portions DI and D2 of FIG. 13(c), the dot data output of the comparator 115 is at a high level, so the edge data output of the comparator 118 is at a low level, and furthermore, E and P are at a low level. - The output of the OR gate 120b is selected. Conversely, in part 3, the output of the OR gate 120a is selected and a dot is recorded. Therefore, the outline of the image that has become unclear due to the use of the dither matrix can be changed to the 13th by using E, P and E, D.
The outline becomes clearer as shown in Figure (d).

<Effects> As explained above, the pixels of the image contour are extracted from the multi-tone pixels by spatial filtering processing, and the output is obtained by converting the pixels to the gradation level of the output means to detect the contour, and then (11
By performing image processing using the output converted to the desired gradation using 1-density modulation, the gradation of the image is improved,
It can emphasize the outline of an image.

[Brief explanation of drawings]

FIG. 1 is a perspective view of the device 1, 1, 1, and 11 of the embodiment of the present invention, and FIG. 2 is a schematic diagram of the device of the embodiment of the present invention. Figure 4 is a timing chart of the sequence, Figure 5 is a sequence flowchart 1, Figure 6-81 is a diagram showing the relationship between the reader's document and the reading synchronization signal, and Figure 6- I) Figure 6
Fig. 6-a is an enlarged view of part A, Fig. 6-c is an explanatory diagram of the positional shift of the reading CCD of each color, Fig. 6-(1 is a diagram showing the relationship between copy paper and recording synchronization signal, Fig. 6-(4) The figures are an enlarged view of part B in Fig. 6-d, Fig. 51 is an explanatory diagram of the positional deviation of the inkjet heads of each color, and Fig. 7-a is a frequency division according to the magnification ratio of the encoder pulse of the reader main scanning motor. Figure 7-b is a diagram showing the timing, and Figure 7-c is a diagram showing the reading pixel interval in the main scanning direction according to the magnification ratio, and Figure 7-c is a diagram showing the interpolation according to the magnification ratio.
7-d is a detailed circuit diagram of the scaling buffer memory 31 of FIG. 3; FIG. 7-e is a detailed circuit diagram of the video data synchronization signal generation circuit 28 of FIG. -f
The figure is a timing chart of the video data synchronization signal, Figure 8-a is a detailed circuit diagram of the image processing circuit 33, and Figure 8-1)
9-a is a diagram showing a convolution mask, FIG. 9-b is an explanatory diagram of contour enhancement by spatial filtering, and FIG. FIG. 10 is an explanatory diagram of a scanning method in a device similar to the embodiment of the present invention, FIG. 1t is an explanatory diagram of a scanning method according to an embodiment of the present invention, FIG. Figure b is a detailed circuit diagram of the binarization circuit, and Figure 13 is an explanatory diagram of contour enhancement processing.

Claims (1)

[Claims] (1) Extract pixels on the outline of the image by spatial filtering from the multi-gradation pixel data that makes up the image,
A contour is detected from the extracted pixel, the multi-gradation pixel data is converted to a gradation level of an output means, and is modulated by area density modulation, and an image is created by the output of the contour detection and the output of the modulation. An image processing method characterized by emphasizing the outline of.