JPS63274547A - Image processing device - Google Patents

Abstract

PURPOSE:To reduce the weight of a binary processing circuit by using a RAM for cumulative storage of errors required per color and providing a single arithmetic integrated printed circuit for binary processing instead of providing a plurality of the circuit according to the number of basic colors to be processed. CONSTITUTION:A divalent circuit is composed of a binary processing unit consisting of a dedicated binary IC 5 or a binary processing CPU, and four independent RAM's 13-16 corresponding to different colors. These four independent RAM's 13-16 are connected to the binary processing units in parallel, and either one of the four RAM's 13-16 is always selected by ports of IC's for parallel output drive by CPU which controls the entire processing system. Consequently, if the binary processing of the equivalent to N line is completed on single color data, another RAM is selected to proceed to another processing. A memory region for error data is protected until a binary processing is started on that color.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention is directed to the binarization of gradated image data using a color serial dot matrix printer, which cannot express Ps tone due to the size of the dot diameter during printing. It relates to a printing method and its processing device.

[Conventional technology]

When printing a gradated image using a serial dot matrix printer, each dot of the input image has II! Binarization processing is required for the data. Among the binarization processing methods, there is a method in which the threshold value of the iI element of interest to be binarized is determined from the accumulation of errors between the gradation values of the dots that have already been binarized and their original pixels. Are better. However, this method requires a memory for recording error data for each dot in terms of height for pixels two lines before the line currently being processed. For this reason, when the above-mentioned binarization processing method is used in a conventional color printer, 100 pages are processed for each negative color, which has the disadvantage of slowing down the processing speed. Also, 1 per color
Two binarization dedicated ICs and t14! ! Circuits with storage memories are disadvantageous in terms of cost and footprint.

[HM Points to be Solved by the Invention The present invention is suitable for the head configuration on the printer side when printing a color M image having gradations by binarization processing using a color printer based on the serial e-dof matrix method. By generating data, the purpose is to speed up the processing order of data to be handled by the binarization processing circuit and to reduce the size of the binarization processing circuit.

Means for Solving the C Interference Point] The present invention performs binarization processing on a plurality of basic colors, and utilizes the visual color mixture of these binarized N elements to create multicolor, gradation The binarization means used in Printing II for color development focuses on the gradation value of one pixel among the above binarization means, and uses this as the threshold value for binarization. A type of binarization method that uses the gradation value of each completed pixel and the value obtained by cumulative calculation of each error caused by binarization is called a gradation method. A RAM for cumulative storage of errors required for each color in an image processing circuit that is applied to color image data and outputs binary image data to a color serial dot matrix printer.
By having logic that switches the processing order of each color,
A feature is that the number of arithmetic ICs for binarization processing is one regardless of the number of colors to be processed.

[Effect]

Most of the serial dot F matrix printers that print using heads such as the Ikpakto dot method and the drop-on-demand inkjet method cannot change the size of the printed dots, so the image information has so-called gradations that grow in brightness and darkness. When printing images,
By creating a difference between dots with the same diameter and one having more dots per unit area and one having fewer dots, this can be expressed as a difference in brightness through human vision.

In this image printing method, each W1 of the input image
Set an appropriate level threshold according to the gradation data of the pixel, and if the gradation value of the pixel of interest exceeds that level, print a dot corresponding to the pixel of interest. A binarization process is required in which printing is not performed unless the adjustment value reaches a threshold level.

Among these binarization processing methods, as a method to obtain a relatively excellent binarized image, the threshold level is set based on the dots that have already been binarized and the actual original pixels corresponding to those dots. There is one that is determined based on the accumulation of errors with the gradation value.

By binarizing processing using this method, the printer fa
In order to print in H, a small amount of memory is required to store error information for pixels that have already been binarized (more than 3 lines).In other words, a printer that prints 1024 pixels in the horizontal direction is required. If the error is to be expressed as a signed 8-bit value, the device must have a dedicated memory area of at least 3072 bytes.The data in this memory must be binarized for one line in the horizontal direction. However, when the serial printer finishes the printing process, it is updated one line at a time.

Today, among the serial dot matrix printers that occupy the majority of the market, it is rare to have a system that performs color printing with a 1-bin or 1-nozzle configuration; The mainstream is a head consisting of several pieces. To print images using these printers, a function called bit image printing, which the printer has, is usually used.

For printers with a 9-bit print head, this means that by transferring 9 lines of pixel binarized data to the printer and printing it once, the printer can print 1 line of data. It is assumed that printing has been performed. One N bin wire dovetail head, yellow, magenta,
Using a color serial printer with ribbons in four colors, cyan and black, we first binarize yellow into l'lk for N947, transfer this to the printer, and print one line of bit image. Then, similar processing is performed for magenta. The magenta ink ribbon may be selected by moving the ribbon mechanism up and down, and the bit image printing operation may be performed again. However, when performing the binarization process on the magenta image data, it is necessary to protect the data in the RAM used for storing error data for the yellow data generated in the previous process. Similarly, the maze/red error data, the cyan error data, and the black error data need to be protected until the processing system starts binarizing the data for that color in the next process. In order to implement this, the present invention does not use four systems of binarization circuits, but instead uses a binarization processing unit consisting of one binarization-dedicated IC or binarization processing CPU, and four independent binarization processing units corresponding to each color. It is assumed that a binarization circuit consisting of a RAM is used. All four RAMs are connected in parallel to the binarization processing unit, and only one of the four RAMs is always selected by the port of the parallel output IC driven by the CPU that controls the entire processing system. shall be As a result, for one color data,
Once the binarization process for N947 is completed, you can select another RAM and move on to processing another color. As a result, the storage area for error data will remain available until the next binarization process for that color begins. protected.

〔Example〕

FIG. 1 is a block diagram of a processing circuit according to one embodiment of the present invention. Here, the binarization processing arithmetic unit, which is a feature of the present invention, is constituted by the ROM 12 and the binarization-dedicated IC 5. In addition, in order to load this binarization processing calculation unit from the 84PUs, the MPU side address bus 38. A host-side system bus 48 consisting of an MPU-side data bus 8, an MPU-side read signal 39, and an MPU-side write signal 40 is developed. MPU
It is also connected to an address decoder 41 with 38 IJ address buses, and outputs selection signals for the binarization-only IC 5 and the parallel output IC 11.

FIG. 2 shows a block diagram of the entire processing system that controls the binarization processing circuit in FIG. 1. The circuit in FIG. 1 corresponds to the binarization unit 7 in FIG. An image obtained by pulling an NTSC video signal on a scanning line was used as a color human image. 1J2 In the figure, 42 indicates a video signal input, 43 is a video signal digitizing circuit,
The color signal is separated into RGB signals, and sampling and A/D conversion are performed. As a result, the digital values of the R signal are sequentially transferred to the RAM 57, and the digitized values of the G and B signals are also transferred to the RAM 57. The data is transferred to RAM 58 and RAM 59. This data processing is performed by hardware that operates in synchronization with the video signal.
Processing such as value conversion is performed by the MPU 8 via the system bus 48, and these operations are written in a program written in the ROM 49.

Further, in this embodiment, the printer-side mechanical body is controlled by another MPU 9. The MPU 9 drives the paper feed motor/carrier motor drive stage 50 via the system bus 47 to cause the printer to perform a printing operation. A program for this purpose is written in the ROM 51, and necessary parameters etc. are created in the RAM 17. Also MP
Parallel input/output IC45 receives and receives data from the U8 side.
, 46, and the seftronic switcher 7 ace 44, and the printer side has a print head for printing yellow, magenta, cyan, and black data generated by processing RGB data, 52 to 5.5.
is the drive circuit for these heads. Furthermore, a bus driver 56 for functional expansion is connected to the printer-side system bus.

In this example, an inkjet serial printer having the configuration shown in FIG. 343 was used as the image printing machine. Here, 18 is a paper feed roller, 19 is printing paper, 20 is a helad carriage, and 21 is a carriage guide shaft. Further, 22 is a yellow printing head, 23 is a magenta printing head, 24 is a cyan printing head, and 25 is a black printing head. The head and head carriage are D
Each head of each color has 12 drop-on-demand type ejection nozzles in a vertical row, which is fixed to a carriage belt driven by a motor C and moves in both directions as indicated by arrow A along a carriage guide shaft 21. The configuration is as follows:
The 12 nozzles in row 111 eject the same color ink. FIG. 4 shows a view of the opening of the head section as viewed from the paper surface direction, 26 is a yellow printing head, 27 is a magenta printing head, 28 is a cyan printing head, 2
9 is a black printing head. Further, 30 is one nozzle. Also, the vertical nozzle spacing 31 for all heads is 1/90 inch.

Further, the dimensions 32 are 5/90 inches, 33 are 55790 inches, and 34 are 5/90 inches.

As is clear from the configuration of its head, the inkjet serial printer used in this example can print four colors at the same time by printing one line in the horizontal direction, and generates a drive signal for each color head. By controlling the time intervals, multiple colors of ink can be overlaid on the same point on the paper surface. This allows ink colors to be mixed on the paper surface, making it possible to produce multiple colors. Also, due to the influence of dots that have not been completely overprinted and surrounding dots, when viewed through the human eye, it may be perceived as a mixed color consisting of multiple 1J dots. When used, it can be used to print images that are close to natural colors. This example actively utilizes this property and uses the printer used in this example to obtain excellent print output of full-color images in combination with the binarization 222) shown in Figure 1. The apparatus was designed to perform a printing operation only when moving in the direction of arrow B in FIG. 3, so that the order of ink reaching the paper surface was always yellow, magenta, cyan, and black. As a result, the order of color mixing was unified, and uniform color mixing was obtained over the entire paper surface.

Next, the printing procedure and operation of each part will be described.

In this embodiment, a general video signal was used as the input source of the gradated color image. Video signal 42 as shown in FIG.
is A/D converted by the video signal digitizing circuit 43, and taken in as 8-bit data into the RAM area for one screen of red, green, and blue. *5 in 2 diagrams
R-RAM, G-RAM, and B-RAM from 7 to 59 indicate memory elements corresponding to red, green, and blue, respectively.The size of the captured image is 512i1 element x 5121if element for each color both vertically and horizontally. It is. MPU8
First, access the B-RAM area, extract one line of data horizontally on the screen, and transfer this to the RAM.
Transfer to the data interpolation area on l0. At this time, data of 100 elements is accessed as having a length of 8 bits, and is stored at an odd address position in the interpolation area. An explanatory diagram of the interpolation area is shown in Fig. 5, where 35 is the start address,
The address will be increased by 1 for each horizontal square in the right direction, and 36 will become the first address +1023, and 35
The square below is located at the starting address +1024. The entire interpolation area grows to 1024×13 bytes), and address 37 becomes the start address +13311. Now, the data transferred per rai/minute is 512 bytes, as shown in Figure 'N5.
Next, access the B-RAM again at the 2nd line, which is located at the ○ mark on the first horizontal line, and save one line of data.
Transfer to the third horizontal line of the interpolation area. Same as below, total 7
The data transfer to the addresses marked with 0 shown in FIG. 5 is completed. After this, the data at the address marked Δ in Figure 5 is generated by interpolation calculation using the data marked ○.
Similarly, interpolation operations are performed for the entire interpolation area. As a result, image data of 1024 dots x 13 lines of yellow was created. Next, 1024 dots x 1 of this data
The two lines are sequentially transferred to the binarization unit 7. The binarization unit shown in FIG. 1 is wired so that 8-bit gradation data for each pixel is manually input from the data bus 6 on the MPU side to the address line side of the ROM 12. Also, at this stage, the parallel output ICII drove A12 to A88 of the address lines of ROM12, and wrote each level of yellow color ink for each level of blue data on the video signal. It is controlled to specify a 256-byte address area. Similarly, ICI 1 for parallel output has output signal line 4 as logical true,
The signals Ill, 2, and 3 are logic voltages. This allows RAM
Among 13 to 1G, only RAM13 becomes active. In this state, the binarization mrc5 binarizes every 8 bytes input, and converts the 0N10FF of each byte into 170 of each bit from the 8th bit to the 1st bit of the data port 6. , binarize and output.

The MPU in FIG. 2, which has received this data, uses the Centronics interface 44 in FIG. 12 to transfer it to the MPUe that controls the printer device. MP
On the Ue side, the bit image printing mode is specified as the operation mode, and the received data is stored in the bit pattern storage area for the yellow head.

Here, the MPUB side again accesses the green data area, destroys all the data in the interpolation area used for the blue data earlier, performs interpolation calculations on the green data, and when this is completed, performs the binarization operation. Therefore, only the signal 113 in FIG. 1 is made logically true, and only the RAM 14 is activated. In addition, PB4 to PBO of parallel output ICII
on the ROM 12 so that the area for generating magenta data from green data is selected. As a result, even while the binarization process is being performed to generate magenta data, it is inevitable that the previously processed yellow error data will be destroyed.
converts the generated binary data for magenta to 1024
Transfer dots x 12 lines to the printer/printer side. The same process as above is performed on the image data from the red signal, and the shear/binarized data is also transferred to the printer side. At this time, on the binary conversion unit shown in FIG.
Make only signals 1 and 2 logically true so that only R
The contents of AM13 and RAM14 are not destroyed. Next, only the lowest level value (that is, the darkest value) of the data in R-RAM, G-RAM, and B-RAM is accessed, and interpolation is performed. Perform value processing. Here, RAM1
Since only G is selected, only the signal mi is set to be logically true.

This creates black data. When this is transferred to the printer side, bit image printing data for each of yellow, magenta, cyan, and black is generated on the printer side in a configuration of 1024 dots x 12 dots.

On the printer side, a buffer area in which data for bit image printing is stored is prepared on the RAM 17 in FIG. 2 in a 16-bit configuration for each color.

These first addresses are YBUF, MBUF, CBUF, yellow, magenta, cyan, black, respectively.
BBUF. Further, in order to access the data of each color, the software on MPU0 has pointers YPT, MPT, and CPT1BPT. On the printer side, when printing, print timing is created using an encoder signal whose level changes every 1/180 inch/inch. When starting printing, YPT=0,
MPT=-5, CPT=-60, BPT=-85, and for each color, add +2 to the pointer every time the print timing occurs, and if the pointer value is negative or 2048 or more. If there is, the data is not output to the head, and when the pointer value is 0 or more and 2048 is not cleared, printing is performed by transferring the data at the address indicated by (iBUF+1PT) to the head a (i
(Y, M, C1B) Here, the initial value of the pointer is in the order of 0, -5, -60, -65 because the print timing occurrence is set to 1 for the head spacing shown in Figure 4. /
This is because it was done every 90 inches. The reason why the pointer is incremented by +2 is because the data output to the head each time is 12 bits, and the buffer is 16 bits per time.
This is because bit data is allocated. This flowchart is shown in FIG.
UF) indicates an address on the memory, and 1HEAD indicates a data output port to a one-color head.

Now, using the above method, we have binarized 12 lines of data on the screen and printed a bit image as one line on the printer, as shown in Figure 1.
All error data for each color on M1e is saved, and then binarization can be started for the next row.This embodiment eliminates the need to create four binarization units by repeating 0 or more. Both printers were able to print full-color images using a color printer consisting of four colors: yellow, magenta, cyan, and black.

〔Effect of the invention〕

The present invention is applicable to the binarization printing means for color images that develops gradations, and even if the binarization-dedicated IC or binarization processing CPU is configured as one, multi-color printing data is generated. made it possible to do so.

Furthermore, since data is received and passed in the optimal processing order for the printer's bit image printing mode, the next row of data is transferred during the printer's printing operation.
Because the value conversion process progresses, the print process can be completed in a shorter time than in the past.

[Brief explanation of drawings]

FIG. 1 is a block diagram of a processing circuit in one embodiment of the present invention, FIG. 2 is a block diagram of a processing system in one embodiment, FIG. 3 is a schematic diagram of a printer used in the embodiment, and FIG. 4 is a schematic diagram viewed from the opening surface of the head section. FIG. 5 is an explanatory diagram of the interpolation area. FIG. 6 is a diagram showing a flowchart during printing. Applicant: Seiko Epson Corporation 19 Figure 3 (Quarterly; inch) Figure 4 Figure 5 σ Rib Figure 6

Claims (1)

[Claims] Each pixel has a gradation value of light and dark, and it corresponds to the above basic colors by inputting the data of a so-called pseudo-natural image in which multiple colors are produced by overlapping 3 to 4 basic colors. When using a color serial dot matrix printer with ink and ink ribbon of one color or more colors as the output device, use a dot matrix printer with a fixed print dot diameter. means to perform binarization processing in accordance with a plurality of basic colors of an input image, and to print a multicolor gradation image by utilizing printing density and visual color mixing of those binarized pixels. , among the binarization methods used for the above means,
Focusing on the gradation value of one pixel, and using the gradation value of each pixel that has already been binarized as a threshold value when binarizing it, In an image processing circuit that implements a type of binarization method that uses values obtained by cumulative calculation of a group of processing errors caused by the presence or absence of dots when converted to a value, it is necessary to calculate the error required for each color. An image processing device comprising a RAM for cumulative storage, and a single integrated circuit for binarization calculation regardless of the number of basic colors to be processed.