JP2000244746A - Image processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To judge whether or not all compression encoded image data can be stored in a memory before they are stored into the memory and to promptly deal with an image processing. SOLUTION: This image processor is provided with a decision means 212 for deciding the attribute information of respective blocks composed of the prescribed number of pixels, a compression encoding means 215 for encoding the respective blocks to a bit number decided by the kind of the attribute information, a counting means 216 for counting the attribute information for the respective kinds, a judgment means 290 for judging whether or not a total code amount can be stored in the memory based on a counted result, a resolution switching means 218 and a selective storage means 230 for storing compression codes in the memory in the case that the total code amount can be stored in the memory and for storing the image data switched to a low resolution in the memory in the case that it cannot be stored.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus for compressing image data, storing the compressed data in a memory, reading the image data stored in the memory, and generating output data for printing or the like.

[0002]

2. Description of the Related Art In a color image forming apparatus such as a color printer, a color copying machine, and a color facsimile, images of print colors Y (yellow), M (magenta), C (cyan), and Bk (black) are superimposed. Therefore, it is necessary to output the image data of each print color to the print head at an optimal timing so that the image of each print color does not shift. Therefore, image data of each print color is temporarily stored in a frame memory, and read out and output at an appropriate timing.

In order to shorten the image forming time in a color image forming apparatus such as a color printer, a color copier, a color facsimile, etc., processing of image data for each of Y, M, C, and Bk print colors is performed. Must be executed concurrently. For this reason, the image data for each print color is temporarily stored in the frame memory, and read out and output simultaneously and in parallel at an appropriate timing.

[0004]

In an apparatus for temporarily storing a large amount of image data in a memory, such as a color image forming apparatus, the image data is compressed so that the memory has a relatively small capacity. All image data. However, in some cases, the degree of compression cannot be increased due to the relationship between the type of image and the compression method, and as a result, the total amount of codes after compression may exceed the storage capacity of the memory. In such a case, in the conventional method, a problem as shown in FIGS. 16 and 17 has occurred.

FIG. 16 shows a processing method in which when the memory becomes full during storage, a warning of memory over is issued to stop processing or stop image formation on paper.
The processing method of FIG. 16 has a disadvantage that a desired image cannot be formed on paper. FIG. 17 shows a processing method in which when the memory becomes full during storage, the resolution is switched to a low resolution with a small data amount and the image is processed again from the beginning. FIG.
The processing method of (1) has a disadvantage that the processing time is unnecessarily long.

The present invention has been made in view of the above circumstances,
A processing method that enables determination of whether or not all of the compression-encoded image data can be stored in the memory before storing the data in the memory. The purpose is to be able to switch quickly to

[0007]

FIG. 1 shows a configuration example of the present invention. An image processing apparatus according to claim 1 is an image processing apparatus that stores image data in a memory, reads out the image data stored in the memory, and generates print data, and includes a predetermined number of pixels cut out from the image data. Attribute determining means 212 for determining attribute information indicating image characteristics of each block for each block; compression encoding means 215 for converting image data of each block into a code having a bit number determined by the type of attribute information; An attribute counting unit 216 for counting each type, a code amount determining unit 290 for determining whether or not the total code amount is an amount that can be stored in the memory based on a counting result for each type of attribute information; Resolution switching means 218 for switching the code, and if the total code amount is an amount that can be stored in the memory, the code converted by the compression coding means 215 is written in the memory. When the total code amount is an amount that cannot be stored in the memory, there is provided a selection storage unit 230 that stores in the memory the image data switched to the low resolution by the resolution switching unit 218 or the compression code thereof. I do. The degree of compression when a photographic image (continuous tone image) is compressed by the compression encoding means 215 is not very high. However, for a photographic image, the data amount can be greatly reduced by switching to a low resolution. Further, even when the resolution is switched to a low resolution, the degree of deterioration of the image quality is generally negligible. An image processing apparatus according to claim 2 is an image processing apparatus that stores image data in a memory, reads out the image data stored in the memory, and generates data for output, and outputs a predetermined number of pixels cut out from the image data. Attribute determining means 212 for determining, for each block, attribute information indicating the image characteristics of the block, and first compression encoding means 215 for converting the image data of each block into a code having a bit number determined by the type of attribute information.
A second compression encoding unit 218 for encoding image data by a compression method different from that of the first compression encoding unit;
Attribute counting means 216 for counting attribute information for each type
And a code amount determining means 290 for determining whether or not the total code amount is an amount that can be stored in the memory based on the counting result for each type of attribute information. The code converted by the first compression encoding means 215 is stored in a memory. If the total code amount is an amount that cannot be stored in the memory, the code is encoded by the first compression encoding means 215 and then the second And a selection storage means for storing the code coded by the compression coding means 218 in a memory. The output data read from the memory and generated may be print data, or may be data for recording on a recording medium such as a smart media. The selection storage means having the configuration of claim 2 is not shown in FIG. In FIG. 1, if the total code amount is an amount that can be stored in the memory, the code converted by the first compression encoding unit 215 is stored in the memory, and if the total code amount is an amount that cannot be stored in the memory, Although the selection storage means 218 for storing the code coded by the second compression coding means 218 in the memory is shown, this configuration is similar to the first compression coding means 215.
Is not high, but the second compression encoding means 21
The compression degree of 8 is effective for high images. An image processing apparatus according to claim 3, wherein the image data is stored in a memory, the image data stored in the memory is read out, and data for output is generated, and a predetermined number of pixels cut out from the image data are used. Determining means 212 for determining, for each block, attribute information indicating the image characteristics of the block composed of
Compression encoding means 215 for converting the image data of each block into a code having a bit number determined by the type of attribute information
Counting means 21 for counting attribute information for each type
6, a code amount determination unit 290 that determines whether the total code amount is an amount that can be stored in the memory based on a count result for each type of attribute information, and a switch that switches a processing method of image data according to the determination result. Means. The output data read from the memory and generated may be print data, or may be data for recording on a recording medium such as a smart media. The switching means of the configuration of claim 3 is not shown in FIG. This switching means may be, for example, a means for coping with memory shortage by switching the encoding method to increase the degree of compression. It may be a means to deal with. According to a fourth aspect of the present invention, in the image processing apparatus according to any one of the first to third aspects, the attribute determining means determines the attribute information for counting by the attribute counting means based on the low-resolution image data. And determining attribute information of each block. An image processing device according to claim 5 is the image processing device according to claim 1, wherein
The attribute determining means determines attribute information of each block based on prescanned image data when determining attribute information for counting by the attribute counting means.

[0008]

DESCRIPTION OF THE PREFERRED EMBODIMENTS [1] Print Image Control Unit (PIC Unit): FIG. 1 is a schematic diagram showing a system configuration of an image forming apparatus according to an embodiment, and FIG. 3 is a print image of the image forming apparatus shown in FIG. FIG. 3 is a block diagram illustrating a control unit (PIC unit; a processing unit that generates print data from image data).

[1-1] Outline of system: The illustrated system includes Y (yellow), M (magenta), C (cyan),
An electrophotographic image forming apparatus that outputs images of four colors Bk (black) on a transfer belt and simultaneously transfers the images onto paper, and includes an image forming unit (photosensitive drum) for each of Y, M, C, and Bk colors (A unit composed of a (PC), a developing unit, an LED array, etc.) are provided at equal intervals (L mm in the illustrated example) along the transfer belt.

[0010] As shown in FIG. 3, the processing in the PIC unit is as follows.
Basic processing is parallel processing of image data of each color of Y, M, C, and Bk. That is, when image data is sent from the scanner image processing unit 12, the image data of each color of Y, M, C, and Bk generated based on one scanning operation is
The data is simultaneously input to the PIC via the image interface 20. On the other hand, when image data is sent from the printer controller 14, a RIP (Raster I
image, Y, M, C,
The image data of each color of Bk is stored in the image interface 20.
Are simultaneously input to the PIC.

The image forming units corresponding to the colors Y, M, C, and Bk are arranged at equal intervals in the traveling direction of the transfer belt in the order of Y, M, C, and Bk. It is necessary to delay the development timings of the toners of the respective colors by a time corresponding to the following so that the toners of the respective colors of Y, M, C, and Bk are superimposed on the transfer belt without color shift. For this reason, in the illustrated system, each image forming timing of Y, M, C, and Bk is
Delay control is sequentially performed for a time corresponding to the above interval.

In the illustrated system, four LEDs are used.
In order to form an electrostatic latent image on the charged surface of each of the photoconductor drums for Y, M, C, and Bk by the scanning of the head, the printing air position shifts in the main scanning direction, and the photoconductor drum and the LED are displaced.
If a skew distortion occurs due to a deviation of the parallelism of the head, a color shift occurs. Therefore, by performing position correction and image correction on each of the Y, M, C, and Bk image data, the above-described phenomenon is prevented, and the occurrence of color shift is prevented.

[1-2] Outline of Processing in PIC Unit Next, a print image control unit (PIC unit) shown in FIG.
An outline of the processing in will be described along the data flow. In FIG. 3, a CPU that manages the flow of the processing in FIG. 3 and sets various setting values is omitted.

Image data of each color of Y, M, C, and Bk transferred from the scanner image processing unit 12 or the printer controller 14 is input to an encoding unit 21. In the encoding unit 21, GBTC is applied to each block of 4 × 4 pixels.
Fixed-length compression processing is performed according to the method. In the case of an A3 size original, after the fixed length compression process, a variable length coded compression process for each attribute is further performed. When performing variable-length coded compression processing, whether the total code amount after variable-length coded compression is an amount that can be stored in the frame memory 22 is determined by counting attribute information for each type. If the amount exceeds the storable amount, a resolution conversion process for switching the resolution to a low resolution is performed instead of the GBTC compression and the variable length coding compression.

In the fixed length compression processing, the image data of each block of 4 × 4 pixels is compressed into 48-bit data. In the variable-length coding compression processing, 48-bit data after the fixed-length compression processing is converted into any of 2 bits / 10 bits / 50 bits according to the attribute. This variable length coded compression processing is performed when the original image size is a size that does not require editing processing such as rotation (A3 size which is the same as the maximum size of output paper). In the resolution conversion process, the average value (8 bits) of the data values (8 bits each) of 2 × 2 pixels is used.
Bit) is a data value of each 2 × 2 pixel. The fixed-length compression processing, the variable-length coding compression processing, and the resolution conversion processing will be described later in detail. 4 by fixed length compression processing
The data of each color converted to 48 bits per 4 × 4 pixel, the data of each color converted to 48 bits per 4 × 4 pixel by fixed length compression processing, and further subjected to variable length coding and compression processing, and 4 The data of each color converted to 32 bits per 4 pixels is output to a 16-bit data bus.

The Y, M, C,
The data of each color of Bk is stored in the frame memory 22 of the sub-scanning delay controller 23. Here, when the original image size is a size (for example, A4 or less) that can be subjected to editing processing such as rotation, data that is subjected to only fixed length compression and has 48 bits per 4 × 4 pixels is stored. You. When the original image size is a size that does not require editing processing such as rotation (for example, A3), data that has been converted to 48 bits by fixed-length compression processing and then further subjected to variable-length coded compression processing is stored. You. Further, when the original image size is a size that does not require editing processing such as rotation (for example, A3), it is determined that the total code amount after the variable-length coding compression processing exceeds the amount that can be stored in the frame memory 22. In this case, data whose resolution has been switched to a lower resolution is stored. In addition, instead of the control of selecting data to be stored from the above when storing the data in the frame memory 22, the control may be performed such that the above selection is performed on the image data input from the image interface 20. .

In the sub-scanning delay control unit 23, the timing of reading the data of each color of Y, M, C, and Bk from the frame memory 22 is determined based on the interval between the developing units of each color (L mm in the illustrated system). It is controlled independently for each of the colors M, C, and Bk. Thereby, the position of the image of each color is corrected. For example, the position of the Bk image is corrected based on the drawing position by the Bk imaging unit at the lowermost stream as the paper reference. With respect to the data of each color of Y, M, and C, the position of each image is corrected based on the Bk data after the image position correction.

The sub-scanning delay controller 23 controls the timing of reading data from the frame memory 22 in accordance with the printing operation. That is, two-sided print mode for printing on both sides of paper / two-in-one mode for printing two-page images in the first half and second half of the same paper / side-sequential mode for sequentially printing images on paper / multiple copies for printing multiple copies The timing of reading data from the frame memory 22 is controlled in accordance with the operation of a print mode such as a memory retention mode for repeatedly outputting page images.

The sub-scanning delay control section 23 controls a read address for reading data from the frame memory 22 in accordance with an image editing mode. For example,
When the mode for rotating the image is set, the image is rotated by controlling the read address as described later.

The frame memory 22 has a capacity capable of storing data of each color of Y, M, C and Bk for drawing A3 size as standard. In other words, the standard has a capacity to store one plane of A3 size image drawing data, and can be expanded to a capacity of four sides. When the original is expanded to four sides, after printing on the first side of the sheet at the time of double-sided printing, the image data of the other original is printed up to three times before being inverted by the reversing unit and printed on the second side. Since the area can be memorized, print productivity can be improved.

Y, read from the frame memory 22,
If each of the data M, C, and Bk is data that has been subjected to variable-length code compression, the fixed-length data conversion unit (variable-length code expansion unit) 25 performs variable-length code expansion. The variable-length coded decompression process is a reverse conversion process of the variable-length coded compression process. As a result, the data of each color of Y, M, C, and Bk is inversely converted into data (= fixed-length compressed data; 48-bit data per 4 × 4 pixels) before being subjected to the variable-length coding compression processing. When each of the Y, M, C, and Bk data read from the frame memory 22 is data subjected to only fixed-length compression, conversion from a 16-bit data bus to 48 bits is performed. If each of the Y, M, C, and Bk data read from the frame memory 22 is data converted to low resolution, 16
The conversion from the data bus of 32 bits to 32 bits is performed. These conversion processes are shown in the left half of FIG. The data subjected to any of the above conversion processes is stored in an internal memory having a storage capacity of a plurality of lines. Variable length coding expansion processing, 16 to 48 bit conversion processing, 16t
The o32-bit conversion processing will be described later in detail.

The data of each color of Y, M, C, and Bk, which is variable-length coded and decompressed to 48 bits per 4 × 4 pixel, or read from the frame memory 22 and read out of the frame memory 22 at 48 bits
The correction unit 27 adjusts the position of the data converted into bits or the data read from the frame memory 22 and converted into 32 bits per 4 × 4 pixel in the main scanning direction and in the sub-scanning direction. Skew correction is performed, and further, fixed-length decompression processing (GBTC decoding) or padding processing is performed. The fixed length extension processing and the padding processing will be described later in detail.

The position adjustment in the main scanning direction is performed by controlling the timing of reading from the internal memory so that a reference position (center position) in the main scanning direction that differs depending on the paper size is adjusted to the center position of the transfer belt. Skew correction in the sub-scanning direction is performed by controlling the timing of reading from the internal memory so as to shift the line in units of a predetermined number of pixels determined by the degree of skew.
Here, the degree of skew is detected by superimposing test patterns of each color of Y, M, C, and Bk on the transfer belt and measuring a shift amount with respect to the Bk component. The measurement of the shift amount is performed using a sensor (not shown) provided on the downstream side of the Bk image forming unit. The fixed length decompression process is an inverse conversion process of the fixed length compression process. The padding process is a process of converting low-resolution image data into a normal resolution. With any one of these, the compressed data of Y, M, C, and Bk is expanded, or image data of normal resolution is reproduced from low-resolution image data, and sent to the tone reproduction unit 29. .

The tone reproduction section 29 corrects the tone distortion due to the γ characteristic in the electrophotographic process after the resolution conversion processing of 2400 dpi for each color of Y, M, C and Bk.
The correction is performed, and the corrected image data of each color is transferred to the corresponding LED driver of the print head control unit 30. Further, prior to the γ correction, the edge detection of the image is performed again.
The intensity modulation of the dot cycle is performed. Further, in the continuous tone portion, tone reproduction is prioritized, and intensity modulation in a two-dot cycle is performed. At this time, in the intensity modulation with a two-dot cycle, different screen angles are set for each color so that the granularity of the image is improved.

[2] Details of encoding process (compression):
Details of the fixed-length compression processing of the GBTC method, the variable-length coding compression processing that is executed as needed after the fixed-length compression of the GBTC method, and the resolution conversion processing of switching the resolution to low resolution are shown in FIGS. It will be described with reference to FIG. 4 is a block diagram mainly showing the encoding unit 21, FIG. 5 is a block diagram showing the fixed length compression unit 211 in FIG. 4, FIG. 6 is a block diagram showing the resolution conversion unit 218 in FIG. 4, and FIG. Attribute determination unit 2
12 and a block diagram mainly showing the variable length coded compression section 215, and FIG. 8 is an explanatory diagram showing the data structure of fixed length compression and variable length coded compression.

[2-1] Block generation: The block generation section 210 shown in FIG. 4 has 8 bits per pixel of Y, M, C,
From the raster data of the image of each color of Bk, block cutting is performed with 4 × 4 pixels as one block. Each cut-out block is composed of 8 × 4 × 4 = 128 bits and sent to the fixed-length compression unit 211 and the resolution conversion unit 218.

[2-2] Fixed length compression processing: The fixed length compression section 211 shown in FIG.
It is compressed by the C method and is converted into 48-bit compressed data. First, a maximum value (Lmax) and a minimum value (Lmin) are calculated from image data corresponding to 4 × 4 pixels.

Next, the internal dividing points P2 and P1 are calculated based on the calculated maximum value and minimum value. The internally dividing point P2 is an internally dividing point that internally divides the gradation between the maximum value and the minimum value into “3: 1”, and is calculated by the following equation (1).
It is a closer inner dividing point. The internal dividing point P1 is an internal dividing point that internally divides the gradation between the maximum value and the minimum value into “1: 3”, and is calculated by the following equation (2) and is closer to the minimum value Lmin. It is.

P1 = (Lmax × Lmin × 3) / 4 (1) P2 = (Lmax × 3 + Lmin) / 4 (1)

Next, the upper limit value Q4 is calculated by the following equation (3) based on the calculated inner dividing point P2, and the lower limit value Q1 is calculated by the following equation (4) based on the inner dividing point P1.

Q4 = (average value of image data of P2 or more)... (3) Q1 = (average value of image data of P1 or less)... (4)

Next, the average value index LA of the block is calculated by the following equation (5) based on the calculated upper limit value Q4 and lower limit value Q1, and similarly, the gradation width of the block is calculated based on the upper limit value Q4 and the lower limit value Q1. The index LD is calculated by the following equation (6).

(3) LA = (Q4 + Q1) / 2 (5) LD = (Q4-Q1) (6)

Next, the average value index LA and the gradation width index LD
And the data value of each pixel, the data value of each pixel is quantized according to the rules of the following equations (7) to (10). A code is assigned.

(7) LA + LD / 4 ≧ data value> LA → 10 (8) LA ≧ data value> LA−LD / 4 → 01 (7) 9) LA-LD / 4 ≧ data value → 00 (10)

By the above processing, 4 × 4 × 8 bits = 1
The 28-bit image data is compressed to a total of 48-bit fixed-length data consisting of an 8-bit LA, an 8-bit LD, and a 32-bit quantization code of 2 × 4 × 4 pixels. You.

[2-3] Variable length coding compression and attribute counting:
The fixed-length-compressed data is further subjected to variable-length coding and compression when the document size is A3. As shown in FIG.
The variable-length coding compression section 215 performs a compression process according to the determination result of the attribute determination section 212. In addition, attribute information is counted for each type by the attribute counting unit 216 for the image data input at the time of the pre-scan. Also, based on the result, the CPU 29 (FIG. 4) determines whether or not the total amount of compressed codes (codes that have been subjected to variable length coding and compression after fixed length compression) of the image data at the time of the main scan can be stored in the frame memory 22. If it is determined that the data cannot be stored, the memory over flag is set to 1. As a result, as will be described later, during the main scan, the data after the resolution conversion processing is selected and the frame memory 22 is selected.
Is stored in

[2-3-1] Attribute discrimination: As shown in FIG. 7, the attribute discriminating unit 212 discriminates the attribute of each block of 4 × 4 pixels based on the average index LA and the gradation width index LD. It is determined according to the rules shown in the section 212, and a 2-bit attribute code is generated for each block. (A) When LD is equal to or smaller than a predetermined gradation width reference value LDref and LA is equal to or smaller than a predetermined average first reference value LAref1, the gradation difference in the block is small and the density level is low (= 0) and the attribute information ATR is set to “00”. (B) When LD is equal to or smaller than the predetermined gradation width reference value LDref and LA is equal to or larger than the predetermined average second reference value LAref2, the gradation difference in the block is small and the density level is high (= around 255). ), The attribute information ATR is "0
1 ". (C) If LD is equal to or smaller than a predetermined gradation width reference value LDref and LA is larger than a predetermined average first reference value LAref1 and smaller than a predetermined average second reference value LAref2, the gradation in the block is used. This is the case where the difference is small and the density level is near the middle, and the attribute information ATR is “10”. (D) In cases other than the above, that is, when LD is larger than the predetermined gradation width reference value LDref, it is a case where the gradation difference in the block is large, and the attribute information ATR is set to “11”.

The three reference values LDre used for the above determination
f, LAref1, and LAref2 are values set in the internal registers by the CPU 29, and have the following meanings. By setting these three reference values to appropriate values, it is possible to adjust the compression ratio of variable-length coded compression in consideration of the balance with image quality. (A) Tone width reference value LDref: used to determine the magnitude of the tone difference. When LDref is set to a large value, “ATR = 1
1 "decreases, the number of blocks having no gradation difference increases, and the number of blocks having a gradation difference decreases. When LDref is set to a small value, "ATR = 11" increases, the number of blocks having no gradation difference decreases, and the number of blocks having a gradation difference increases. (B) Average first reference value LAref1: Used to determine a block having a low density level. When LAref1 is set to a large value, “ATR = 00” increases. For example, LAref1
Assuming that the density level of all pixels in the block is 5 or less, “ATR = 00” is set. (C) Average second reference value LAref2: used for determining a block having a high density level. When LAref2 is set to a small value, “ATR = 01” increases. For example, LAref2
Assuming that = 250, if the density level of all the pixels in the block is 250 or more, "ATR = 01" is set.

[2-3-2] Variable-length coding: The variable-length coding compression section 215 outputs data of each block of 4 × 4 pixels according to the value of the attribute information ATR input from the attribute determination section 212. Are converted according to the rules shown in the lower right column in FIG. That is, (a) the block of “ATR = 00” is converted into only 2-bit attribute information “00”. (B) The block of “ATR = 01” is converted into only 2-bit attribute information “01”. (C) The block of “ATR = 10” is converted into 10-bit data including 2-bit attribute information “10” and 8-bit average index LA following the attribute information “10”. (D) The block of “ATR = 11” is converted into 50-bit data consisting of 2-bit attribute information “11” and 48-bit fixed-length compressed data following the attribute information “11”.

According to the value of the attribute information ATR, 10 bits consisting of 2-bit attribute information / 2-bit attribute information / 2-bit attribute information and 8-bit average value index LA from 48-bit fixed-length compressed data The variable-length compressed data converted into either 50-bit data consisting of data / 2-bit attribute information and 48-bit fixed-length compressed data is stored in an internal register as shown in the upper right column in FIG. ,
The data is serially converted and output so as to be compatible with a 16-bit data bus.

[2-3-3] When not Variable-Length Coded: After variable-length coded compression, the compressed data has a variable length, and the degree of compression can be increased. However,
Specifying an arbitrary image position (random address control in a compressed image) at the time of storage in the memory, cannot perform reading, and performs editing processing such as image rotation by address control when reading from the frame memory 22. Can not. Therefore, when an image editing process is required, it is necessary to select fixed-length compression that enables random address control with compressed data without performing variable-length coding. For example, when rotation processing is not required (for example, A3 size output), variable length coding is performed to increase the degree of compression, and rotation processing is required (for example, A4 size output).
In this case, fixed-length compression is performed without performing variable-length coding, thereby enabling random address control with compressed data. In addition, instead of selecting according to the output size, depending on whether or not the editing process is necessary, at an arbitrary size,
A setting may be made so that variable-length coding is not performed. Note that the compression ratio of only the fixed-length compression processing is smaller than that of the variable-length coding compression processing. However, since the A4 size image data amount is smaller than the A3 size image data amount,
This is the amount of data that can be sufficiently stored in the frame memory.
FIG. 8 shows the data structure of the compression code by the fixed-length compression processing and the data structure of the compression code by the variable-length coding compression processing.
Shown in

[2-3-4] Attribute counting (at the time of pre-scan): Image data input from the image interface 20 is image data at the time of pre-scan (pre-scan is generally executed at a lower resolution than main scan) In the case of (1), the attribute counter 216 counts the number of each piece of the attribute information for each type. This is because the code amount after variable length compression is uniquely determined according to the type of attribute information (ATR = 00 → 2 bits, ATR = 01 → 2 bits, ATR = 10 → 10 bits, ATR = 11 → 50 bits) ) Is used to determine the total code amount after variable-length coded compression. The counting result is sent to the CPU 29, and based on this, the CPU 29 takes into account the difference in resolution between the pre-scan and the main scan, and compresses the image data at the time of the main scan (variable-length compression after fixed-length compression. Code) is obtained. Further, it is determined whether or not the obtained total code amount is an amount that can be stored in the frame memory 22. If the total amount cannot be stored, the memory over flag is set to 1.

[2-4] Resolution Conversion Processing: FIG. 6 shows the processing of the resolution conversion unit 218. As shown in the figure, a block of 4 × 4 pixels is divided into four small areas each having 2 × 2 pixels, and for each small area, an average value of data values (four 8-bit data per small area) Is calculated. Further, the calculated average value (8-bit data) is used as the data value of the small area, and is output to the 16-bit data bus. As a result of this resolution conversion, the data amount per 4 × 4 pixels (one block) is reduced from 128 bits to 32 bits.

[3] Memory control: Which one of data subjected to fixed-length compression only, data subjected to variable-length coding compression after fixed-length compression, and data subjected to resolution conversion, A description will be given of write control for selecting whether or not to write to the image memory, and control for rotating an image by controlling an address when reading from the frame memory 22.

[3-1] Memory write control (switch control):
The fixed-length compressed data of 48 bits per 4 × 4 pixel is further subjected to variable-length coding compression and written to the frame memory 22, or written to the frame memory 22 without variable-length coding compression, or after resolution conversion processing. The selection of whether to write data to the frame memory 22 is made according to a fixed length / variable length switching signal CODESEL set in an internal register by the CPU 29 and a memory over flag controlled by the CPU 29, as shown in FIG. (A) When the memory over flag is set to 1, the data after the resolution conversion processing is stored in the frame memory 22
Write to. (B) When the memory over flag is not set to 1 and “CODESEL = 0”, the fixed-length compressed data is further subjected to variable-length code compression and written into the frame memory 22. (C) When the memory over flag is not set to 1 and “CODESEL = 1”, the data after the fixed length compression is written to the frame memory 22 without performing the variable length coding compression. The control is performed as follows.

The above CODESEL is determined according to the document size. That is, when the document size is A3 size, “CODESEL = 0” is set. Also, A4
If the size is smaller than the size, "CODESEL = 1" is set.

According to the above switching control, when the original is smaller than A4 size, fixed-length compressed data of 48 bits per 4 × 4 pixels is stored in the frame memory 22. Therefore, by performing address control for exchanging the row address and the column address when reading from the frame memory 22,
The image can be rotated 90 degrees. That is, the editing process can be performed in a compressed state. If the original is A3 size, the fixed length compressed data of 48 bits per 4 × 4 pixel is further compressed into variable length code on condition that the code amount can be stored in the frame memory 22. It is stored in the frame memory 22. Therefore, the data amount can be further reduced, a relatively small-capacity and inexpensive memory can be used as the frame memory 22, and the cost can be reduced. For example, in the system of the embodiment,
By configuring the frame memory 22 using two 64 MB DRAMs per color, an A4 size 600 dp
The image data amount of i can be stored. If the original is A3
In the case of the size, if it is determined that the total code amount after the variable length coded compression exceeds the amount that can be stored in the frame memory 22, the data converted to low resolution and the data amount is reduced Stored in the memory 22. For this reason, the data cannot be stored during the storage in the frame memory 22. As a result, problems such as the inability to form an image and the rescanning of an image are prevented. Further, since the image for which the total code amount is determined to exceed the storage capacity of the frame memory 22 is a photographic image (continuous tone image), even when converted to a low resolution, the deterioration of the image quality is negligible.

[3-2] Memory Read Control (Rotation of Image): FIG. 12 is an explanatory diagram showing data stored at each address in the frame memory 22. As shown,
Each address in the frame memory 22 is specified by a row address (row address) and a column address (column address). As described above, the 48-bit fixed-length compressed data of one block is an 8-bit average value index LA.
, An 8-bit gradation width index LD, and 2-bit code information P0 to Pf each having a total of 32 bits.

FIG. 13 shows the principle of rotating an image by exchanging a row address (row address) and a column address (column address) when reading from the frame memory 22 and reading the column address in reverse order. As shown in the upper part of FIG. 13, in the normal state where the image is not rotated, reading is performed in the order of the row address from the left end of the row and in the order of the row from the top row. On the other hand, at the time of 90-degree rotation in which the image is rotated by 90 degrees, as shown in the lower part of FIG. 13, reading is performed in the reverse order of the column address from the lower end of the column and in the normal order of the column from the first column. Further, at this time, the rows and columns are exchanged for the 32-bit code information P0 to Pf of the 48-bit code data.
That is, Pc, P8, P4, P0, Pd, P9, P5
P1, Pe, Pa, P6, P2, Pf, Pb, P7, P
In the order of 3, lower 32 bits of data in each address are read.

[4] Decoding: FIG. 9 shows the frame memory 22
FIG. 10 is a block diagram showing a process of decoding data read out from a memory according to its encoding method (fixed length compression, fixed length compression & variable length coding compression, resolution conversion).
FIG. 11 is an explanatory view showing the principle of decoding a code which has been subjected to fixed-length compression or fixed-length compression and variable-length code compression into code data of 48 bits per 4 × 4 pixels. The upper part of FIG. 11 shows a code of 48 bits per 4 × 4 pixels. FIG. 11 is a block diagram showing a process of performing GBTC decoding (expansion from fixed-length compression) of data, and FIG.
It is a block diagram showing processing which inflates low-resolution data of 32 bits per × 4 pixels to 128 bits.

From the frame memory 22, Y, M, C,
Data of each color of Bk is sequentially read out at the timing determined by the interval (Lmm in this example) between the image forming units of each color, and is first written into the input buffer.

[4-1] Variable-length code expansion (fixed-length data conversion): When the fixed-length / variable-length switching signal (CODESEL) given by the CPU 29 is 0 and the low-resolution data selection signal is 0, This is the case where the data which has been subjected to the variable length code compression after the fixed length compression is stored in the frame memory 22. In this example, this is the case of an A3 size document. In this case, CODESEL shown in the upper section of FIG.
The process when = 0 is selected as the decoding process. That is, the output of the variable length code conversion unit 255 in FIG. 9 is selected by the selection unit.

More specifically, 2 read from the input buffer
The following process is performed in which 48-bit data (new fixed-length compressed data) is generated and written to the output buffer according to an algorithm corresponding to the value of the bit attribute information ATR. (A) When the attribute information ATR is “00”, the average value index LA = 00h, the gradation width index LD = 0h, and the code information = 00
48h data of 00h is generated. (B) When the attribute information ATR is “01”, the average value index LA = ffh, the gradation width index LD = 0h, and the code information = 00.
48h data of 00h is generated. (C) When the attribute information ATR is "10", the 8-bit data F1h following the 2-bit attribute information is the average value index L
A, this “LA = F1h” and the gradation width index LD =
48-bit data of 0h and code information = 0000h is generated. (D) When the attribute information ATR is “11”, the 48-bit data following the 2-bit attribute information is generated as new fixed-length compressed data.

Next, the data decoded to the code data (GBTC compressed data) of 48 bits per 4 × 4 pixel as described above is the average value index L as shown in the upper part of FIG.
A, gradation width index LD, and 2 bits per pixel (4 × 4
The inverse transform is performed based on the quantization code of (32 bits per pixel). That is, for each pixel, one of the following (11) to (14) is selected according to the value of the 2-bit quantization code of the pixel, and the LA and LD of the pixel are substituted into the selected formula. The value is selected and output as the data value of the pixel.

(A) LA + LD / 2 (11) (b) LA + LD / 6 (12) (c) LA-LD / 6 (13) (d) LA- LD / 2 (14) The above processing is an inverse transformation of the fixed-length compression processing (item [2-2]).

[4-2] Fixed-length decoding (GBTC inverse conversion): C
Fixed length / variable length switching signal (CO
When DESEL) is 1, the code compressed only by the fixed length compression is stored in the frame memory 22, and in this example, it is the case of an original of A4 size or less.
In this case, first, the CODESE shown in the lower column of FIG.
The process when L = 1 is selected. In other words, the output of the “16 to 48 conversion unit” that cuts out 48-bit code data (data of one block of 4 × 4 pixels) from the stream input from the 16-bit data bus is selected by the selection unit.

Next, 48 bits of coded data (GBTC compressed data) per 4.times.4 pixel obtained by the above processing.
Are fixed-length decoded as in the case of the above-described 48-bit data after variable-length code expansion. That is, the average value index LA, the gradation width index LD, and 2 bits (4 ×
1 based on a quantization code of 32 bits per 4 pixels).
1, the image data is reproduced. In other words, for each pixel, one of the above (11) to (14) is selected according to the value of the 2-bit quantization code of the pixel, and the LA and LD of the pixel are substituted into the selected mathematical expression. A process of selecting and outputting the value obtained as the data value of the pixel is executed.

[4-3] Inflated (Reverse resolution conversion): CPU2
When the low-resolution data selection signal given from 9 is 1,
In other words, when the memory over flag is set to 1, it means that the image data converted to a low resolution is stored in the frame memory 22.
This is the case of a photo manuscript of a size. In this case, first
32 from the stream input from the 6-bit data bus
The output of the “16 to 32 conversion unit” that cuts out low-resolution image data of bits (low-resolution image data of one block of 4 × 4 pixels) is selected by the selection unit.

Next, the process of inflating the low-resolution image data of 32 bits per 4 × 4 pixel obtained by the above process to the image data of 128 bits is performed by the inflating section 273 as shown in the lower part of FIG. Be executed. That is, the low-resolution data is obtained by dividing 4 × 4 pixels into four small areas of 2 × 2 pixels and calculating an average value for each small area to obtain image data of each small area. This average value is calculated as
The processing is performed so as to be adopted as image data of each pixel constituting the × 2 pixel small area. In the example of FIG.
The average value of the × 2 pixels is “A0 (8 bits)”, and the “A0 (8 bits)” is the data value of the four pixels that constitute the 2 × 2 pixels.

Either the data value thus obtained or the data value obtained by GBTC decoding as described above is selected and output by the selection unit. This choice is
This is performed according to the value of the low resolution data selection signal.

[5] Processing Procedure: FIGS. 14 and 15 are flowcharts collectively showing the procedure of each processing described above. For example, if the document size is A4 or smaller (S0
1: NO), the code compressed only by the GBTC compression is written to the frame memory 22 (S43). When the document size is larger than A4, the attribute count is counted for each type at a low resolution (S03 to S09), and the total code amount at a normal resolution can be stored in the storage capacity of the frame memory 22 based on the count result. It is determined whether or not the memory is over, and if the storage is not possible, the memory over flag is set to 1 (S11).

When the memory over flag is set to 1 (S21: YES), the image data read at a low resolution is stored in the frame memory 22 (S3).
5). On the other hand, when the memory over flag is 0 (S2
1: NO), and after the fixed length compression (GBTC compression), the data further subjected to variable length coding and compression is stored in the frame memory 22 (S25).

[0059]

According to the present invention, it is possible to determine whether or not all of the compression-encoded image data can be stored in the memory before storing the data in the memory.
It is possible to switch to a compression method or the like that further reduces the data amount, and it is possible to store all necessary image data in a memory and proceed with the processing. Further, since the image printing does not stop halfway due to the memory over and it is not possible to print the entire image, the user does not feel uncomfortable. Further, an image having a small compression effect due to variable length coding that causes memory over is a photographic image having many blocks having a large gradation difference, and in the present invention, the resolution is reduced only in that case. Let it. Therefore, the resolution of a character image having few blocks having a large gradation difference does not decrease. Therefore, the character image can maintain the resolution, and the photographic image can maintain the gradation, and the image quality can be maintained. On the other hand, by counting the attribute determination result at the time of variable length coding, it is possible to easily know whether or not the memory is over, and it is easy to simplify the circuit for determining the memory over.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration example of the present invention.

FIG. 2 is a system configuration diagram of an image forming apparatus in which the image processing apparatus according to the embodiment is mounted.

FIG. 3 is a block diagram illustrating the image processing apparatus according to the embodiment;

FIG. 4 is a block diagram mainly showing an encoding unit 21 of FIG. 3;

FIG. 5 is a block diagram showing a fixed length compression unit 211 of FIG. 4;

FIG. 6 is a block diagram showing a resolution converter 218 in FIG. 4;

FIG. 7 is a block diagram mainly illustrating an attribute determining unit 212 and a variable-length coding / compressing unit 215 of FIG. 4;

FIG. 8 is an explanatory diagram of fixed-length compression and variable-length coded compression.

FIG. 9 shows a fixed-length data converter 25 and 600 dpi of FIG.
FIG. 3 is a block diagram illustrating a skew correction unit 27.

FIG. 10 is an explanatory diagram showing the principle of processing in the variable-length encoding / decompression unit 255 of FIG. 9;

FIG. 11 is a block diagram showing an extension part 271 and a padding part 273 from the fixed length compression in FIG. 9;

FIG. 12 is an explanatory diagram showing a data configuration stored at each address in a frame memory 22.

FIG. 13 is an explanatory diagram showing a reading order from the frame memory 22 at the time of 0-degree rotation (normal) and at the time of 90-degree rotation.

FIG. 14 is a part of a flowchart showing a control procedure of the apparatus according to the embodiment;

FIG. 15 is the remaining part of the flowchart showing the control procedure of the apparatus according to the embodiment.

FIG. 16 is a first example showing a case of memory over in a conventional image processing apparatus.

FIG. 17 is a second example showing a case of memory over in a conventional image processing apparatus.

[Explanation of symbols]

 211 Fixed length compression unit 212 Attribute discrimination unit 215 Variable length coding compression unit 216 Attribute counting unit 218 Resolution conversion unit 25 Fixed length data conversion unit (Variable length coding expansion unit)

Claims (5)

[Claims] 1. An image processing apparatus which stores image data in a memory, reads out the image data stored in the memory, and generates print data, comprising: Attribute determining means for determining, for each block, attribute information indicating the image data; compression encoding means for converting image data of each block into a code having a bit number determined by the type of attribute information; and counting attribute information for each type. Attribute counting means; code amount determining means for determining whether or not the total code amount is an amount that can be stored in the memory based on the counting result for each type of attribute information; resolution switching means for switching the resolution of image data; If the code amount is an amount that can be stored in the memory, the code converted by the compression encoding unit is stored in the memory, and if the total code amount is an amount that cannot be stored in the memory, An image processing apparatus comprising: a selection storage unit that stores, in a memory, image data or a compression code thereof switched to a low resolution by the resolution switching unit. 2. An image processing apparatus for storing image data in a memory, reading out the image data stored in the memory and generating data for output, comprising: an image of a block including a predetermined number of pixels cut out from the image data Attribute determining means for determining attribute information indicating characteristics for each block; first compression encoding means for converting image data of each block into a code having a bit number determined by the type of attribute information; first compression code Second encoding means for encoding image data by a compression method different from the encoding means, attribute counting means for counting attribute information for each type, and total code based on a counting result for each type of attribute information. Code amount determining means for determining whether the amount is an amount storable in the memory; and, if the total code amount is an amount storable in the memory, the code converted by the first compression encoding means is recorded. If the total code amount is an amount that cannot be stored in the memory, the code encoded by the first compression encoding unit is further stored in the memory after encoding by the first compression encoding unit. An image processing apparatus comprising: 3. An image processing apparatus for storing image data in a memory, reading out the image data stored in the memory and generating data for output, comprising: an image of a block including a predetermined number of pixels cut out from the image data Attribute determining means for determining attribute information indicating characteristics for each block; compression encoding means for converting image data of each block into a code of a bit number determined by the type of attribute information; counting attribute information for each type Attribute counting means, code amount determining means for determining whether or not the total code amount is an amount that can be stored in the memory based on the counting result for each type of attribute information, and a method for processing image data according to the determination result. An image processing apparatus, comprising: switching means for switching. 4. An attribute determining means according to claim 1, wherein said attribute determining means determines each attribute based on low-resolution image data when determining attribute information for counting by said attribute counting means. Determining the attribute information of the image processing apparatus. 5. The attribute determining means according to claim 1, wherein the attribute determining means determines the attribute information for counting by the attribute counting means based on the image data of the prescan. Determining the attribute information of the image processing apparatus.