JP2015055944A - Image processor and image rotation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processor capable of reducing a deterioration in work memory use efficiency and a memory band load during main memory access in the case of performing rotation processing of image data of multiple gradation in a rectangular block unit of squares.SOLUTION: The image processor includes a first memory 114 for storing image data, and a rotation processing part 119 for performing rotation processing by using image data transfer parts for reading the image data in a predetermined rectangular pixel unit and performing writing back, a second memory for storing a rectangular image data area, and a rotation part for performing control of writing and reading the image data read in the rectangular pixel unit. In the rotation processing part 119, one of the image data transfer parts performs burst transfer of data of the rectangular pixel with a length n in a continuous address direction of the first memory 114 in accordance with a bit gradation number n (n is a positive integer) of the image data, and stores n rectangular blocks composed of the rectangular pixels in the second memory to perform rotation processing.

Description

  The present invention relates to an image rotation method in an image processing apparatus, and more particularly to an image processing apparatus and an image rotation method for dividing an image into rectangular blocks and rotating the image in rectangular blocks.

  In an image processing apparatus typified by a digital multi-function peripheral, a digitized image is taken into the apparatus, and various image processing is performed according to the use conditions of the apparatus and user settings, and an image is output. For example, when print processing is performed by a digital multi-function peripheral, image rotation processing is performed inside the apparatus in order to align the paper orientation of the print data (eg, A4 paper landscape) and the output paper orientation (eg, A4 paper portrait). There is a need. In general, image rotation processing can be realized by switching the vertical and horizontal pixel data of an image in accordance with the rotation direction, that is, storing image data in a certain memory and performing a predetermined address operation.

  Here, in recent years, with the demand for higher image quality of printed output, the image resolution handled by the digital multi-function peripheral has increased, and the number of pixels handled by the apparatus has increased accordingly. For example, when the image resolution is doubled from 600 dpi (dot per inch) to 1200 dpi, the number of pixels for one page of the image is quadrupled. Further, the memory capacity necessary to hold one pixel is increased by the number of color gradations per pixel for the pixels constituting the image. Therefore, in order to reduce the cost related to the memory used for the image rotation process, it is desirable to perform the rotation process using a memory for a smaller image unit instead of one page of the image. In order to realize this, Patent Document 1 discloses a method of dividing an image into predetermined rectangular block units and performing a rotation process using a memory area for image data constituting the rectangular block.

  According to Patent Literature 1, when the memory capacity required for rotating image data for one page is insufficient, the image rotation processing can be performed in units of rectangular blocks. Further, even when the pixels constituting the image data are expressed by a color gradation of N bits (N is a power of 2) per pixel, the width of the rectangular block cut out from the image data according to the color gradation, and The height can be set appropriately, and image rotation processing can be performed in block units.

JP 2005-178033 A

  Here, it is assumed that image data to be subjected to rotation processing is divided into square rectangular blocks that are image data represented by N-bit color gradation per pixel. The reason why the rectangular block is a square is that the vertical and horizontal pixel replacement numbers of the images to be replaced at the time of rotation are the same, and the rotation processing setting is facilitated.

  A memory area (hereinafter referred to as a work memory) used for image rotation processing corresponding to a square rectangular block is 32 words × 32 bits.

  In the case of 1-bit image data per pixel (N = 1), the maximum square image size that the work memory can hold is 32 (vertical) × 32 (horizontal) pixels, and all 32 words × 32 bits of the work memory. Rotate using.

  Next, in the case of 2-bit image data (N = 2) per pixel, the maximum square image size that can be held in the work memory is 16 (vertical) × 16 (horizontal) pixels. The reason why the number of vertical and horizontal pixels of a square rectangular image is 16 is because there are 2 bits of data per pixel. For example, if the 2 bits of data are arranged in the horizontal direction, = 16 × 2) and the work memory is full. At this time, rotation processing is performed using 16 words × 32 bits which is half of the work memory.

  Similarly, in the case of 4-bit image data (N = 4) per pixel, the maximum square pixel size that can be held in the work memory is 8 (vertical) × 8 (horizontal) pixels. Rotation processing is performed using a quarter of 8 words × 32 bits.

  As described above, the use efficiency of the work memory used when N-bit image data per pixel is rotated in units of square rectangular blocks is 1 / N, and multi-value gradation image data in which N is 2 or more. This causes a problem that the use efficiency of the work memory is lowered during rotation.

  In addition, when the rotation is performed in units of rectangular blocks, the number of rotation pixels that can be rotated at a time is 1 / N, and therefore the number of times of processing relating to rectangular block cutout is N times. For this reason, in the system configuration in which the image data held in the main memory is acquired as a rectangular block and rotated, the number of accesses to the main memory increases, and the memory bandwidth load increases accordingly.

  The present invention has been made in view of the above-described problems. The purpose of this is to reduce the work memory usage efficiency and reduce the memory bandwidth load when accessing the main memory when rotating multiple-tone image data in units of square rectangular blocks. It is to provide a processing device.

A first memory (114) for storing image data;
For the image data (1101) on the first memory (114),
An image data transfer unit (601, 604) for reading the image data (1101) in units of a predetermined rectangular pixel and writing back to the first memory;
A second memory (603) for storing a rectangular image data area;
Using the image data transfer unit (601, 604), the rotation unit (602) performs writing of image data read out in units of the rectangular pixels and control of reading out to the second memory (603). When,
An image rotation unit (119) for performing rotation processing using
With
The image rotation unit (119), according to the bit gradation number n (n is a positive integer) of the image data (1101),
One of the image data transfer units (601, 604) burst-transfers the rectangular pixel data in a length n in the continuous address direction (1111 to 1115) of the first memory (114). A second memory (603) stores n rectangular blocks made up of the rectangular pixels and performs rotation processing.
It is characterized by that.

  According to the present invention, the image rotation means performs rotation processing using the entire work memory, and the use efficiency of the work memory is not lowered. Also, burst reading is possible when reading M rectangular blocks, and the read access efficiency to the main memory can be improved and access overhead can be reduced as compared to reading rectangular blocks in units of one.

1 is a block diagram illustrating an overall configuration of an image processing apparatus according to the present invention. 4 is a flowchart illustrating processing in which print data from a host computer is encoded and stored by the image processing apparatus according to the present invention. 4 is a flowchart illustrating a process in which the image processing apparatus according to the present invention decodes encoded data and prints it out. It is the figure which showed typically the pixel which comprises intermediate image data in this invention. It is a schematic diagram which shows the structure of the image process part in this invention. It is a schematic diagram which shows the structure of the rotation process part in this invention. It is a schematic diagram which shows the structure of the work memory which the rotation process part has in this invention. It is a sequence diagram which shows the mutual processing relationship between the blocks which concern on a rotation process in this invention. It is a flowchart which shows the process which a rotation process part performs in this invention. It is a flowchart which shows the process which the input data transfer part which a rotation process part has in this invention performs. It is a schematic diagram which shows the data read-out process before the rotation process which a rotation process part performs in this invention. It is a flowchart which shows the process which the rotation part which a rotation part process part has in this invention performs. It is a schematic diagram which shows the data writing process after the rotation process which a rotation process part performs in this invention.

  The best mode for carrying out the present invention will be described below with reference to the drawings.

(First embodiment)
The first embodiment of the present invention will be described below.

  FIG. 1 is a block diagram showing the overall configuration of the image processing system in the present embodiment.

  The image processing system shown in the figure is largely composed of a host computer (101) and an image processing apparatus (102).

  The host computer (101) is a computer such as a general PC (Personal Computer) or WS (Work Station), and images and documents created on these computers are input to the image device (102) as PDL data. The

  The image processing apparatus (102) receives data output from the host computer (101). That is, it means that a network capable of data communication is formed between the image processing apparatus (102) and the host computer (101), but the configuration of this network is not particularly limited.

  The image processing apparatus (102) performs various types of image processing based on the data received from the host computer (101), and outputs the image-processed data to the print engine unit (123). Details of processing performed by the image processing apparatus (102) will be described later. The print engine unit (123) performs a printing process on a storage medium such as paper based on the image processed data output from the image processing apparatus (102).

  In this embodiment, the host computer (101) is used to input data to the image processing apparatus (102). However, other apparatuses may be used. For example, the data may be output from a scanner input unit (not shown). The data may be input to the image processing apparatus (102).

  Next, the configuration of the image processing apparatus (102) will be described.

  The controller unit of the image processing apparatus includes a host I / F unit (111), a PDL processing unit (112), a CPU (113), a RAM (114), a ROM (115), and an encoding unit (116). A decoding unit (117), an image processing unit (118), a rotation processing unit (119), a storage device control unit (120), a storage device (121), and an engine I / F unit (122). And a print engine unit (123).

  The host I / F unit (111) functions as an interface for receiving PDL data output from the host computer (102). The host I / F unit (111) differs depending on the network connecting the image processing apparatus (102) and the host computer (101), and is configured by, for example, an Ethernet (registered trademark) or a serial interface.

  The PDL processing unit (112) performs processing for expanding the PDL data received by the host I / F unit (111).

  The CPU (113) controls the entire image processing apparatus (102) using programs and data stored in the RAM (114) and ROM (115), and will be described later performed by the image processing apparatus (102). Execute each process.

  A RAM (114) is a system memory of the image processing apparatus (102), and temporarily stores data received from the host computer (101) via the host I / F unit (111), as well as a CPU (113). ) Is a work area used when executing various processes.

  The ROM (115) stores programs and data executed when the CPU (113) controls the entire image processing apparatus (102), setting data of the image processing apparatus (102), and the like.

  The image processing unit (118) performs various types of image processing such as color space conversion, filter, and halftone processing on the image composed of the data output from the host computer (101).

  The rotation processing unit (119) performs rotation processing on the image data according to the present invention in units of rectangular blocks. Details of the image processing unit (119) will be described later.

  The encoding unit (116) compresses the image developed in the PDL processing unit (112) using a compression method such as JPEG for multi-valued image data and JBIG for binary image data. Perform the encoding process.

  The decoding unit (117) performs a decoding process based on the compression method encoded by the encoding unit (116), and generates an image before encoding.

  The storage device control unit (120) performs control for recording the image data handled by the image processing device (102) in the storage device (121).

  The storage device (121) is a large-capacity storage medium including a hard disk and a semiconductor memory, and stores image data received by the image processing device (102) from the host computer (101) and intermediate image data generated during processing.

  The engine I / F unit (122) performs a series of processes for outputting data processed by the image processing apparatus (102) to the print engine unit (123).

  The above-described units are connected to each other by an internal bus (131) of the image processing apparatus (102). Data transfer between the respective units is performed using a data transfer unit (not shown) formed by a DMAC (Direct Memory Access Controller) provided between each unit and the internal bus (131).

  Next, processing performed by the image processing apparatus (102) when data is output from the host computer (101) to the image processing apparatus (102) will be described using the flowchart of FIG.

  The actual state of the processing shown in the flowchart of FIG. 2 is a program stored in the ROM (115), and the CPU receives the PDL data output from the host computer (101) via the host I / F unit (111). (113) starts upon detection. Note that the actual flow charts described below are programs stored in the ROM (115) unless otherwise specified.

  First, in step (S201), the CPU (113) temporarily stores the received PDL data in the RAM (114). Next, the process proceeds to step (S202), and after the CPU (113) expands the received PDL data to DL (Display List), the process proceeds to step (S203) and is temporarily stored in the RAM (114). . In step (S204), the PDL processing unit (112) performs image expansion processing based on the stored DL, and generates intermediate image data that is image data to be handled in the image processing apparatus (102). In step (S205), the encoding unit (116) encodes the intermediate image data and performs data compression. In step (S206), the encoded data is stored in the storage device (121) via the storage device control unit (120). Through the above processing, one page or more of the encoded and compressed image data is stored in the storage device (121) in accordance with the content of the image data transmitted from the host computer (101).

  Next, a process of reading the encoded and compressed intermediate image data stored in the storage device (121) and outputting it to the print engine unit (123) will be described using the flowchart shown in FIG.

  In the processing shown in the flowchart of FIG. 3, the CPU (113) instructs the decoding unit (117) to issue a print start command after receiving the encoded intermediate image data storage processing end notification from the storage device control unit (120). To start.

  First, in step (S301), operation settings for the image processing unit (118) and the rotation processing unit (119) are performed. Specifically, the CPU (113) writes necessary setting values into the image processing module (not shown) included in the image processing unit (118) and the registers included in the rotation processing unit (119).

  Next, the process proceeds to step (S302), and the CPU (113) instructs to output the encoded intermediate image data stored in the storage device (121) to the RAM (114). At the same time, the CPU (113) instructs the decoding unit (117) to read the encoded intermediate image data from the RAM (114). The decoding unit (117) sequentially decodes the read encoded intermediate image data and outputs the decoded intermediate image data to the RAM (114). Note that data reading and output performed by the decryption unit (117) are performed using a data transfer unit (not shown) included in the decryption unit (117).

  Next, the process proceeds to step (S303), and the image processing unit (118) reads intermediate image data from the RAM (114) by using a data transfer unit (not shown) provided therein, and performs image processing for print output. The data is converted into print data and output to the RAM (114) again. The print data is data after various types of image processing (not shown) included in the image processing unit (118), and is data subjected to halftone processing. The image data is composed of pixel data having a lower gradation than the intermediate image data in accordance with the halftone processing setting.

  Next, the process proceeds to step (S304), and the rotation processing unit (119) reads the print data from the RAM (114) using the internal data transfer unit (not shown) and performs rotation processing again. Output to the RAM (114). Details of processing performed by the rotation processing unit (119) will be described later.

Then, the process proceeds to step (S305), the print data is output to the RAM (114) using a data transfer unit (not shown), and then the engine I / F unit (122) includes the data (not shown). The data is transferred to the print engine unit (123) using the data transfer unit.
Through the above processes, data received from the host computer (101) can be output to the print engine unit (123).

  In the above description, the spool data is held in the storage device (121). However, apart from this, all the data is held only in the RAM (114), and the various processes described above are performed, and the print engine unit ( 122).

  FIG. 4 is a diagram schematically showing pixels constituting the intermediate image data.

  The data of each pixel constituting the intermediate image data is constituted by image information indicating the bit gradation for each color component. In this embodiment, it is assumed that the gradation is 8 bits with a monochrome pixel. Note that the data length of the image information and the color space information are not limited to this, and may be, for example, a color pixel having a gradation of 8 bits in an R, G, B color space.

  Next, the block configuration of the image processing unit (118) will be described with reference to FIG.

  The image processing unit (118) is roughly composed of a block for performing data transfer and a block for performing image processing. In this embodiment, an input data transfer unit (501) and an output data transfer unit (502) are used as data transfer blocks, and image processing blocks 1 to 4 (503 to 506) are used as image processing blocks. Shall be. Here, the image processing block constituting the image processing unit (118) includes at least one block for performing halftone processing, and in this embodiment, the image processing block 3 (505) corresponds to this. In the other image processing blocks in this embodiment, the image processing block 1 (503) is background processing, the image processing block 2 (504) is color conversion processing, and the image processing block 4 (506) is trimming processing. Note that the processing content and the number of stages of the image processing block are examples, and may be increased or decreased depending on the image processing to be realized.

  The input data transfer unit (501) is connected to the internal bus (131), and reads the intermediate image data stored in advance in the RAM (114) which is also connected to the internal bus (131) into the image processing unit (118). .

  The settings required for the operation of the input data transfer unit (501) include the start address address of the image information constituting the intermediate image data existing on the RAM (114), the input data size, etc., and the CPU (113) performs image processing. This is set in a register (not shown) of the section (118).

  The output data transfer unit (502) is connected to the internal bus (131), and transfers the image data processed by the image processing unit (118) to the RAM (114) connected to the internal bus (131). . The settings necessary for the operation of the output data transfer unit (502) include the start address address and the output data size for writing the image processed print data on the RAM (114). This is set in a register (not shown) of the processing unit (118).

  The image processing blocks 1 to 4 (503 to 506) perform image processing unique to each image processing block on the pixel data output from the input data transfer unit (501) in a pipeline manner, and print the processing results. Data is output to the output data transfer unit (502). In this embodiment, the halftone processing output setting of the image processing block 3 (505) that performs halftone processing is set to 2 bits, and the 8-bit gradation intermediate image data acquired from the RAM (114) has a 2-bit gradation. Converted to print data.

  Next, the configuration of the rotation processing unit (119) will be described using FIG. 6 and FIG.

  FIG. 6 is a block diagram of the rotation processing unit (119).

  The input data transfer unit (601) is connected to the internal bus (131), reads print data stored in advance in the RAM (114) into the image processing unit (119) in units of a predetermined rectangular block, and stores the rectangular block pixels. It outputs to a rotation part (603) per predetermined line unit.

  The rotation unit (602) transfers the pixels constituting the rectangular block received from the input data transfer unit (601) to the work memory (603) in a predetermined line unit. In addition, the rotation unit (602) performs rotation processing by transferring the rectangular block data necessary for the rotation processing to the work memory (603) and then reading the data from the work memory (603) by a predetermined address scanning. The read data is transferred to the output data transfer unit (602).

  FIG. 7 shows a block diagram of the work memory (603).

  The work memory (603) is a memory area used for image rotation processing in units of rectangular blocks. In this embodiment, the capacity of the work memory (603) is 32 words × 32 bits. Assuming that the work memory (603) is a square bit of 32 bits × 32 bits, the rotation unit (602) can write in the word direction (direction indicated by 610) and arbitrarily in the bit direction (direction indicated by 611). It is assumed that data can be read with the data length. It should be noted that the work memory (603) may be mounted in either SRAM (Static RAM) or FF (Flip Flop) as long as the access control is possible.

  The output data transfer unit (604) is connected to the internal bus (131), receives the rotated data rotated using the work memory (603) from the rotation unit (602), and is connected to the internal bus (131). The data is transferred to the RAM (114).

  In addition to the above, the rotation processing unit (119) has a register (not shown). The register setting values include the start address address of the image information of the intermediate image data existing on the RAM (114), the address address for writing the rotated data, the main scan image size, and the sub scan for data reading and writing. There are image size and rectangular cutout size.

  Furthermore, there are registers for setting the rotation angle and setting the bit gradation of the data to be processed. For example, 90, 180, and 270 degrees can be set clockwise, and in this embodiment, the rotation angle is set to 90 degrees. The setting value of the bit gradation of the processing target data corresponds to the halftone output setting of the image processing block 3 (505) that performs the halftone process of the image processing unit (118), and in this embodiment, 2 bits. Suppose that

  Next, the rotation processing control in units of rectangular blocks performed by the rotation processing unit (119) will be described in detail with reference to FIGS.

  First, the mutual processing relationship between the components (601 to 604) of the rotation processing unit (119), the CPU (113), and the RAM (114) will be described with reference to the sequence diagram of FIG. The processing sequence of FIG. 8 is started after the CPU (113) detects that the print data processed by the image processing unit (118) has been output to the RAM (114) in the above-described step (S303). Shall.

  Prior to the start of the processing sequence of FIG. 8, it is assumed that the register values necessary for the operation of the rotation processing unit (119) have been set in the above-described step (S301), and the set values in this embodiment are the data to be processed. The bit gradation is 2 bits, and the rectangular cut-out size is 16 × 16 pixels.

  First, the CPU (113) instructs the image rotation unit (119) to start a rotation process (801). Specifically, the CPU (113) writes 1 to a processing start setting register bit value (initial value is 0) (not shown) included in the rotation processing unit (119).

  When the rotation processing unit (119) confirms that the processing start setting register bit value is 1, the input data transfer unit (601) reads out the print data as the processing target data from the RAM (114) (802). ), The read print data is transferred to the rotation unit (602) (803). Details of the processing of the input data transfer unit (601) will be described later.

  The rotation unit (602) writes the print data received from the input data transfer unit (601) to the work memory (603) (804). Data writing (804) is performed until a predetermined amount of data is written to the work memory (603) while determining the amount of written data (805). The rotation processing unit (119) instructs the input data transfer unit to interrupt data transfer after writing the predetermined amount of data in the work memory (603) (806), and reads data from the work memory (603) ( 807). The rotation unit (602) realizes a rotation process according to the rotation angle setting by combining address scanning and data control in the data write (804) and the data read (807) direction (611). Details of the rotation process will be described later.

  The rotation processing unit (602) transfers the print data read from the work memory (603) to the output data transfer unit (808). Then, the output data transfer unit (604) writes the print data received from the rotation unit (602) to the RAM (114) (810). The rotation processing unit (602) performs data transfer (808) until all data written by data writing (804) is read while determining the amount of data read from the work memory (603) (809). The rotation unit (602) cancels the data transfer interruption instruction to the input data transfer unit (601) (811).

  The processing (802 to 811) described above is processing for one rectangular block, and this processing is repeated for the image size set in the input data transfer unit, thereby performing rotation processing for print data. When the input data transfer unit (601) determines the end of the rotation processing for all image sizes, the CPU notifies the CPU of the end of the rotation processing (812), and the rotation processing unit (119) performs the rotation processing of the print data. finish.

  FIG. 9 shows a processing flowchart of the rotation processing unit (119). The processing shown in this flowchart is executed at the start timing of the above-described step (S304).

  In step (S901), the input data transfer unit (601) reads the processing target data on the RAM (114). In step (S902), the rotation unit (602) performs rotation processing using the work memory (603). Then, the output data transfer unit writes the rotated data to the RAM (114) in step (S903). It is determined whether or not a series of these processes has been performed for the set size in step (S904). Details of input data transfer (S901) and rotation (S902) will be described later.

  FIG. 10 shows a flowchart of the process (S901) performed by the input data transfer unit (601).

  In step (S1001), from the register values set in the rotation processing unit (119) by the CPU (113), the bit gradation of the data to be processed, the rectangular cutout size, the read data start address, the main scan, the sub scan, Each scan size is acquired. In this embodiment, the set values are bit gradation 2, rectangular cutout size 16 × 16 pixels, and top address 0x1000. The main scanning size is 80 pixels and the sub-scanning size is 96 pixels.

  Next, in steps (S1002 to S1006), the unit data length (hereinafter referred to as transfer size) of data transfer performed by the input data transfer unit (119) according to the set value of the bit gradation of the processing target data is set. Set. In the case of 1-bit gradation (YES in S1002), the main scan transfer size is set to 32 bits in step (S1003). This is a transfer size of burst length 1 on the bus (131). If it is not 1-bit gradation (NO in S1002) and 2-bit gradation (YES in S1004), the main scan transfer size is set to 64 bits. This is the transfer size of burst length 2 on the bus (131). If it is not 2 bit gradation (NO in S1004), the main scanning transfer size is set to 128 bits. This is the transfer size of burst length 4 on the bus (131).

  After setting the transfer size, the input data transfer unit (601) proceeds to step (S1007), reads the processing target data on the RAM (114) by the unit rectangular block, and ends the data transfer process.

  Here, referring to FIG. 11, the processing performed when the rotation processing unit (119) reads the unit rectangular block area (1102) from the print data (1101) that is the rotation processing target data on the RAM (114). The operation will be schematically described.

  The print data (1101) is image data processed by the image processing block 3 (505) that performs halftone processing of the image processing unit (118). The print data (1101) in this embodiment has a main scanning size of 160 bits and a sub-scanning size of 192 bits, and assuming that 2-bit gradation data per pixel is placed in the main scanning direction, 80 pixels (main scanning). ) × 96 pixels (sub-scanning).

  The input data transfer unit (601) cuts out the print data (1101) in the unit rectangular block areas (1102 to 1104) and reads the data. The unit rectangular block area is a rectangular area determined by the previously set main scanning transfer size and the sub-scanning direction size of the rectangular cut-out size acquired in step (S1001). In this embodiment, 64 (bits) × 16 (Pixel). To read the unit rectangular block area (1102), first, 64-bit data that is the main scanning transfer size is read from the start address (1110) (1111).

  Specifically, data reading is performed at continuous addresses corresponding to two burst lengths from address 0x1000 to address 0x1007. Next, an address jump corresponding to (main scanning size−main scanning transfer size) is performed from address 0x1008, which is the current address position, and data reading for the next main scanning transfer size is performed (1112). A similar address jump and data reading with the main scanning transfer size are performed for a total of 16 pixel lines, and data reading of the unit rectangular block area (1102) is performed. When the last data reading (1113) of the unit rectangular block area (1102) is completed, an address jump corresponding to (main scanning size × pixel line of the unit rectangular block area) is performed, and the next unit rectangular block area (1103) is started. Data is read from the address (1114).

  Thereafter, data reading is performed for 16 lines, and after data reading (1115) for the 16th line is completed, data is read from the start address of the next unit rectangular block area (1104) (1116). Similarly, the input data transfer unit reads data in a unit rectangular block for all print data by reading continuous address data with a burst length of 2.

  FIG. 12 shows a flowchart of processing relating to the rotation (S903) performed by the rotation unit (602).

  In step (S1201), the CPU (113) acquires the bit gradation and rotation angle setting of the processing target data from the register value set in the rotation processing unit (119). In this embodiment, the bit gradation is 2 and the rotation angle is 90 degrees.

  Next, in steps (S1202 to S1206), the rotating unit (602) determines how many pieces of data in the rectangular block area composed of square pixels are written to and read from the work memory (603) at the same time. Set the number of rectangular blocks shown. In the case of 1-bit gradation (YES in S1202), the number of rectangular blocks is set to 1 in step (S1203). At this time, the rotation unit (602) performs a rotation process by writing one square rectangular block region of 32 × 32 pixels to the work memory (603) and reading it out.

  If it is not 1-bit gradation (NO in S1202) and 2-bit gradation (YES in S1204), the number of rectangular blocks is set to 2 in step (S1205). At this time, the rotation unit (602) performs rotation processing by writing two square rectangular block regions of 16 pixels × 16 pixels into the work memory (603) and reading them out. Further, the data amount of the square rectangular block area is 32 × 16 bits assuming that 2-bit gradation data is placed in the main scanning direction per pixel, and the work memory (603) has two square rectangular block areas. Can store data. If it is not a 2-bit gradation (NO in S1204), the number of rectangular blocks is set to 4 in step (S1206).

  At this time, the rotation unit (602) performs rotation processing by writing four 8-pixel by 8-pixel square rectangular block areas to the work memory (603) and reading them out. The data amount of the square rectangular block area is 32 × 4 bits assuming that 4-bit gradation data is placed in the main scanning direction per pixel, and the work memory (603) stores data for four square block areas. Can be stored. In this embodiment, the bit gradation is 2 and the number of rectangular blocks is 2.

  Thereafter, the process proceeds to step (S1207), and the rotation unit (602) writes the print data transferred by the input data transfer unit (601) to the work memory (603) according to the previously set number of rectangular blocks. In step (S1208), it is determined whether the number of rectangular blocks has been written. In this embodiment, after determining data writing for two rectangular blocks, the rotating unit (602) reads data from the work memory (603) by address scanning capable of rotating by 90 degrees according to the rotation angle setting (S1209). ).

  Here, referring to FIG. 13, in the steps (S1207 to S1209) described above, the rotation unit (602) performs rotation by writing data to and reading data from the work memory (603) in units of rectangular blocks. The process will be schematically described. In the following, description will be made based on the setting in the present embodiment, which is bit gradation 2 and rotation angle setting 90 degrees.

  First, the data writing process in step (S1207) will be described. The rotation unit (602) receives the 64-bit data (1301) read out in the above-described step (S901) by the input data transfer unit (601), and then converts the data into the lower 32 bits (1302) and the upper 32 bits (1303). Divide and write to work memory (603). At this time, the lower 32 bits of data (1302) are written to address 0x00 (1304) of the work memory (603), and the upper 32 bits of data (1303) are written to address 0x10 (1306).

  Similarly, the rotation unit (602) divides 64-bit data received from the input data transfer unit (601) into upper and lower 32 bits, the lower 32 bits are from the address 0x00 (1304), and the upper 32 bits are Write in ascending order from address 0x10 (1306). Data is written for 16 pixels (lines) corresponding to the unit rectangular block area, and finally, the work memory (603) corresponds to two unit rectangular block areas (1310, 13) corresponding to the entire memory area. 1311) is stored.

  Next, the data reading process in step (S1209) will be described. The rotation unit (602) reads the data stored in the work memory (603) in the bit direction (1308, 1309) and the 16-bit data for each unit rectangular block area (1310, 1311) (1308, 1309). . The rotation unit (602) outputs the read 16-bit length data (1308, 1309) to the output data transfer unit (604), and the output data transfer unit (604) determines the RAM (114) according to the rotation angle setting. Write to the rotation destination data storage area secured in advance above.

  For example, the unit rectangular block area (1310) corresponds to a 16 × 32-bit data area (1323) on the rotated data (1320). The output data transfer unit (604) writes the 16-bit length data (1308) read from the work memory (603) by the rotation unit (602) to the 16-bit length data (1321) of the post-rotation data (1320). The unit rectangular block area (1311) corresponds to the area (1324) of the post-rotation data (1320), and the rotation unit (602) rotates 16-bit data (1309) read from the work memory (603). Write to the 16-bit length data (1322) of the subsequent data (1320).

  Note that the data write to the RAM (114) performed by the output data transfer unit (604) is a single write with a burst length of 1. Through the above processing, the rotation unit (602) reads the data written in the work memory (603) and realizes rotation.

  As described above, the rotation processing unit (119) can rotate the print data (1101) on the RAM (114) using the work memory (603) provided therein. In this rotation processing, the work memory (603) can store two square rectangular pixels, and the memory usage rate is 100%. Further, the print data (1101) of the bit gradation 2 on the RAM (114) can be continuously read out with the burst length of 2. The present invention is not limited to the various settings described in this embodiment.

  In this embodiment, continuous address reading with a burst length of 2 is performed on data with a bit gradation of 2, and two square rectangular pixels (16 × 16) are stored in the work memory (603) to perform rotation processing. went.

  As a setting different from this, in the case of setting to rotate data with a bit gradation of 4, the square rectangular pixels stored in the work memory (603) are set to 8 × 8 pixels, and the rotation processing is performed. At this time, the number of bits of the square pixel (8 × 8) is (8 (pixel) × 4 (bit gradation)) × 8, that is, 32 × 4 bits, and the work memory (603) has four squares. If the pixels are stored, the rotation process can be performed with a memory use efficiency of 100%. Further, reading of four square pixels from the RAM (114) can be performed by continuous address reading with a burst length of 4.

  That is, in the present invention, when data consisting of bit gradation n (n is a positive integer) on the RAM is rotated in units of m × m (m is a multiple of n) square rectangular pixels, one side is m N square rectangular pixels are stored in a temporary memory for rotation of × n bits, and the rotation process is performed in this unit. As a result, all the temporary memory for rotation can be used, and continuous address reading with a burst length n can be performed for data transfer from the RAM to the temporary memory for rotation. Then, the data transfer efficiency can be increased and the bus bandwidth load can be reduced.

  The burst transfer of data performed between the RAM (114) and the work memory (603) in the present invention is not limited to the data read performed by the input data transfer unit (601), but the output data transfer unit (604). It may be performed on the side. For example, consider that the image data (1101) is generated by rotating the rotated image data (1320) by 270 degrees. At this time, the input data transfer unit (601) performs single transfer and reads data (1321, 1322), and the output data transfer unit (604) can perform burst transfer of the rotated data with a burst length of 2 (1111 to 1115). ). Therefore, when the data access of the RAM (114) is slower than the data reading, for example, the image data related to the rotation processing is handled on the RAM (114) so as to perform burst transfer on the data writing side. As a result, data reading and data writing to the RAM (114) can be leveled.

102 Image Processing Device 119 Rotation Processing Unit 601 Input Data Transfer Unit 602 Rotation Unit 603 Work Memory 604 Output Data Transfer Unit

Claims (3)

A first memory (114) for storing image data;
For the image data (1101) on the first memory (114),
An image data transfer unit (601, 604) for reading the image data (1101) in units of a predetermined rectangular pixel and writing back to the first memory;
A second memory (603) for storing a rectangular image data area;
Using the image data transfer unit (601, 604), the rotation unit (602) performs writing of image data read out in units of the rectangular pixels and control of reading out to the second memory (603). When,
An image rotation unit (119) for performing rotation processing using
With
The image rotation unit (119), according to the bit gradation number n (n is a positive integer) of the image data (1101),
One of the image data transfer units (601, 604) burst-transfers the rectangular pixel data in a length n in the continuous address direction (1111 to 1115) of the first memory (114). A second memory (603) stores n rectangular blocks made up of the rectangular pixels and performs rotation processing.
An image processing apparatus. The area of the rectangular pixel unit is a square rectangular pixel, and the second memory (603) performs a rotation process in a square rectangular pixel unit.
The image processing apparatus according to claim 1, wherein: The square rectangular pixel is m × m (m is a multiple of n), and the second memory (603) is a square bit region having one side of m × n bit length.
The image processing apparatus according to claim 1, wherein: