JP4200884B2 - Image decompression apparatus and method, and image processing system - Google Patents

Description

  The present invention relates to a technique for performing image processing (particularly image decompression processing) in parallel using a plurality of processors.

  Generally, in a laser printer system, a print object described in PDL (page description language) is developed into a print image by a RIP (raster image processor), and sequentially sent to a print engine for each scan line (raster) for printing. Execute.

  At this time, if all the print images are developed and stored for one page, a very large memory area is required. Therefore, it is common to divide one page into bands composed of a plurality of scanning lines and develop a print image in that unit.

In order to perform such band-by-band expansion processing at high speed, a configuration has been proposed in which each band is assigned to each of a plurality of processors and the expansion processing is performed in parallel (Patent Document 1).
JP 2003-51019 A

  When the RIP is provided in the host device, the developed print image needs to be transmitted to the printer side. In this case, since the print image has a large amount of data, it is usual to adopt a framework in which compression processing is performed before transmission by the host device and decompression is performed on the printer side.

  Here, when the host device or the printer includes a plurality of processors, a configuration in which each band is assigned to each of the plurality of processors and compression processing or decompression processing is performed in parallel as in Patent Document 1 can be considered.

  However, in such a method of parallelizing in units of bands, data of individual scanning lines (individual bands) is not parallelized. Therefore, there remains a problem that the time required for obtaining the processing result for the first one scanning line (one band) is the same as when processing is performed by a single processor. In other words, despite the parallelization by a plurality of processors, no advantage is obtained with respect to the rise until the actual printing is executed.

  In addition, in the case of the method of parallelization in units of bands, in order to improve the throughput, a memory area for storing the processing results for the number of processors so that each processor can execute processing in parallel for the band allocated to itself. Is required. Therefore, if the number of processors is increased to increase the degree of parallelism, there arises a problem that the memory area necessary for processing increases accordingly.

  The inventors of the present application have studied the above problem, and as a result, the individual scanning lines (individual bands) are divided into a plurality of partial regions (hereinafter referred to as “channels”) perpendicular to the direction of the scanning lines. It has been considered that the above object can be achieved by assigning at least one of the plurality of processors to each of the channels and performing image processing (compression processing and decompression processing) in parallel.

  Here, when the compressed data of each channel is transferred from the host device to the printer side, burst transfer is performed in units of data sets including compressed data of the same capacity (for example, n bytes) for each channel in order to transfer data at high speed. It is desirable to have a configuration. In particular, by storing parallel compressed data in SDRAM (Synchronous Dynamic Random Access Memory) on the host device side, burst transfer can be performed at higher speed.

  On the other hand, when performing parallel decompression using a plurality of processors on the printer side, each processor can always read compressed data from the same location using a FIFO (First In First Out) memory in order to simplify the processing of each processor. Is desirable.

  However, if the burst transfer and the FIFO memory are simply combined, that is, if the compressed data subjected to the burst transfer is stored in a single FIFO memory and decompression processing is performed, the following problem occurs.

  Normally, the size of compressed data for one channel differs depending on the compression status. Therefore, when burst transfer is performed for each channel in units of data sets having the same capacity, the data set transferred in one burst transfer A situation may occur in which compressed data of different scan lines (different bands) is included in each channel.

  Such a situation will be described with reference to FIGS. 16 to 18 as an example in which the scanning line is divided into four channels. As shown in FIG. 16, compressed data is stored in the host device. As shown in FIG. 17, burst transfer is performed in units of a 4-byte data set including 1-byte compressed data for one channel. Shall. In this case, the first to sixth burst-set data sets include compressed data related to the first scan line for each channel. However, in the data set that is burst transferred at the seventh time, the first, second, and fourth channels are compressed data related to the first scan line, and the third channel is compressed data related to the second scan line. There is a situation where the compressed data is included.

  When the result of burst transfer is stored in the FIFO memory as it is, the scan line sequence is not maintained as shown in FIG. 18 (that is, in a state where a partial inversion occurs with respect to the scan line sequence). Will be stored. When reading data from the FIFO memory in such a state and trying to synchronize the operation of each processor in units of scan lines, the processor in charge of the third channel reads the data in the sixth byte of the first scan line, third channel ( At the time when the code 1-3-6) is read and the expansion process is performed, the expansion process for the first scan line is terminated and stopped, and the start of the expansion process for the second scan line is awaited. In this case, since the data (reference numeral 2-3-1) of the second byte of the second scan line and third channel remains unread in the FIFO memory, the processor in charge of the fourth channel performs the first scan line and the fourth channel of the first channel. The 7th byte data (reference numeral 1-4-7) cannot be read out.

  As described above, in the configuration in which compressed data transferred in bursts is stored in the FIFO memory as it is and decompression processing is performed, there is a problem that decompression processing cannot be performed in parallel while synchronizing in units of scan lines (or bands). I found out that

  Therefore, the present invention achieves both a configuration for receiving burst-transferred compressed data and a configuration for reading compressed data from a FIFO memory and decompressing in parallel when decompressing compressed data in parallel using a plurality of processors. In addition, an object of the present invention is to provide a framework capable of executing parallel processing at high speed while synchronizing the operation of each processor in units of scan lines or bands.

  In order to solve the above-described problem, the image processing apparatus of the present invention compresses each partial area when the scan line (or band) constituting the image is divided into a plurality of partial areas perpendicular to the direction of the scan line. An image processing apparatus that performs parallel decompression processing on processed data, and includes a plurality of FIFO type memories provided corresponding to each partial area, and compressed data of the same capacity for each partial area A storage means for distributing the burst-transferred data in units of sets to the compressed data of each partial area and storing it in the corresponding memory, and each processor from the memory corresponding to the allocated partial area A parallel processor that reads compressed data in a region and executes decompression processing in parallel in synchronization with each scan line (or band). Characterized in that it comprises a unit.

  According to another aspect of the present invention, there is provided a printer apparatus comprising the image expansion apparatus.

  The image processing system of the present invention is an image processing system that includes an image compression unit and an image decompression unit, each having a function of performing image processing in parallel using a plurality of processors, and the image compression unit In a case where a scan line (or band) constituting an image is divided into a plurality of partial areas perpendicular to the direction of the scan line, at least one of a plurality of processors is assigned to each of the partial areas in parallel. Image compression processing is performed, and a burst is transmitted to the image decompression unit in units of data sets containing the same amount of compressed data for each partial area. The image decompression unit receives the data burst-transferred from the image compression unit, and receives this data. The FIs provided for each partial area are allocated to the compressed data for each partial area. Each processor stores the compressed data of the partial area from each memory allocated to the O-type memory, and controls to perform image decompression processing in parallel in units of scan lines (or bands). It is characterized by that.

  Further, the image processing method of the present invention targets the compressed data for each partial area when the scan line (or band) constituting the image is divided into a plurality of partial areas perpendicular to the direction of the scan line. An image decompression method for decompressing in parallel using a plurality of processors, receiving data burst-transferred in units of data sets including compressed data of the same capacity for each partial area, The process of allocating the compressed data for each area and storing it in a FIFO type memory provided corresponding to each partial area, and each processor reading and scanning the compressed data of the partial area from the memory allocated to each area And a step of performing control so that image expansion processing is performed in parallel in synchronization in units of lines (or bands). .

  In addition, the image decompression method of the present invention can be implemented by a computer, and a computer program therefor is installed or loaded on the computer through various media such as a CD-ROM, a magnetic disk, a semiconductor memory, and a communication network. Can do. It also includes the case where a computer program is recorded and distributed on a printer card or printer option board.

  According to the present invention, when a plurality of processors are used to decompress compressed data in parallel, a configuration for receiving burst-transferred compressed data and a configuration for reading compressed data from a FIFO memory and decompressing in parallel are made compatible. In addition, it is possible to execute parallel processing at high speed while synchronizing the operations of the processors in units of scan lines or bands.

(First embodiment)
FIG. 1 is a block diagram showing a hardware configuration of a printer system 1 according to an embodiment of the present invention. As shown in FIG. 1, the printer system 1 may be any of a host device 10 and a communication network (LAN, Internet, dedicated line, packet communication network, a combination thereof, etc., and includes both wired and wireless. ), A printer device (image forming device) 20 configured to be communicable with the host device 10.

  The host device 10 includes hardware such as a main CPU, a parallel processing unit 15 including parallel processors 11 to 14, a ROM, a RAM, a user interface, and a communication interface. In the present embodiment, the parallel processing unit 15 includes four processors 11 to 14, but the number of processors can be any number greater than or equal to 2 (for example, 8) depending on the design.

  The host device 10 includes a printer driver unit 16 as a normal control function necessary for causing the printer device 20 to execute printing.

  The printer driver unit 16 has the same functional configuration as that of a normal printer driver, and is described in a predetermined printer control language such as a postscript in response to a print request from an application program operating on the host device 10, for example. RIP means for generating a raster image based on print target data, image processing means for creating a print image by performing predetermined image processing (screen processing or the like) on the raster image, and the like.

  However, as will be described later, the printer driver unit 16 of the present embodiment includes a compression control unit 17 that executes image compression processing in parallel with respect to individual scanning lines (or individual bands) using the parallel processing unit 15. This is different from the conventional configuration in that it includes transfer means 18 for burst-transferring the data compressed by the parallel processing unit 15 to the printer device 20 (see FIG. 2).

  Each of these means is functionally realized by the main CPU executing a program stored in a ROM or RAM in the host device 10 or an external storage medium.

  The printer device 20 includes a power mechanism unit and a printer controller 26.

  The power mechanism unit includes a paper feed mechanism that supplies paper into the printer, a print engine that performs printing, and a paper discharge mechanism that discharges paper outside the printer. As the print engine, for example, a serial printer that prints in units of characters, such as an inkjet printer or a thermal transfer printer, a line printer that prints in units of lines, a page printer that prints in units of pages, and the like are used. it can.

  The printer controller 26 includes a main CPU, a parallel processing unit 25 including processors 21 to 24, a ROM, a RAM, a user interface, a communication interface, and the like. In the present embodiment, the parallel processing unit 25 includes four processors 21 to 24. However, the number of processors can be any number of 2 or more (e.g., 8) depending on the design. In addition, the power mechanism unit may include an independent CPU. In this case, the CPU of the power mechanism unit communicates with the main CPU of the information processing unit via a predetermined communication path to control the print engine. Thus, a printing operation is performed.

  The printer controller 26 has the same functional configuration as that of a printer controller in a normal printer. For example, the printer controller 26 receives commands and data from an engine control unit that controls a power mechanism unit to execute printing and a host device 10 and stores them in a reception buffer. The receiving means 27 etc. to store are provided.

  However, when the receiving means 27 of this embodiment receives burst-transferred data, the receiving means 27 sorts the received data into compressed data for each partial area, and a plurality of FIFO types provided corresponding to the respective partial areas. Stored in the receiving buffer (input buffer).

  Further, the printer controller 26 according to the present embodiment executes decompression processing in parallel for each scanning line (or each band) using the parallel processing unit 25 based on the compressed data stored in each reception buffer. The present embodiment is different from the conventional configuration in that it includes a decompression control means 28 for transferring, a transfer means 29 for transferring the decompression processing results by the parallel processing unit 25 to the print engine in the order used for printing (see FIG. 2).

  Each of these means is realized by the main CPU executing a program stored in a ROM or RAM in the printer device 20, an external storage medium, or the like.

  Hereinafter, the printing process in the printer system 1 will be described with reference to the flowcharts shown in FIGS. Each step (including a partial step to which no code is assigned) can be executed in any order or in parallel as long as no contradiction occurs in the processing contents.

(Processing in the host device 10)
When the printer driver unit 16 receives a print request from an application program operating on the outside or the host device 10, the printer driver unit 16 transmits a print instruction command to the printer device 20 (printer controller 26), and also to the RIP unit or the like. To start the process.

  When the RIP unit receives an instruction to start processing, the RIP unit generates a raster image based on print target data described in a predetermined printer control language such as a postscript received from the application program. If the print target data can be received in the form of a raster image from an application program or the like, the processing by the RIP unit can be omitted.

  The image processing means performs predetermined image processing (screen processing or the like) on the generated raster image, generates a print image, and stores it in a predetermined area of the RAM.

  The compression control means 17 selects lines in the main scanning direction (hereinafter simply referred to as “scanning lines”) constituting the generated print image in the order of the sub-scanning direction (FIG. 3: step S100). The main scanning direction and the sub-scanning direction are determined based on scanning in the printer device 20.

  Next, the compression control means 17 divides the selected scanning line into a plurality of partial areas perpendicular to the direction of the scanning line (FIG. 3: step S101). Hereinafter, such a partial area is referred to as a “channel”, and channel numbers are added in order of arrangement in the scanning line direction, that is, from the left to identify each channel (see FIG. 4). In the present embodiment, the concept of right and left is used on the assumption that the scanning line is scanned horizontally from left to right.

  Next, the compression control means 17 specifies the head position of the print image (hereinafter referred to as “partial data”) included in each channel in the selected scanning line (FIG. 3: step S102). For example, when one scan line is 2048 pixels, one scan line is equally divided into four partial areas according to the number of processors, and the first pixel, the 512th pixel, the 1024th pixel, and the 1536th pixel are respectively partial. It may be possible to specify the head position of data.

  Note that steps S101 to S102 can be omitted if channels are defined in common for all scanning lines and the head position of partial data of each channel is specified in advance.

  Next, the compression control means 17 assigns at least one of the processors 11 to 14 to each of the plurality of channels (FIG. 3: step S103). Specifically, a combination of a channel and a processor is fixedly assigned, such as a processor 11 for channel 1 and a processor 12 for channel 2.

  Next, the compression control means 17 determines whether or not the allocated processor can process new data. If the allocated data can be processed, the compression control means 17 passes the head position of the corresponding partial data to the allocated processor, and the partial data Is read (FIG. 3: Step S104).

  Here, the case where the processor can process new data is a case where the processor can newly write the next processing result. For example, when a partial area of the output buffer is assigned to each of the processors 11 to 14 (channels 1 to 4) as shown in FIG. 16 as an area for writing the processing result, is there an empty area in the output buffer allocation area? Alternatively, it may be determined whether or not new data can be processed based on whether or not the transfer by the transfer means 18 is completed and can be overwritten on the processing result already written in the allocation area.

  Note that the output buffer is preferably constituted by an SDRAM capable of performing burst transfer at high speed. This is because the processing result (compressed data) written in the output buffer is burst transferred by the transfer means 18 as will be described later.

  Next, when there is an unselected scan line among the scan lines constituting the generated print image, the compression control means 17 returns to Step S100 (FIG. 3: Step S105).

  When each of the processors 11 to 14 of the parallel processing unit 15 receives the start position and read instruction of the partial data from the compression control means 17 (FIG. 5: step S200: YES), a predetermined area of the RAM storing the generated print image The partial data is read out based on the head position (FIG. 5: step S201).

  Next, a predetermined compression process is performed on the read partial data to generate partial compressed data (FIG. 5: step S202). At this time, predetermined boundary information is added to the end (or first) of the partially compressed data so that the boundary of the partially compressed data can be detected. As the predetermined compression processing, various conventional compression algorithms can be adopted depending on the design. For example, if the print image is binary data, JBIG (Joint Bi-level Image Experts Group). It is conceivable to adopt an algorithm.

  Then, the generated partial compressed data is written in a partial area assigned to the output buffer (FIG. 5: step S203). When assigning a partial area of the output buffer to each processor as shown in FIG. 16, for example, a counter that stores the current address for each processor so that each processor can refer to the address (current address) to be written next. It is conceivable to adopt a configuration in which a counter is added according to the written capacity (number of bytes).

  The transfer unit 18 determines whether there is untransferred data in the output buffer for the generated print image (FIG. 6: step S300). If there is no untransferred data (transfer has ended), the process ends.

  On the other hand, if there is untransferred data, the transfer unit 18 can determine whether the printer device 20 (printer controller 26) can receive the partial compressed data based on the communication result with the printer device 20 (printer controller 26), for example. It is determined whether or not (FIG. 6: step S301).

  If reception is possible, data transfer from the output buffer to the printer device 20 (printer controller 26) is controlled so that burst transfer is performed in units of data sets including partially compressed data of the same capacity for each channel ( FIG. 6: Step S302).

  Although the size of the data set for burst transfer can be determined according to the design, in this embodiment, as shown in FIG. 7, a total of 32 bytes (4 bytes including partial compressed data for 8 bytes for each channel). It is assumed that burst transfer is performed in units of a data set of (channel × 8 bytes). When burst transfer is performed in this way, there is a possibility that partially compressed data relating to a plurality of scan lines is included in the data set to be burst transferred. Referring to FIG. 7, it can be seen that the first burst transfer includes partially compressed data relating to the first scan line and the second scan line.

(Processing in the printer device 20)
When a print instruction command is sent from the host device 10, the printer controller 26 receives this via the receiving unit 27 and controls the power mechanism unit by the engine control unit to prepare for printing.

  When receiving the print instruction command and receiving the burst-transferred data set (FIG. 8: Yes in step S400), the receiving unit 27 executes the storing process in the reception buffer.

  Here, as described above, the reception buffer is a buffer for allocating and storing data that is burst-transferred from the host device 10 in units of data sets including the partial compressed data of each channel as partial compressed data for each channel. Therefore, for example, a FIFO type buffer as shown in FIG. 9 may be configured for each channel on the RAM (or a plurality of FIFO type memories may be provided for each channel). . In the example shown in FIG. 9A, four reception buffers corresponding to the number of channels are configured so that burst-transferred data can be sorted and stored for each channel.

  The receiving means 27 distributes the burst-transferred data to the partial compressed data for each channel according to the setting of the data set for burst transfer (N channels × M bytes). Specifically, the receiving means 27 selects a receiving buffer in the order of channel arrangement (FIG. 8: step S401). Then, data for one channel capacity (partially compressed data for one channel) is extracted from the received data and stored in the selected reception buffer (FIG. 8: step S402). For example, in this embodiment, from the host device 10 Since the data is transferred in bursts in units of a total of 32 bytes including 8 bytes of partial compressed data for each channel, the reception means 27 selects the reception buffer 1 corresponding to channel 1 and receives the received data. 8 bytes are extracted from the received data and stored in the reception buffer 1.

  The receiving means 27 determines whether or not the storage process for the data of all channels has been completed (FIG. 8: step S403), and if it is determined that the storage process has not been completed (FIG. 8: step S403). No), the process returns to S401 to process the remaining channel data. On the other hand, if it is determined that the storage process has been completed for all the channel data (FIG. 8: Yes in step S403), the process ends, and the expansion control unit 28 is instructed to start the expansion process. As a result, the process of storing the data received by the first burst transfer in each reception buffer is completed.

  When a plurality of reception buffers are configured in accordance with the number of channels in this way, the partial compressed data of channel 1 is continuously stored in the reception buffer 1, and the partial compressed data of channel 2 is continuously stored in the reception buffer 2. The partially compressed data of channel N is continuously stored in the reception buffer N. That is, since the partial compressed data of each channel is stored in the reception buffer corresponding to each channel as shown in FIG. 9B, each processor of the parallel processing unit 25 always specifies the same location (reception buffer). Thus, the partial compressed data of the channel assigned to each processor can be read.

  When the expansion control unit 28 receives an instruction to start expansion processing, it sets “unfinished” as an initial value to the expansion flag prepared for each channel (FIG. 10: step S500). As will be described later, the decompression flag is changed to “end” when the decompression of the partial compressed data relating to one scanning line is finished in each channel.

  Next, the decompression control means 28 instructs each of the processors 21 to 24 of the parallel processing unit 25 to read out the partially compressed data from the reception buffer corresponding to the allocated channel (FIG. 10: Step S501). . Then, it is determined whether or not the expansion flags of all the channels are “end” (FIG. 10: step S502). If all the expansion flags are “end”, one scanning line is related. It is determined that the decompression of the partial compressed data has been completed, and it is determined whether or not the partial compressed data is still stored in the reception buffer (FIG. 10: Step S503). If the partial compressed data is stored in the reception buffer, the process returns to S500 to perform the partial compressed data expansion process for the next scan line.

  On the other hand, if the decompression control means 28 determines in S502 that all the decompression flags are not “finished”, it regards that the decompression of the partial compressed data related to one scanning line has not yet been finished, The decompression flag is checked (FIG. 10: Step S502).

  Next, decompression processing of each of the processors 21 to 24 of the parallel processing unit 25 will be described. In this embodiment, each processor 21 to 24 is assigned a channel to be processed in advance and a reception buffer from which partial compressed data is to be read. However, when the decompression control means 28 starts processing, each processor 21 It is good also as a structure which allocates the channel which should be processed with respect to -24.

  When each processor 21 to 24 receives a partial compressed data read instruction from the decompression control means 28 (FIG. 11: S600), it reads out the partial compressed data byte by byte from the receiving buffer to which each processor is assigned (FIG. 11: step). S601).

  Next, each of the processors 21 to 24 executes a predetermined decompression process on the read 1-byte data to generate a part of the corresponding partial data (FIG. 11: step S602). Store in the output buffer (FIG. 11: step S603). The decompression process needs to employ the decompression process corresponding to the compression algorithm employed in the parallel processing unit 15 of the host device 10.

  Next, each of the processors 21 to 24 determines whether or not the read 1-byte data includes information indicating the end of the partial compressed data (FIG. 11: step S604). If the read 1-byte data does not include information indicating the end of the partial compressed data (FIG. 11: Yes in step S604), the process returns to S601 to continue the decompression process.

  On the other hand, when the information indicating the end is included (FIG. 11: No in step S604), the expansion flag is changed to “end” (FIG. 11: step S605), and synchronization on the scanning line is realized. Therefore, the process is stopped.

  As described above, in this embodiment, since a reception buffer is provided for each channel, that is, for each processor, each processor stops processing when decompression processing of partial compressed data for one scan line is completed (that is, scanning). Even in a configuration that operates synchronously in line units, it is possible to perform decompression processing in parallel.

  The transfer unit 29 selects scanning lines in the order of the sub-scanning direction in order to transmit data in the order used for printing (FIG. 12: Step S700).

  Next, the transfer unit 29 determines whether or not the partial data for one scanning line is stored in the output buffer in a state where all the partial data for the selected scanning line has been expanded (step S701 in FIG. 12).

  If it is determined that the data is stored in an expanded state, partial data is read from the output buffer for the selected scan line and transferred to the print engine in order to transmit the data in the order used for printing. FIG. 12: Step S702).

  Thereafter, when there is an unselected scan line for the generated print image, the transfer unit 29 returns to S700 (FIG. 12: step S703).

  According to the configuration of the present embodiment, the compression processing is performed in parallel for each channel constituting one scan line (that is, for each partial data), so the compression processing time for each scan line Can be shortened. In particular, since the compression processing time for the first scan line is shortened, the compression processing for the first scan line is completed after the compression processing for the print image is started, and the processing result is the printer device 20 (printer controller 26). It is possible to shorten the time until the data is transferred to (), and it is possible to realize a compression process (and thus a print process) that starts quickly.

  Similarly, since decompression processing is executed in parallel for each channel constituting one scan line (that is, for each partial compressed data), the decompression processing time for each scan line can be shortened. it can. In particular, since the decompression time of the first scan line is shortened, the decompression process is completed for the first scan line after the decompression process of the compressed print image is started, and the processing result is transferred to the print engine. It is possible to shorten the time until the image is processed, and it is possible to realize a decompression process (and thus a print process) that starts quickly.

  Further, in the printer controller 26, the reception buffer that stores the burst-transferred partial compressed data is configured as a plurality of FIFO-type reception buffers corresponding to the respective channels, so that the burst-transferred partial compressed data is received. Provide a framework capable of executing parallel processing at high speed while combining a configuration and a configuration in which partially compressed data is read from a FIFO type reception buffer and synchronized in parallel in units of scan lines (or bands). be able to. As a result, it is possible to respond in real time to a laser printer or the like that needs to continuously supply print data in units of scanning lines to the engine at a constant speed.

(Second Embodiment)
In the first embodiment, the configuration in which each processor stops processing when the partial compressed data decompression processing for one scan line is completed (that is, the configuration that operates synchronously in units of scan lines) has been described. In this embodiment, each processor performs expansion processing without synchronizing in units of scan lines, thereby increasing the processing speed of each processor.

However, if each processor is configured to perform decompression processing without synchronization, a situation occurs in which the output buffer includes partial decompression data for different scan lines depending on the channel. Therefore, in the second embodiment, In addition, an intermediate output buffer for storing the partial decompression data of each processor is provided for each channel, and the partial decompression data stored in the intermediate output buffer is controlled to be stored in the output buffer in units of scanning lines. It was. Hereinafter, the decompression processing according to the second embodiment will be described in detail with reference to FIGS. Note that the processing in the host device 10 has the same configuration as the processing in the first embodiment, and thus description thereof is omitted here.

  First, in addition to the functions described in the first embodiment, the printer controller 26 according to the present embodiment is provided with one scan line from an intermediate output buffer that stores partially expanded data expanded by each processor for each channel. Is provided with data order adjusting means 30 for reading out the partial decompressed data and transferring them to the output buffer (see FIG. 13).

  Next, processing in the printer device 20 in the present embodiment will be described. Similarly to the first embodiment, the receiving unit 27 sorts the partially compressed data burst-transferred into the partially compressed data for each channel and stores them in the respective receiving buffers (see the flowchart of FIG. 8). If the reception unit 27 determines that the storage process has been completed for all the channel data, the reception unit 27 terminates the process, and from the reception buffers assigned to the processors 21 to 24 of the parallel processing unit 25, respectively. Is instructed to read the partially compressed data.

  When each of the processors 21 to 24 receives an instruction to read partial compressed data from the receiving unit 27, the processors 21 to 24 read the partial compressed data byte by byte from the reception buffer to which each processor is assigned. Then, each of the processors 21 to 24 executes a predetermined decompression process on the read 1-byte data to generate a part of the corresponding partial data, and each corresponding intermediate output buffer as shown in FIG. 1 to 4 are stored. FIG. 14 is a diagram for explaining an intermediate output buffer provided for each channel.

  At this time, each of the processors 21 to 24 determines whether or not the read 1-byte data includes information indicating the end of the partial compressed data. When the read 1-byte data includes information indicating the end of the partial compressed data, a predetermined boundary is set at the end (or first) of the expanded partial data so that the boundary of the expanded partial data can be detected. Add information.

  Next, the data order adjusting unit 30 sets “unfinished” as an initial value in a transfer flag prepared for each channel (FIG. 15: step S800). As will be described later, the transfer flag is changed to “end” when the transfer of the partial compressed data related to one scan line (transfer from the intermediate output buffer to the output buffer) is completed in each channel.

  Next, the data order adjusting means 30 selects the channels in the order of the channels (FIG. 15: Step S801). Then, the transfer flag of the selected channel is checked, and if the transfer flag is “end”, the process returns to S801 (FIG. 15: step S802).

  On the other hand, when the transfer flag is “unfinished”, the data order adjusting unit 30 specifies the intermediate output buffer corresponding to the selected channel among the intermediate output buffers 1 to 4 and specifies the specified intermediate output. One byte of data is selected from the untransferred partially expanded data stored in the buffer (FIG. 15: step S803).

  Next, the data order adjusting unit 30 transfers the selected 1-byte data to the output buffer (FIG. 15: step S804), and whether the selected 1-byte data includes information indicating the end of the partial decompression data. It is determined whether or not (FIG. 15: Step S805).

  If the selected 1-byte data does not include information indicating the end of the partial decompression data, the data order adjustment unit 30 returns to step S803 and executes transfer processing for the next 1-byte data. On the other hand, if the selected 1-byte data includes information indicating the end of the partial decompression data, the transfer flag of the selected channel is changed to “end” (FIG. 15: step S806).

  Next, when there is a channel whose transfer flag is “incomplete” (FIG. 15: Yes in step S807), the data order adjusting unit 30 transfers partial expanded data of all channels for one scanning line. If not completed, the process returns to S801 to select the next channel.

  On the other hand, when the transfer flags of all the channels are “end” (FIG. 15: No in step S807), the data order adjusting unit 30 transfers the data (untransferred scan) to be transferred to the intermediate output buffers 1 to 4. It is determined whether or not (line expansion data) remains (FIG. 15: step S808). If it remains, the process returns to S800 to transfer the next scan line.

  Next, as in the first embodiment, the transfer unit 29 selects scan lines in the sub-scanning direction in order to transmit data in the order used for printing, and the selected scan line is equivalent to one scan line. It is determined whether or not all partial data is stored in the output buffer in a decompressed state. If it is determined that the data is stored in an expanded state, partial data is read from the output buffer for the selected scan line and transferred to the print engine in order to transmit the data in the order used for printing. (See the flowchart in FIG. 12).

  When the processing is configured in this way, each processor can individually decompress the data of each channel, and a series of data transferred until the transfer flag of all the channels is “finished” is 1 Since it belongs to partial data relating to one scanning line, partial data belonging to one scanning line is continuously stored in the output buffer in a state where the arrangement order of the scanning lines is maintained. As a result, the image data is transferred to the print engine while being synchronized in units of scan lines, and the decompression process of each processor can be executed at high speed.

(Other embodiments)
The present invention is not limited to the above-described embodiment, and can be variously modified and applied. For example, the present invention can be applied to a system other than a printer system as long as the system compresses / decompresses an image.

  For example, in the above-described embodiment, the printer device 20 includes the parallel processing unit 25 and the printer controller 26. However, the present invention is not necessarily limited to such a configuration. For example, the parallel processing unit 25 and the printer controller 26 may be configured as external devices that can be connected to the printer device 20. Furthermore, the parallel processing unit 25 and the printer controller 26 may be configured as devices that can be connected to the host device 10 according to a standard such as a PCI bus. When the parallel processing unit 25 is connected to the host device 10, the functions of the printer controller 26 may be realized by the main CPU or the like of the host device 10.

  For example, in the above-described embodiment, the configuration in which one scan line is equally divided into partial areas according to the number of processors has been described. However, for example, the size of the partial area to be allocated is changed according to the processor specifications and the like. It is not necessary to divide. Further, the number of divisions (number of channels) of the partial area is not necessarily equal to the number of processors.

  Further, for example, in the above-described embodiment, the configuration in which the combination of the channel and the processor is fixedly described has been described. However, for example, the processor that has completed the processing may be allocated to the next channel. In this case, the correspondence between the processor and the channel may be different for each scan line. Note that not only the processors of the parallel processing units 15 and 25 but also the main CPU may be assigned channels to execute parallel processing.

  Further, for example, in the above embodiment, the configuration in which the processors 11 to 14 read partial data from the RAM has been described. However, for example, the host device 10 transfers means for transferring partial data from the RAM to the processors 11 to 14 (for example, DMA transfer means). ), The compression control means may instruct the means. In this case, the processors 11 to 14 execute processing upon receiving the transfer from the means. Similarly, for example, when the printer device 20 includes means for transferring partially compressed data from the data reception buffer to the processors 21 to 24 (for example, DMA transfer means), the decompression control means may instruct the means. The processors 21 to 24 execute processing in response to the transfer from the means.

  Further, for example, in the above-described embodiment, the configuration in which the parallel processing is performed for all the scan lines of the print image has been described. However, if the present invention is applied to at least one scan line constituting the print image, the compression processing time for the scan line is determined. / An effect of shortening the extension processing time can be obtained.

  Further, for example, in the above-described embodiment, a configuration has been described in which the transfer unit 29 transfers data to the print engine when data for one scan line is prepared. However, depending on the type of print engine, data for one scan line is described. Alternatively, the data may be transferred to the print engine without waiting for the data to be aligned or when the data for a plurality of scanning lines has been prepared.

  For example, in the above embodiment, the compression processing, the decompression processing, and the unit to be synchronized are all configured based on the scanning line, but the present invention is not necessarily limited to such a configuration. For example, the processing may be configured on the basis of a band including a predetermined number of scanning lines. In this case, in principle, the processing may be configured by replacing “scan lines” with “bands” in the above-described embodiment, but some steps need to be changed within a range obvious to those skilled in the art. For example, for step S102, the compression control means 17 needs to specify the head address for each scanning line included in the band as the head position of the partial data of each channel in the selected band. Assuming a configuration with a band as a reference, the above embodiment can be considered as an aspect in which one scanning line = 1 band.

2 is a block diagram illustrating a hardware configuration of a printer system. FIG. 2 is a block diagram illustrating a functional configuration diagram of the printer system. FIG. 4 is a flowchart showing processing contents of a compression control means 17; It is a figure for demonstrating a channel. 3 is a flowchart showing processing contents in a parallel processing unit 15. 3 is a flowchart showing processing contents in a transfer means 18; It is a figure for demonstrating the data set transmitted by burst. 4 is a flowchart showing processing contents of a receiving means 26. It is a figure for demonstrating each receiving buffer. 4 is a flowchart showing processing contents in an expansion control means 28. 4 is a flowchart showing processing contents in a parallel processing unit 25. 4 is a flowchart showing processing contents of a transfer means 29. FIG. 10 is another block diagram illustrating a functional configuration diagram of the printer controller 26. It is a figure for demonstrating an intermediate | middle output buffer. 4 is a flowchart showing processing contents in a data order adjusting unit 30. It is a figure explaining the output buffer in the host apparatus 10 (compression side). It is a figure for demonstrating the condition by which burst transfer is carried out. FIG. 10 is a diagram illustrating a state in which a burst-transferred data set is stored in a FIFO memory in the related art.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Printer system, 10 Host apparatus, 11-14 Processor for parallel processing, 15 Parallel processing unit, 16 Printer driver means, 17 Compression control means, 18 Transfer means, 20 Printer apparatus, 21-24 Processor for parallel processing, 25 Parallel processing Unit, 26 printer controller, 27 receiving means, 28 decompression control means, 29 transfer means

Claims (5)

An image processing device that performs decompression processing in parallel on data compressed for each partial area when the scanning line (or band) constituting the image is divided into a plurality of partial areas perpendicular to the direction of the scanning line Because
A plurality of FIFO type memories provided corresponding to the respective partial areas;
Storage means for allocating the data burst-transferred in units of data sets including the same amount of compressed data for each partial area, and storing the data in the corresponding memory by allocating the compressed data of each partial area;
Each processor, and a parallel processor unit for executing expansion processing in parallel in synchronization with the scan line (or band) unit compressed data reads of each partial area from the memory corresponding to the allocated portion region, respectively, the Prepared,
The compressed data of each partial area includes predetermined boundary information added when the compressed data is generated at the end of the compressed data,
The burst transferred data set includes compressed data of scan lines (or bands) that vary depending on the partial area.
Each of the processors determines whether or not the compressed data of each partial area read from the memory includes the predetermined boundary information. When the predetermined boundary information is included, the processor performs its own decompression process. An image decompression apparatus that performs decompression processing in parallel in synchronization with each scanning line (or band) by stopping .   A printer apparatus comprising the image expansion apparatus according to claim 1. An image processing system including an image compression unit and an image expansion unit, each having a function of performing image processing in parallel using a plurality of processors,
The image compression unit
When scanning lines (or bands) constituting an image are divided into a plurality of partial areas perpendicular to the direction of the scanning lines, at least one of a plurality of processors is assigned to each of the partial areas and image compression is performed in parallel. Perform processing, burst transmission to the image expansion unit in units of data sets containing the same amount of compressed data for each partial area,
Image decompression section
The data transferred in burst from the image compression unit is received, the received data is distributed into compressed data for each partial area, and stored in a FIFO type memory provided for each partial area. The compressed data of the partial area is read from each memory allocated to each , and controlled so as to perform image expansion processing in parallel in synchronization with each scanning line (or band) ,
The image compression unit adds predetermined boundary information to the end of the compressed data of each partial area, and burst-transmits a data set including compressed data of scan lines (or bands) that differ depending on the partial area to the image expansion unit. And
The image decompression unit
Each processor determines whether or not the compressed data of each partial area read from the memory includes the predetermined boundary information. If the processor includes the predetermined boundary information, the processor performs its own decompression process. by stopping the image processing system characterized that you execute the decompression processing in parallel in synchronization with the scan line (or band) units. A scan line (or band) constituting an image is divided into a plurality of partial areas perpendicular to the direction of the scan line, and the compressed data compressed for each partial area is used as a target in parallel using a plurality of processors. An image decompression method that decompresses
Receives data transferred in bursts in units of data sets containing the same amount of compressed data for each partial area, and distributes the received data to compressed data for each partial area. and the step of storing the FIFO type of memory was,
Each processor reads out the compressed data of the partial area from the memory allocated to each processor, and controls to perform image expansion processing in parallel in synchronization with each scan line (or band) ,
In the step of storing in the memory, predetermined boundary information is added to the end of the compressed data of each partial area, and a data set including compressed data of different scanning lines (or bands) depending on the partial area is sent to the image decompression unit. Including burst transmission,
In the controlling step, each processor determines whether or not the compressed data of each partial area read from the memory includes the predetermined boundary information, and when the predetermined boundary information is included An image decompression method comprising: controlling the decompression process in parallel in synchronization with each scanning line (or band) by stopping the self decompression process.   A program for causing a computer to execute the image decompression method according to claim 4.

Images