JP4158695B2 - Image processing method and image processing apparatus - Google Patents

Description

  The present invention relates to an image processing method and an image processing apparatus capable of rotating image data stored in a page memory at high speed.

  In an image processing apparatus that processes image data, such as a facsimile machine, a digital copying machine, a printer, or a multifunction machine that integrates these functions, image data read by the image processing apparatus or an image input from an external device Some have a function of rotating data 90 degrees clockwise or counterclockwise and recording the data on a sheet (for example, see Patent Document 1). Using such image data rotation processing, the long and short directions of the graphic drawn by the image data are made to coincide with the long and short directions of the paper to be printed out, and the paper to be printed out set in a predetermined direction is out of paper. Even in such a case, it is possible to continue the print output process using the other paper set with the paper changed 90 degrees.

  The rotation processing of image data in the conventional image processing is performed as follows using a page memory. In other words, image data read by the image processing device or image data input from an external device is temporarily stored in a buffer, and when a predetermined page of image data is stored in the buffer, later reading is considered. The image data for one page stored in the buffer is developed at a predetermined address of the page memory calculated in this way, and the developed image data is read out and sent to the recording unit.

Since the capacity of color image data handled as image data is generally large, it is necessary to read or write image data at a higher speed. Therefore, for example, when performing burst reading or burst writing with 4 pixels as a unit and accessing in units of banks, address control is performed based on a signal that circulates the bank, for example, until the bank selection signal circulates. A random access device that can be precharged and speeds up reading or writing by accessing a memory without interruption has been disclosed (for example, see Patent Document 2).
JP 2000-268169 A JP 2002-184175 A

  However, a conventional image processing apparatus that performs rotation processing of image data requires an arithmetic device for calculating the address of a page memory to which image data is to be written by, for example, multiplication processing. There is a problem that cost increases are inevitable. In addition, it is necessary to have a buffer having at least the same capacity as the page memory, which is one of the causes of increasing the bulk and cost of the device configuration.

  In addition, in order to speed up reading or writing of image data, for example, when performing burst reading or burst writing with 4 pixels as a unit, continuity of bank access is not necessarily guaranteed, and the precharge time is sufficient. There is also a problem that there is a case where the memory cannot be secured and the memory cannot be accessed at high speed.

  The present invention has been made in view of such circumstances, and even if the configuration is smaller and lower in cost than the conventional device, the data reading or writing speed can be surely increased at the time of burst reading or burst writing. An object of the present invention is to provide an image processing method and an image processing apparatus capable of performing data rotation processing at high speed.

In order to achieve the above object, an image processing method according to a first aspect of the present invention is an image for performing burst reading or burst writing with a plurality of pixels as one bank on one page of image data expanded two-dimensionally in a memory. In the processing method, when burst read or burst write is performed across each line of the memory, it is determined whether or not the bank numbers of the first banks of the lines of the memory are continuous, and are determined not to be continuous. In this case, one bank of empty addresses not having actual data is added to the end of each line of the memory, and a plurality of lines corresponding to the number of pixels in one row direction developed in the memory are set to 1 When performing burst read or burst write across the block, the bank number of the first bank of each line of the memory is continuous. If the Most is determined, characterized by adding one bank a free address with no actual data at the end of the first block of the memory.

In the image processing method according to the first aspect of the invention, burst reading or burst writing is performed by using a plurality of pixels as one bank, and burst reading or burst writing is performed on one page of image data expanded two-dimensionally in a memory. If the bank number of the first bank of each line in the row direction is not consecutive, an empty address that does not have actual data is added to the end of each line for one bank, and the bank number of the first bank of each line in the row direction To be continuous. As a result, an arithmetic unit for calculating the write address in the page memory is unnecessary, and the configuration of the apparatus can be reduced in size and cost. Also, since the bank numbers to be accessed are continuous during burst reading or burst writing of image data, it is possible to secure a precharge time during the circulation of the bank selection number, and even during burst reading or burst writing. Data reading or writing speed can be increased, and image data rotation processing can be performed at high speed.
If the bank number of the first bank of each line in the row direction is not continuous when burst reading or burst writing is performed, an empty address that does not have actual data at the end of one block that is a memory reading or writing unit Is added for one bank so that the bank numbers of the first bank of each line in the row direction are consecutive. As a result, since the bank numbers to be accessed are continuous at the time of burst reading or burst writing of image data, it is possible to secure a precharge time during the circulation of the bank selection number, and at the time of burst reading or burst writing. In addition, the data reading or writing speed can be increased, and the rotation processing of the image data can be performed at a high speed.

According to a second aspect of the present invention, there is provided a memory for developing and storing image data in two dimensions, and burst reading or burst writing for a plurality of pixels as one bank with respect to the image data stored in the memory. Determining whether or not the bank numbers of the first bank of each line of the memory are consecutive when performing burst reading or burst writing across each line of the memory . and 1 determining means, the case where the bank number is determined not to be continuous in the first determination means, the first additional adding one bank free address at the end of each line of the memory without a real data means and, a plurality of lines corresponding to the number of one unit in the row direction of the pixels to expand the memory as one block, burst read or burst write across the block When determining that the bank numbers of the first bank of each line of the memory are consecutive, and when the second determining means determines that the bank numbers are not consecutive And a second adding means for adding one bank of empty addresses having no real data at the end of one block of the memory .

In the image processing apparatus of the second invention, one page of image data is expanded in a page memory, burst reading or burst writing is performed on the expanded image data with a plurality of pixels as one bank, burst reading or If the bank number of the first bank of each line is not consecutive when burst writing is performed, an empty address not having actual data is added to the end of each line for one bank, and the bank number of the first bank of each line is consecutive. To do. As a result, an arithmetic unit for calculating the write address in the page memory is unnecessary, and the configuration of the apparatus can be reduced in size and cost. Also, since the bank numbers to be accessed are continuous during burst reading or burst writing of image data, it is possible to secure a precharge time during the circulation of the bank selection number, and even during burst reading or burst writing. Data reading or writing speed can be increased, and image data rotation processing can be performed at high speed.
If the bank number of the first bank of each line in the row direction is not continuous when burst reading or burst writing is performed, an empty address that does not have actual data at the end of one block that is a memory reading or writing unit Is added for one bank so that the bank numbers of the first bank of each line in the row direction are consecutive. As a result, since the bank numbers to be accessed are continuous at the time of burst reading or burst writing of image data, it is possible to secure a precharge time during the circulation of the bank selection number, and at the time of burst reading or burst writing. In addition, the data reading or writing speed can be increased, and the rotation processing of the image data can be performed at a high speed.

  According to the present invention, a circuit for performing complicated address calculation is unnecessary, and it is not necessary to provide a large-capacity buffer, and image data rotated by the storage capacity of the page memory can be generated. By providing a small buffer with a minimum required capacity, it is possible to perform image data rotation processing. In addition, since the banks to be accessed are continuous during burst reading or burst writing of image data, it is possible to secure a precharge time during the circulation of the bank selection signal. High-speed reading or writing can be realized, and image data rotation processing can be performed at high speed. As a result, it is possible to realize an image processing apparatus that can read or write image data at high speed even with a small-scale configuration and can perform image data rotation processing at high speed.

  Hereinafter, the present invention will be described in detail with reference to the drawings showing embodiments thereof. FIG. 1 is a block diagram showing a configuration of a facsimile composite apparatus 20 as an image processing apparatus of the present invention.

  The facsimile multifunction device 20 includes a control unit 1, a reading unit 2, a recording unit 3, a display unit 4, an operation unit 5, a ROM 6, a RAM 7, a codec unit 8, an image memory 9, a page memory 10, a buffer 11, a modem 12, an NCU ( Network Control Unit) 13, a PC interface unit 14 and the like. The facsimile multifunction device 20 responds to a reading function for reading an original by the reading unit 2 to acquire image data, read image data, image data received by facsimile communication, or image data received from an externally connected PC. A recording function for recording an image in the recording unit 3 and a transmission function for facsimile transmission of read image data or received image data are provided.

  Specifically, the control unit 1 is configured by a CPU, and is connected to the above-described hardware units of the facsimile multifunction peripheral 20 via the bus 15. The control unit 1 controls the above-described hardware units, and the ROM 6. Various software functions are executed in accordance with the control program stored in. The reading unit 2 reads a document using a CCD image sensor, for example, and outputs the read image data. Specifically, the optical unit irradiates and scans the document fed to the platen glass, and the reflected light from the document is captured into a CCD image sensor via a mirror, a lens, etc., and the scanned document image data is captured. Read. The read image data is output after, for example, shading processing.

  The recording unit 3 is an electrophotographic printer device, and displays images corresponding to image data of a document read by the reading unit 2, image data received by facsimile communication, image data sent from an external PC, and the like. Print on paper. The display unit 4 is a display device such as a liquid crystal display device or a CRT display. The display unit 4 displays an operation state of the facsimile multifunction device 20, displays a screen for prompting an operation input to the user, or reads an image of a document read for transmission. Data, image data transmitted from other facsimile machines and PCs, etc. are displayed.

  The operation unit 5 includes character keys, numeric keys, abbreviated dial keys, one-touch dial keys, various function keys, and the like necessary for operating the facsimile multifunction device 20. The operation unit 5 includes a rotation processing setting key 5a for setting whether or not to perform image data rotation processing using the page memory 10 and the buffer 11 as function keys. In addition, the display part 4 can substitute for a part or all of the various keys of the operation part 5 by making the display part 4 into a touch panel system. The ROM 6 stores in advance various software programs necessary for the operation of the facsimile composite apparatus 20. The RAM 7 is composed of SRAM, flash memory, etc., and stores temporary data generated when software is executed. The codec unit 8 encodes and compresses the image data and decodes the encoded and compressed image data. The image memory 9 stores image data obtained by reading and encoding a document.

  The page memory 10 is composed of SDRAM, and stores one page of image data read by the reading unit 2 and one page of image data received from another facsimile apparatus or PC. The image data for one page accumulated and developed in the page memory 10 is read to the recording unit 3, and an image corresponding to the read image data is recorded. The buffer 11 has 24 flip-flops for every 16 pixels configured in a 4 × 4 matrix, and multi-valued data consisting of 8 flip-flops is converted into R, G, and B data. Hold as value. Data can be written or read between the page memory 10 and the buffer 11 with 4 × 4 16-pixel data of the page memory 10 as a unit, and the above-described data writing can be performed as necessary. Alternatively, transposed image data in the image data rotation process is generated by performing read control.

  The modem 12 is connected to the bus 15 and is a facsimile modem capable of facsimile communication. The modem 12 is also connected to the NCU 13 connected to the bus 15. The NCU 13 is hardware that performs operations of closing and opening the line L1 with a public switched telephone network (PSTN), and connects the modem 12 to the PSTN as necessary. The facsimile composite apparatus 20 is connected to other facsimile apparatuses by PSTN so that normal facsimile communication can be performed. The PC interface unit 14 is connected to an external PC via, for example, a LAN communication line L2, and transmits / receives data to / from the PC. Various image data created and edited by an external PC are transmitted to the facsimile multifunction apparatus 20 (PC interface unit 14) via the communication line L2, and an image corresponding to the transmitted image data is transmitted to the facsimile multifunction apparatus 20. The data is printed out on a sheet at (recording unit 3).

  The operation of the recording process in the facsimile multifunction apparatus 20 having the above-described configuration will be described below. FIG. 2 is a flowchart showing an operation procedure of recording processing in the image processing apparatus according to the embodiment of the present invention, and FIGS. 3 to 5 are diagrams for explaining 90-degree rotation processing of image data in the clockwise direction. FIG.

  As shown in FIG. 2, one page of image data read by the reading unit 2 or image data received from another facsimile machine or PC is stored in the page memory 10 (step S1).

  Next, the control unit 1 determines whether or not the rotation process is set based on whether or not the rotation process setting key 5a is pressed (step S2). When rotation processing is not set (step S2: NO), image data is read from the page memory 10 to the recording unit 3 in the same order as written in the page memory 10 (step S8), and corresponds to the read image data. The image to be printed is printed out on a sheet by the recording unit 3 (step S9).

  When rotation processing of image data is set (step S2: YES), 16 pixels of 4 × 4 square shape in the image data stored in the page memory 10 are set as one unit (block), and Data of one block is read from the page memory 10 and written to the buffer 11 (step S3). In the writing process to the buffer 11, data is written in the row direction of the buffer 11. After the write processing to the buffer 11 is completed, the data written to the buffer 11 is read from the column direction, and the read data is written to the original 4 × 4 memory area of the page memory 10 (step S4). Thereby, transposed image data is generated as intermediate data for the rotation process. The control unit 1 determines whether or not transposed image data generation with 4 × 4 pixels as a unit over one page stored in the page memory 10 is completed (step S5). (Step S5: NO), the operation returns to Step S3 to generate transposed image data in the next block. When transposition image data generation is completed over one page (step S5: YES), the operation proceeds to step S6.

  FIG. 3 is a diagram showing a first data accumulation state in the page memory 10. As shown in FIG. 3, the image data for one page is divided into a total of six blocks, three in the row direction and two in the column direction. FIG. 4 is a diagram showing transposed image data obtained by reading or writing data with the buffer 11. The hatched portion in FIGS. 3 and 4 indicates where the data read before transposition is written after transposition.

  Next, the read address is controlled, the transposed image data is read from the page memory 10 to the recording unit 3 (step S6), and the rotated image corresponding to the read transposed image data is printed out on a sheet by the recording unit 3 (step S6). S7). Reading of each line from the page memory 10 is performed in the order described in the leftmost column of FIG. 4, and the lines are alternately read out by two blocks. By controlling the read address, an image rotated 90 degrees clockwise as shown in FIG. 5 is obtained.

  By the way, in the present embodiment, transposed image data is generated with 4 × 4 pixels as a unit. Therefore, after storing all the image data for one page in the page memory 10 is completed, It is not necessary to start the generation, and it is possible to start generating the transposed image data every time the four lines of data are stored. FIG. 6 is a flowchart showing the operation procedure of the rotational recording process, and FIG. 7 is a diagram showing a data write state of the page memory 10. In FIG. 7, hatched areas with a wide pitch indicate image data before transposition, and hatched areas with a narrow pitch indicate transposed image data. Note that the following description will be given by taking the case of using color image data as an example.

  In FIG. 6, when accumulation of image data in the page memory 10 is started (step S61, FIG. 7A), the control unit 1 determines whether image data for four lines has been accumulated (step S62). ). When four lines have been accumulated (step S62: YES, FIG. 7B), the block memory consisting of 16 pixels of 4 × 4 square shape accumulated in the page memory 10 is read from the page memory 10. Reading and writing in the row direction of the buffer 11 (step S63), then reading the data written in the buffer 11 from the column direction and writing it into the original 4 × 4 memory area of the page memory 10 (step S64). Thus, transposed image data is generated (FIG. 7C). In this case, the image data accumulation processing in the subsequent lines is performed in parallel with the transposed image data generation processing.

  The control unit 1 determines whether or not the transposed image data generation has been completed for one page (step S65), and if not completed (step S65: NO), the operation returns to step S62, and similarly. Transposed image data in the next four lines is generated (FIGS. 7D and 7E). When the generation of the transposed image data is completed over one page (step S65: YES, FIG. 7 (f), FIG. 7 (g)), the transposed image data is read by controlling the read address (step S66, FIG. 7 ( h)), an image rotated 90 degrees clockwise is printed out (step S67).

  As a result, the image data accumulation process in the page memory 10 and the transposed image data generation process can be performed in parallel, and therefore a delay from the completion of the accumulation of the image data for one page to the start of reading the transposed image data. The time can be reduced to the time required for generating the transposed image data for one page, and the entire time required for the rotation processing of the image data can be greatly reduced.

  Next, address control of the page memory 10 using SDRAM will be described. FIG. 8 is a timing chart showing a data read operation in the SDRAM. FIG. 9 is a timing chart showing a data write operation in the SDRAM. 8 and 9, / ACK is a DMA transfer permission signal, / RAS is a row address strobe, / CAS is a column address strobe, / WE is a write permission signal, BS is a bank selection signal, and A10 is an address. A bus signal, ADD indicates a read address or write address signal, Din indicates an input, and Dout indicates an output. As shown in FIGS. 8 and 9, burst length is 4 words (4 pixels), and burst reading or writing is performed with 4 words as a set. When burst read and burst write of a set of 4 words of data is completed, the next set of burst read or burst write performs address control to access different banks.

  10 (a), 10 (b), 11 (c), and 10 (d), k is an integer, and one line is composed of 4k words, (4k + 1) words, (4k + 2) words, and (4k + 3) words. FIG. 5 is a diagram showing a relationship between an address and a bank when burst reading or burst writing is performed with 4 words as one set. When one line is composed of 4k words, that is, when the number of words in one line is a multiple of four, as shown in FIG. Does not generate a fractional address at the rear end of the line. However, when one line is not composed of 4k words, that is, when the number of words in one line is not a multiple of 4, as shown in FIGS. 10 (b), 11 (c), and 11 (d), 4 When a burst read or burst write is performed with a set of words, a fractional address is generated at the rear end of the line. Therefore, when burst reading or burst writing is performed, data across a plurality of lines cannot be read or written correctly.

  Therefore, when one line is composed of other than 4k words, a predetermined number of invalid addresses with no actual data are added at the end of each line. That is, a predetermined number of invalid addresses are added at the end in order to make the number of words in each line a multiple of four. 10 (b), 11 (c), and (d), respectively, when one line is composed of (4k + 1) words, (4k + 2) words, and (4k + 3) words, It is a figure which shows the relationship between an address at the time of adding the invalid address of 3 words, 2 words, and 1 word, and a bank. As is clear from FIGS. 10B, 11C, and 11D, by adding an invalid address of 3 words, 2 words, and 1 word to the end, burst reading for the next line or When burst writing is performed, it is always possible to start from the top address of the bank, and burst access with a set of 4 words can be performed correctly.

  Next, a case where transposed image data is generated will be described. When generating transposed image data, burst writing or burst reading of data is performed not in the horizontal direction (main scanning direction) but in the vertical direction (sub-scanning direction). 12 (a), 12 (b), 13 (c), and (d), p is an integer, and one line is 4p-3 to 4p words, 4 (p + 1) -3 to 4 (p + 1) words, Address when 4 bits (p + 2) -3 to 4 (p + 2) words and 4 (p + 3) -3 to 4 (p + 3) words are used for burst writing or burst reading with 4 words as a unit It is a figure which shows the relationship between a bank. When one line is composed of 4 (p + 1) -3 to 4 (p + 1) words or 4 (p + 3) -3 to 4 (p + 3) words, the access banks in the first column are 0, 1, 2, 3,... Or 0, 3, 2, 1,..., And continuous bank access is possible. On the other hand, when one line is composed of 4p-3 to 4p words or 4 (p + 2) -3 to 4 (p + 2) words, the access banks in the first column are 0, 0, 0, 0,... Or 0, 2, 0, 2,..., And continuous bank access cannot be performed.

  Therefore, when one line is composed of 4p-3 to 4p words or 4 (p + 2) -3 to 4 (p + 2) words, there are four actual data at the end of each line, that is, one bank. Add the address. That is, in order to make each line have an odd number of banks, an invalid address for one burst access in which no actual data exists is added at the end. FIGS. 14 (a) and 14 (b) show one bank at the end of each line when each line is composed of 4p-3 to 4p words or 4 (p + 2) -3 to 4 (p + 2) words. It is a figure which shows the relationship between an address and a bank at the time of adding the invalid address of minutes (for 4 words). As is clear from FIGS. 14A and 14B, by adding an invalid address for one bank (for four words) at the end, the access bank in the first column becomes 0, 1, 2, .., Or 0, 3, 2, 1,..., And it becomes possible to perform continuous bank access.

  Next, a case where the generated transposed image data is read from the page memory 10 as rotated image data will be described. FIG. 15 is a diagram showing banks in each line of each block. Since transposed image data is generated with 4 × 4 pixels as a unit, in FIG. 15, 1 set of 1 line of 1 block, 1 set of 1 line of 2 blocks, 1 set of 1 line of 3 blocks, .., 1 set of 1 line of the last block, 1 set of 2 lines of 1 block, 2 sets of 2 lines of 1 block, ..., 2 sets of 1 line of the last block, ... 4 of 1 block Access is made in the order of the first set of lines, the first set of four lines of the last block, the first set of one line of one block, the second set of four lines of the last block. At this time, the banks to be accessed are 0, 0, 0,..., 0 in the first set of lines of each block, 1, 1, 1,... Therefore, since the same bank is circulated, continuous bank access cannot be performed.

  Therefore, one set (one bank worth) of invalid addresses is added at the end of the fourth line (final line) in each block. That is, an invalid address for one bank, which is a unit of burst reading or burst writing, is added to the end of the fourth line (final line) (the address indicated by the black arrow in FIG. 16). FIG. 16 is a diagram illustrating a relationship between an address and a bank when an invalid address for one bank (4 words) is added to the end of the fourth line (final line). FIG. 16 shows an example in which one line is composed of 4q-3 to 4q words (q is an integer), and an invalid address for one bank is added at the end of the fourth line of each block. . Since one line is composed of 4q-3 to 4q words, an invalid address for one bank is provided at the end of each line as described above, and an invalid address for two banks (8 words) is provided for the fourth line. Will be provided. Although the case where one line is composed of (4k + 1) words, (4k + 2) words, and (4k + 3) words is not shown in the drawing, the same as the case where one line is composed of 4k words, the fourth line of each block. Finally, an invalid address for one bank is added.

  By adding invalid addresses for one bank to the 4th line (final line) in this way, as shown in FIG. 17, 1 block 1 line 1 set, 2 blocks 1 line 1 set, 3 blocks In this order, the access banks are 0, 1, 2,... In the order of the first set of the first line,..., And the bank circulates in different banks, thereby enabling continuous bank access.

  In the above-described embodiment, the case where the transposed image data is generated with 4 × 4 pixels as one unit is described. However, the present invention is not particularly limited to 4 × 4 pixels, and for example, 8 × 8 pixels are used. The value of n may be arbitrary as long as the unit is a square pixel block.

  In this embodiment, the facsimile complex apparatus is described as an example. However, it goes without saying that the present invention can be applied to all apparatuses that rotate and output acquired image data, such as a facsimile apparatus, a digital copying machine, and a printer. Yes.

1 is a block diagram illustrating a configuration of an image processing apparatus (facsimile multifunction apparatus) according to the present invention. 5 is a flowchart showing an operation procedure of recording processing in the image processing apparatus according to the embodiment of the present invention. It is a figure for demonstrating the 90 degree rotation process to the clockwise direction of image data. It is a figure for demonstrating the 90 degree rotation process to the clockwise direction of image data. It is a figure for demonstrating the 90 degree rotation process to the clockwise direction of image data. It is a flowchart which shows the operation | movement procedure of a rotation recording process. It is a figure which shows the data write state of a page memory. 6 is a timing chart showing a data read operation in the SDRAM. 5 is a timing chart showing a data write operation in the SDRAM. It is a figure which shows the relationship between an address and a bank in case one line is comprised by 4k word or (4k + 1) word. It is a figure which shows the relationship between an address and a bank in case one line is comprised by (4k + 2) word or (4k + 3) word. It is a figure which shows the relationship between an address and a bank in case 1 line is comprised by 4p-3-4p word or 4 (p + 1) -3-4 (p + 1) word. It is a figure which shows the relationship between an address and a bank in case 1 line is comprised by 4 (p + 2) -3-4 (p + 2) word or 4 (p + 3) -3-4 (p + 3) word. It is a figure which shows the relationship between an address at the time of adding the invalid address in each line in this invention, and a bank. It is a figure which shows the bank in each line of each block. It is a figure which shows the relationship between the address and bank at the time of adding the invalid address for 1 bank to the tail end of the 4th line. It is a figure which shows that a continuous bank access is possible.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Control part 2 Reading part 3 Recording part 5 Operation part 5a Rotation process setting key 6 ROM
7 RAM
10 page memory 11 buffer

Claims (2)

Burst read or burst write across each line of the memory in an image processing method for performing burst read or burst write with a plurality of pixels as one bank for one page of image data expanded two-dimensionally in a memory When determining whether the bank number of the first bank of each line of the memory is continuous, if it is determined that the bank number is not continuous, there is real data at the end of each line of the memory. When an unused address is added for one bank , a plurality of lines corresponding to the number of pixels in one row developed in the memory are set as one block, and burst reading or burst writing is performed across the block. When it is determined that the bank numbers of the first bank of each line of the memory are not consecutive, at the end of one block of the memory Image processing method characterized by a free address adding one bank having no data. In an image processing apparatus comprising: a memory for developing and storing image data in two dimensions; and means for performing burst reading or burst writing with a plurality of pixels as one bank for the image data stored in the memory, when performing a burst read or burst write over each line of said memory, a first determining means for determining whether the bank number of the first bank of each line of the memory is contiguous, the first determining means When it is determined that the bank numbers are not continuous, first addition means for adding one bank of empty addresses not having real data at the end of each line of the memory, and one unit to be developed in the memory When a plurality of lines corresponding to the number of pixels in the row direction are set as one block, and burst reading or burst writing is performed across the block, each line of the memory is recorded. Second determining means for determining whether or not the bank numbers of the first bank of the bank are consecutive, and when the second determining means determines that the bank numbers are not consecutive, the end of one block of the memory An image processing apparatus comprising: a second adding unit that adds one bank of empty addresses that do not have actual data .

Images