JP4790545B2 - Image processing apparatus and image processing method - Google Patents

Description

  The present invention relates to an image processing apparatus and an image processing method.

  Conventionally, there is a technique for dividing an input image, processing each divided image, and outputting the processed image. For example, Patent Document 1 discloses a technique for processing an image with a small amount of data obtained by dividing an image in order to reduce the storage capacity of a built-in memory used when processing input image data. Is disclosed. Patent Document 2 discloses a technique for controlling an offset of an external memory address in input / output control of divided images.

  An example of such a hardware configuration in the prior art includes a CPU (Central Processing Unit), a DMAC (Direct Memory Access Controller), an SDRAM (Synchronous DRAM), a built-in buffer, and a signal processing circuit as shown in FIG. It has a configuration. In this configuration, when data is transferred from the SDRAM to the built-in buffer, according to the conventional technique, as shown in FIG. 16, the image indicated by the image data stored in the SDRAM is horizontally rectangular (striped). After the division width (hereinafter referred to as strip width) of the divided image (hereinafter referred to as strip width) is determined, the DMAC performs addressing in the SDRAM.

In this case, the CPU needs to make various settings for the DMAC. First, the number of transfers (burst width and the number of block column bursts) in one line at the input / output strip width is set. Next, an address width (line offset) for jumping after the transfer of one line is completed is set. Further, the number of blocks indicating the number of vertical divisions of the strip and the number of block rows indicating the number of rows in the block are set. Also, an address width (block offset) for jumping after the transfer of a predetermined number of block rows is completed is set. Then, an address width (line offset) that jumps when transfer of one strip is completed is set.
JP 2002-304624 A Japanese Patent Laid-Open No. 2004-220484

  As described above, in the conventional technique for transferring image data when an image is divided and processed, the setting performed by the CPU is complicated.

  In view of the above problems, an object of the present invention is to provide an image processing apparatus and an image processing method that simplify settings performed by a CPU.

In order to achieve the above object, the invention of claim 1 functions as a work area capable of storing input image data and output image data, and displays a two-dimensional image indicated by the input image data and the output image data on an XY plane. the image above, the coordinates in the XY plane, a first storage means and the address of the memory space is associated one-to-one at the work area, the predetermined signal processing on the input image data A second storage means used when applying, a dividing means for dividing the two-dimensional image indicated by the input image data into a plurality of two-dimensional images in the X direction in the work area, and the two-dimensional obtained by the division first setting means and a first start address of the work area for setting the first position coordinates indicating a position at the work area corresponding to the pixel of the image data of the divided image, the X-direction width size dimension divided images, the pixel size of one pixel in the image data of the two-dimensional image segment and on the basis of the start coordinates of the two-dimensional image segment corresponding to the first coordinates in which the set A first address generation unit configured to generate an address in the memory space; and the input image data stored in the first storage unit according to the address generated by the first address generation unit First transfer means for transferring each divided image to the second storage means; signal processing means for performing the signal processing on the image data of the two-dimensional divided image transferred to the second storage means; A second setting means for setting a second position coordinate indicating a position in the work area corresponding to a pixel of the image data subjected to the signal processing; and a second head address of the work area. The width size in the X direction of the two-dimensional image indicated by the image data subjected to the signal processing, the pixel size of one pixel in the image data subjected to the signal processing, and the image data subjected to the signal processing A second address generation unit configured to generate an address in the memory space corresponding to the set second position coordinate based on the leading coordinates of the two-dimensional image shown; and generated by the second address generation unit According to the address, the image data subjected to the signal processing is transferred as the output image data from the second storage means to the first storage means, and at the time of the transfer, transferred to the second storage means The input image data transferred to the second storage means next to the last input image data that has been processed and the output image data that has been subjected to the signal processing and transferred to the first storage means Second transfer that shifts the input image data to be transferred next in the X direction in the work area so as not to overlap in the work area, and overwrites the output image data subjected to the signal processing in the shifted trace area. Means .

Here, in the present invention, when transferring from the second storage means to the first storage means, it is transferred to the second storage means next to the last image data before signal processing transferred to the second storage means. The input image data to be transferred next is X direction in the work area so that the input image data to be output and the output image data that has been subjected to signal processing and transferred to the first storage means do not overlap in the work area. To provide an image processing apparatus capable of greatly reducing the memory area used in the DMA transfer of image data because the output image data subjected to signal processing is overwritten on the shifted trace area. Can do.

On the other hand, in order to achieve the above object, the invention of claim 2 functions as a work area capable of storing the input image data and the output image data, and the two-dimensional image indicated by the input image data and the output image data. the image on the XY plane, the coordinates in the XY plane, the method comprising the address of the memory space providing a first storage means associated with one-to-one at the work area, in advance in the input image data A step of preparing a second storage means to be used when performing predetermined signal processing, and a division step of dividing a two-dimensional image indicated by the input image data into a plurality of two-dimensional images in the X direction in the work area; a first setting step of setting the first position coordinates indicating a position at the work area corresponding to the pixel of the image data of the two-dimensional division images obtained by the divided, the The first head address of work area, the X-direction width size of the two-dimensional image segment pixel size of one pixel in the image data of the two-dimensional image segment and on the basis of the start coordinates of the two-dimensional image segment wherein A first address generation stage for generating an address in the memory space corresponding to the set first position coordinate, and the first storage means according to the address generated in the first address generation stage. A first transfer step of transferring the stored input image data to the second storage means for each of the two-dimensionally divided images; and image data of the two-dimensional divided images transferred to the second storage means A signal processing stage for performing the signal processing on the second stage, and a second setting stage for setting a second position coordinate indicating a position in the work area corresponding to a pixel of the image data subjected to the signal processing. A second head address of the work area, a width size in the X direction of the two-dimensional image indicated by the image data subjected to the signal processing, a pixel size corresponding to one pixel in the image data subjected to the signal processing, and A second address generation step of generating an address in the memory space corresponding to the set second position coordinate based on the leading coordinates of the two-dimensional image indicated by the image data subjected to the signal processing; According to the address generated in the second address generation stage, the image data subjected to the signal processing is transferred as the output image data from the second storage means to the first storage means, and the transfer In this case, the input image data transferred to the second storage means next to the last input image data transferred to the second storage means, and the first storage hand subjected to the signal processing. The input image data to be transferred next is shifted in the X direction in the work area so that the output image data transferred to the stage does not overlap in the work area, and the signal processing is applied to the shifted trace area. A second transfer stage for overwriting the output image data .

Here, since the invention of claim 2 operates in the same manner as the invention of claim 1, the same effect as the invention of claim 1 can be obtained.

  According to the present invention, it is possible to provide an image processing apparatus and an image processing method that simplify the settings performed by the CPU.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. Note that the image processing apparatus according to the present embodiment is assumed to be mounted on a device that handles image data, such as a digital camera.

  FIG. 1 is a diagram illustrating a hardware configuration of an image processing apparatus according to the present embodiment. The image processing apparatus shown in the figure includes an address generation circuit 10, a built-in buffer 12, a DMAC 14, an SDRAM 16, a signal processing circuit 18, and a CPU 20. The CPU 20, the DMAC 14, the SDRAM 16, the address generation circuit 10, and the built-in buffer 12 are electrically connected to each other via a bus. Further, the signal processing circuit 18 is electrically connected to the built-in buffer 12, and thereby image data to be subjected to signal processing (hereinafter referred to as input data) and image data subjected to signal processing. Exchange of processed image data (hereinafter referred to as output data) between the SDRAM 16 and the built-in buffer 12 is possible. The input / output data used in the following description indicates input data and output data.

  Next, the flow of input / output data will be described. In the image processing apparatus according to this embodiment, the DMAC 14 transfers the input data stored in the SDRAM 16 to the built-in buffer 12, and the signal processing circuit 18 performs image processing on the input data. The data is again transferred to and stored in the SDRAM 16 by the DMAC 14.

  Next, each detail in the said structure is demonstrated. First, the SDRAM 16 stores input / output data as described above, and the position of the input / output data on the SDRAM can be expressed using coordinates corresponding to addresses.

  The address generation circuit 10 generates an input / output data address to be notified to the DMAC 14. Specifically, the address generation circuit 10 generates an address in the SDRAM 16 corresponding to the coordinates from the coordinates indicating the position of the input data that the DMAC 14 transfers to the internal buffer 12. Similarly, the address generation circuit 10 generates an address in the SDRAM 16 in which the DMAC 14 stores output data. Therefore, an address output circuit, which will be described later, provided in the address generation circuit 10 is provided for each input / output data. Details of the address generation method in the address generation circuit 10 will be described later.

  The built-in buffer 12 temporarily stores the input / output data and has a function for exchanging with the DMAC 14.

  The DMAC 14 transfers input data from the SDRAM 16 to the internal buffer 12 and transfers output data from the internal buffer 12 to the SDRAM.

  Next, an address generation method will be described with reference to FIG. As shown in the figure, the address generation circuit 10 receives an address corresponding to the coordinates (x, y) when the horizontal image size, the one-pixel data size, the coordinates (x, y), and the image head address are input. Is included. Here, the coordinates in the present embodiment are coordinates on the XY plane, and are for expressing the two-dimensional image indicated by the image data as an image on the XY plane, as shown in FIG. This coordinate is also a coordinate indicating the position of the image data in the SDRAM 16, and is a coordinate whose direction of the Y axis is opposite to that of the normal XY coordinate. These coordinates are used in common for input / output data. That is, since all the input / output data is stored on the SDRAM 16, it can be expressed on the same coordinates.

  Therefore, the coordinates (x, y) are the coordinates indicating the position on the SDRAM 16 where the DMAC 14 transfers the input data to the internal buffer 12, and the coordinates indicating the position on the SDRAM 16 where the DMAC 14 transfers the output data from the internal buffer 12. It is. The horizontal image size is the size of the image in the X-axis direction in bytes, as shown in FIG.

  The 1-pixel data size is the size of one pixel in bytes (Byte Per Pixel). The one-pixel data size may vary depending on input / output data. Further, the image head address indicates the head address of input / output data as shown in FIG.

  The address output circuit includes multipliers 11a and 11b and adders 13a and 13b. The multiplier 11a multiplies the horizontal image size by y and outputs the obtained value to the adder 13a. The adder 13a adds the value output from the multiplier 11a and x, and outputs the value obtained thereby to the multiplier 11b. The multiplier 11b multiplies the value output from the adder 13a and the one-pixel data size, and outputs the value obtained thereby to the adder 13b. The adder 13b adds the value output from the multiplier 11b and the image head address, and outputs the value obtained thereby as an address corresponding to the coordinates (x, y).

  Therefore, when the horizontal image size is L, the pixel data size is M, and the image head address is A, the address output circuit is a circuit for obtaining an address (Ly + x) M + A.

  The input / output data is transferred for each strip obtained by dividing the image in the horizontal direction as shown in FIG. Since the input / output data is transferred based on the output of the address output circuit, as shown in FIG. 4, it is transferred from the start 1 which is the top of the strip 1 to end 1 while returning one line at a time, and then the top of the strip 2 The transfer is performed for each of the divided strips, such as the same transfer from start2 to end2. The horizontal image size L in this case indicates the size (number of bytes) of the horizontal width of each strip.

  Details of the transfer processing described above will be described with reference to the flowchart of FIG. The process shown in FIG. 5 is a process executed by the CPU 20 and the address generation circuit 10.

  First, in step 101, the CPU 20 sets image information in the DMAC 14. Here, the image information is the leading address of the input / output data, the horizontal width of the strip in the input / output data, and the one pixel data size in the input / output data. Further, in step 102, the CPU 20 sets the strip leading coordinates (x, y) of the input / output data in the address generation circuit 10.

  In the next step 103, the address generation circuit 10 performs the calculation by the address output circuit described with reference to FIG. 2 on the basis of the strip leading coordinates (x, y), the horizontal image size, and the one-pixel data size. 12, the transfer source address corresponding to the coordinates (x, y) of the input data transferred from the SDRAM 16 and the transfer destination address corresponding to the coordinates (x, y) for storing the output data in the SDRAM 16 from the built-in buffer 12 are obtained.

  Next, the address generation circuit 10 notifies the DMAC 14 of the leading address of the input / output data in step 104, so that DMA transfer is performed in step 105.

  Here, as shown in FIG. 6, the handshake between the built-in buffer 12 and the address generation circuit 10 with the DMAC 14 first notifies the DMAC 14 of the address described above, and then the built-in buffer 12 makes a data request. The built-in buffer 12 receives image data from the DMAC 14.

  Returning to the flowchart of FIG. 5, when the transfer of the input / output data is completed, the address generation circuit 10 increments the X coordinate in the input / output data by one in step 106 in order to transfer the next input / output data. Next, in step 107, the address generation circuit 10 determines whether or not the incremented X coordinate matches the value obtained by subtracting 1 from the top X coordinate of the next strip. This determination is for determining whether or not the transfer has been completed to the right end of the strip.

  If the address generation circuit 10 makes a negative determination in step 107, the process of step 104 is executed again. If the address generation circuit 10 makes an affirmative determination, the address generation circuit 10 increments the Y coordinate in the input / output data by one in step 108, and sets the X coordinate to the first X coordinate of the strip in step 109.

  Next, the address generation circuit 10 determines in step 110 whether the Y coordinate of the output data transferred to the SDRAM 16 matches the vertical image size. Here, the vertical image size indicates the size of the strip in the Y-axis direction in bytes. If the address generation circuit 10 makes a negative determination in step 110, the process of step 103 is executed again. If the address generation circuit 10 makes an affirmative determination, the CPU 20 determines in step 111 whether processing for all strips has been completed. If the CPU 20 makes a negative determination in step 111, the object to be transferred in step 112 is moved to the next strip in the input data, and the process of step 102 is executed again. If the CPU 20 makes a positive determination, the process ends.

  Note that in a digital camera, the image data 24 shown in FIG. 7 is input data, and the image data 26 excluding a hatched area (so-called margin) from the input data is often output data. In this case, as shown in the figure, the size of the strip in the X-axis direction (hereinafter referred to as strip width) is different between the input data and the output data. This margin is used for the image processing using the peripheral pixels of the strip edge portion (outer peripheral portion) pixels in the signal processing circuit 18.

  The process shown in the flowchart of FIG. 5 described above is a basic process, and a process in which this process is associated with a work area will be described. As shown in FIG. 8, the processing corresponding to the work area stores image data 32 as input data in the work area 30, and obtains an address using coordinates with the start address of the work area as the origin. It is processing.

  An address generation method in this case will be described with reference to FIG. As shown in the figure, the address generation circuit 10 receives a work area horizontal size, a pixel data size, coordinates (x, y), image data start coordinates (p, q), and a work area start address. And an address output circuit for outputting an address corresponding to the coordinates (x, y).

  The coordinates (x, y) are the coordinates indicating the position on the SDRAM 16 where the DMAC 14 transfers the input data to the built-in buffer 12 in the coordinates shown in FIG. 8, and the DMAC 14 outputs the output data from the built-in buffer 12. It is a coordinate which shows the position on SDRAM16 to transfer. The work area horizontal size is the size of the work area in the X-axis direction in bytes, as shown in FIG. Also, the work area head address indicates the head address of the work area (in FIG. 8, 0x002A0000: hexadecimal notation), as shown in FIG.

  The address output circuit includes adders 15a, 15b, 15c and 14d and multipliers 17a and 17b. The adder 15a adds q and y, and outputs the value obtained thereby to the multiplier 17a. The multiplier 17a multiplies the value output from the adder 15a by the work area horizontal size, and outputs the value obtained thereby to the adder 15c.

  On the other hand, the adder 15b adds p and x, and outputs the value obtained thereby to the adder 15c. The adder 15c adds the value output from the multiplier 17a and the value output from the adder 15b, and outputs the value obtained thereby to the multiplier 17b. The adder 15d adds the value output from the multiplier 17b and the work area head address, and outputs the value obtained thereby as an address corresponding to the coordinates (x, y).

  Therefore, when the work area horizontal size is K, the pixel data size is M, and the image head address is A, the address generation circuit 10 obtains an address (K (y + q) + x + p) M + A as shown in FIG. Circuit. The obtained value indicates an address corresponding to the coordinates (x, y).

  By corresponding to the work area in this way, the SDRAM 16 can be used efficiently as shown in FIG. FIG. 10 shows output data 34, input data 36, and strips 38 and 40. As shown in the figure, the output data 34 and the input data 36 overlap, but this can be overwritten because the strip 38 of the output data 34 does not overlap the strip 40 of the input data 36 to be transferred next. It has become. In fact, as shown in FIG. 10, each strip of output data 34 does not overlap each strip of input data 36 to be transferred next.

  With this processing, as can be seen from comparison with the memory map in the case of not overwriting shown in FIG. 11, the amount of use of the SDRAM 16 can be greatly reduced.

  The processing corresponding to the work area described above is the same as the image information set in step 101 in the flowchart described with reference to FIG. The strip size and the one-pixel data size of the input / output data are used, and the address generation processing described in FIG.

  Next, a configuration for realizing high-speed processing will be described. FIG. 12 shows a configuration in which an output width counter 42 is further added to the configuration of the image processing apparatus described in FIG. The output width counter 42 is integrated with the address generation circuit 10, and information such as strip leading coordinates set in the address generation circuit 10 can be referred to by the output width counter 42.

  The output width counter 42 outputs the X coordinate of the strip leading coordinates of the input / output data to the address generation circuit 10. The configuration of the output width counter 42 will be described with reference to FIG.

  As shown in the figure, the output width counter 42 has an X coordinate that is the X coordinate of the leading coordinate of the output data, an X coordinate that is the X coordinate of the leading coordinate of the input data, an output strip width that is the strip width of the output data, and an output The horizontal width of the output data indicating the length of the data in the X-axis direction in bytes and the input shift amount are input, and the X coordinate of the strip start coordinates of the input data and output data is output.

  The output width counter 42 includes selectors 41a and 41b, adders 43a and 43b, and a comparator 45. The selector 41a outputs either X1 or the output of the adder 43a, and outputs the value input from the adder 43a after first outputting X1.

  The adder 43 a adds the output from the selector 41 a and the output strip width in the output data, and outputs the result to the address generation circuit 10 and the comparator 45. The comparator 45 compares the output data horizontal width with the output from the adder 43a. If the value output from the adder 43a is equal to or larger than the output horizontal width, the comparator 45 outputs a signal indicating that the processing is completed.

  On the other hand, the selector 41b outputs either X2 or the output of the adder 43b, and outputs the value input from the adder 43b after first outputting X2.

  The adder 43 b adds the output from the selector 41 b and the input shift amount in the input data, and outputs the result to the address generation circuit 10.

  Details of the transfer process using the output width counter 42 will be described with reference to the flowchart of FIG. The process shown in FIG. 14 is a process executed by the CPU 20, the address generation circuit 10, and the output width counter 42.

  First, in step 201, the CPU 20 sets image information in the DMAC 14. The image information here is the work area start address, the start coordinates of the input / output data, the horizontal width of the strip in the input / output data, the one pixel data size in the input / output data, the input shift width, and the strip of output data. Width.

  Further, the CPU 20 sets the strip leading coordinates (x, y) in the address generation circuit 10 in step 202. In the next step 203, the address generation circuit 10 performs the transfer source address corresponding to the coordinates (x, y) of the input data by the calculation by the address output circuit described with reference to FIG. And a transfer destination address corresponding to the coordinates (x, y) for storing the output data. Next, the address generation circuit 10 notifies the DMAC 14 of the leading address of the input / output data in step 204, so that DMA transfer is performed in step 205.

  When the transfer of the input / output data is completed, in order to transfer the next input data, in step 206, the address generation circuit 10 increments the X coordinate in the input / output data by one. Next, in step 207, the address generation circuit 10 determines whether or not the incremented X coordinate matches the value obtained by subtracting 1 from the top X coordinate of the next strip. This determination is for determining whether or not the transfer has been completed to the right end of the strip.

  If the address generation circuit 10 makes a negative determination in step 207, the process of step 204 is executed again. If the address generation circuit 10 makes an affirmative determination, the address generation circuit 10 increments the Y coordinate in the input / output data by one in step 208, and sets the X coordinate to the first X coordinate of the strip in step 209.

  Next, the address generation circuit 10 determines in step 210 whether the Y coordinate of the output data transferred to the SDRAM 16 matches the vertical image size. If the address generation circuit 10 makes a negative determination in step 210, the process of step 203 is executed again. When the address generation circuit 10 makes an affirmative determination, the output width counter 42 again sets X1 plus the output strip width to X1 in step 211, and X2 plus X2 plus the input shift width.

  Next, the output width counter 42 determines in step 212 whether the size of the output data matches the output data horizontal width size. This determination is made by the comparator 45 described above, and is a determination of whether all output data has been stored in the SDRAM 16. If the affirmative determination is made in step 212, the comparator 45 outputs a signal to the CPU 20, and thus the processing ends.

  On the other hand, if the comparator 45 makes a negative determination in step 212, the process of step 202 is executed again.

  Comparing the processing in FIG. 14 described above with the processing described in FIG. 5 described above, the CPU 20 performs processing in step 111 and step 112 in FIG. 5, but these processing is performed by hardware in FIG. Since an address generation circuit 10 and an output width counter 42 execute, processing can be executed at high speed.

  As described above, in the present embodiment, the two-dimensional image indicated by the image data is an image on the XY plane, and the coordinates in the XY plane and the first storage means (SDRAM 16) capable of storing the image data are used. There is a one-to-one correspondence with an address in the memory space, a first head address that is a head address in the first storage means of the image data, and the two-dimensional image on the XY plane. Based on the set value of the first width size that is the size with respect to the X-axis direction and the first pixel size that is the size of one pixel in the image data, the memory space corresponding to the coordinates from the set coordinates First address generation means (address generation circuit 10) for generating the first address, the first head address, the first width size, and the first pixel size Setting means (CPU 20) for setting the first address generation means, and coordinate setting means (CPU 20) for setting the coordinates where the two-dimensional image exists on the XY plane for the first address generation means. ) And the first data used for performing predetermined signal processing on the image data stored in the first storage unit by the first pixel size from the address generated by the first address generation unit. Transfer means (DMAC 14) for transferring to two storage means (built-in buffer 12).

  In the present embodiment, when the processed image data, which is the image data subjected to the signal processing, is transferred from the second storage means (built-in buffer 12) to the first storage means (SDRAM 16). A second start address that is a start address in the first storage means in which the transferred processed image data is stored, and the X-axis direction of the two-dimensional image indicated by the processed image data on the XY plane On the basis of a set value of a second width size that is a size for and a second pixel size that is a size of one pixel in the processed image data, in the memory space corresponding to the coordinates from the set coordinates Second address generation means (address generation circuit 10) for generating an address is further included, and the setting means (CPU 20) is configured to provide the second head address, the second address And the second pixel size are further set for the second address generation means, and the coordinate setting means (CPU 20) is configured to display the two-dimensional image indicated by the processed image data on the XY plane. Is set for the second address generation means, and the transfer means (DMAC 14) sets the second pixel size from the address generated by the second address generation means by the second pixel size. The processed image data stored in the storage unit is transferred to the first storage unit.

  In the present embodiment, the image data is stored in the first storage means (SDRAM 16) for each divided image (strip) that is an image obtained by dividing the two-dimensional image by a line segment parallel to the Y axis. When transferring from the first storage unit (internal buffer 12) to the second storage unit, the setting unit (CPU 20) sets the first width size set in the first address generation unit to the X axis of the divided image. On the other hand, the coordinate setting means (CPU 20 or output width counter 42), when there is a divided image to be transferred next, sets the coordinates on the XY plane where the divided image exists as the first size. It is set in the address generation means (address generation circuit 10).

  Furthermore, in the present embodiment, the processed image data is stored in the second storage means (for each divided image (strip) that is an image obtained by dividing the two-dimensional image by a line segment parallel to the Y axis. When transferring from the built-in buffer 12) to the first storage means (SDRAM 16), the setting means (CPU 20) sets the second width size set in the second address generation means to the value of the divided image. On the other hand, the coordinate setting means (the CPU 20 or the output width counter 42) determines the coordinates on the XY coordinate plane where the divided image exists when there is a divided image to be transferred next, while the size is set in the X axis direction. The second address generation means (address generation circuit 10) is set.

  On the other hand, in the present embodiment, the two-dimensional image indicated by the image data is an image on the XY plane, the coordinates on the XY plane, and the memory space in the first storage means (SDRAM 16) capable of storing the image data. 1-to-1 address is associated with the first start address that is the start address in the first storage means of the image data, and the X-axis direction of the two-dimensional image on the XY plane Based on the set value of the first width size that is the size and the first pixel size that is the size of one pixel in the image data, an address in the memory space corresponding to the coordinates is generated from the set coordinates The first head address, the first width size, and the first pixel size are set for first address generation means (address generation circuit 10) that performs A setting step (step 101), a coordinate setting step (step 102) for setting the coordinates where the two-dimensional image exists on the XY plane to the first address generation unit, and the first address generation unit A transfer step of transferring the image data stored in the first storage unit by the first pixel size from the address generated in step (2) to a second storage unit used when performing predetermined signal processing ( Step 105).

  In the present embodiment, when the processed image data, which is the image data subjected to the signal processing, is transferred from the second storage means (built-in buffer 12) to the first storage means (SDRAM 16). A second start address that is a start address in the first storage means in which the transferred processed image data is stored, and the X-axis direction of the two-dimensional image indicated by the processed image data on the XY plane On the basis of a set value of a second width size that is a size for and a second pixel size that is a size of one pixel in the processed image data, in the memory space corresponding to the coordinates from the set coordinates For the second address generation means (address generation circuit 10) for generating an address, in the setting step (step 101), the second head address, the previous address A second width size and the second pixel size are further set, and in the coordinate setting step (step 102), coordinates where the two-dimensional image indicated by the processed image data exists on the XY plane are set as the XY. The coordinates in the plane are further set for the second address generation means, and in the transfer step (step 105), the second pixel size is set by the second pixel size from the address generated by the second address generation means. The processed image data stored in the storage means is transferred to the first storage means.

  According to a seventh aspect of the present invention, the image data is stored in the first storage means (SDRAM 16) for each divided image (strip) which is an image obtained by dividing the two-dimensional image by a line segment parallel to the Y axis. ) To the second storage means (internal buffer 12), in the setting step (step 101), the first width size set in the first address generation means is set to the first width size of the divided image. In the coordinate setting step (step 102), when there is a divided image to be transferred next, the coordinates on the XY plane where the divided image exists are stored in the first address generation unit. Set.

  Furthermore, in the present embodiment, the processed image data is stored in the second storage means (for each divided image (strip) that is an image obtained by dividing the two-dimensional image by a line segment parallel to the Y axis. When transferring from the built-in buffer) to the first storage means (SDRAM 16), in the setting step (step 101), the second width size set in the second address generation means (address generation circuit 10) is set. In the coordinate setting step (step 102), when there is a divided image to be transferred next, the coordinates on the XY coordinate plane where the divided image exists are set as the size of the divided image in the X-axis direction. Set to the second address generation means.

It is a figure which shows the hardware constitutions (the 1) of the image processing apparatus which concerns on embodiment. It is a figure which shows an address output circuit (the 1). It is a figure which shows XY coordinate which concerns on embodiment. It is a figure which shows a mode that the transfer of image data is performed for every divided strip. It is a flowchart (the 1) which shows a transfer process. It is a figure which shows the handshake performed by an address generation circuit, DMAC, and a built-in buffer. It is a figure which shows the margin. It is a figure which shows the image in the coordinate at the time of using a work area. It is a figure which shows an address output circuit (the 2). It is a figure which shows a memory map in case output data is overwritten on the input data which complete | finished transfer. It is a figure which shows a memory map when not overwriting output data with input data. It is a figure which shows the hardware constitutions (the 2) of the image processing apparatus which concerns on embodiment. It is a figure which shows an output width counter. It is a flowchart (the 2) which shows a transfer process. It is a figure which shows an example of the hardware constitutions which concern on the data transfer in a prior art. It is a figure which shows a mode that image data is transferred for every strip in a prior art.

Explanation of symbols

10 Address generation circuit 12 Built-in buffer 14 DMAC
18 Signal processing circuit 20 CPU
30 Work area 42 Output width counter

Claims (2)

While functioning as a work area capable of storing input image data and output image data, the two-dimensional image indicated by the input image data and the output image data is an image on the XY plane, the coordinates on the XY plane, and the work area First storage means in which the addresses in the memory space are associated one-to-one ;
Second storage means for use in performing a predetermined signal processing on the input image data,
Dividing means for dividing the two-dimensional image indicated by the input image data into a plurality of two-dimensional images in the X direction in the work area;
First setting means for setting first position coordinates indicating positions in the work area corresponding to pixels of image data of the two-dimensional divided image obtained by the division;
The first head address of the working area, on the basis of the start coordinates of the X-direction width size of the two-dimensional image segment pixel size of one pixel in the image data of the two-dimensional image segment and the two-dimensional image segment First address generation means for generating an address in the memory space corresponding to the set first position coordinate;
First transfer for transferring the input image data stored in the first storage means to the second storage means for each of the two-dimensionally divided images according to the address generated by the first address generation means. Means,
Signal processing means for performing the signal processing on the image data of the two-dimensionally divided image transferred to the second storage means;
Second setting means for setting a second position coordinate indicating a position in the work area corresponding to a pixel of the image data subjected to the signal processing;
A second head address of the work area, a width size in the X direction of the two-dimensional image indicated by the image data subjected to the signal processing, a pixel size corresponding to one pixel in the image data subjected to the signal processing, and Second address generation means for generating an address in the memory space corresponding to the set second position coordinate based on the leading coordinates of the two-dimensional image indicated by the image data subjected to signal processing;
According to the address generated by the second address generation means, the image data subjected to the signal processing is transferred as the output image data from the second storage means to the first storage means, and the transfer In this case, the input image data transferred to the second storage means next to the last input image data transferred to the second storage means, and the signal processing is performed and transferred to the first storage means Output image data obtained by shifting the input image data to be transferred next in the X direction in the work area and performing the signal processing on the shifted trace area so that the output image data does not overlap in the work area. A second transfer means for overwriting,
An image processing apparatus comprising: While functioning as a work area capable of storing input image data and output image data, the two-dimensional image indicated by the input image data and the output image data is an image on the XY plane, the coordinates on the XY plane, and the work area Preparing a first storage means having a one-to-one correspondence with an address in a memory space in
Preparing a second storage means used when applying predetermined signal processing to the input image data,
A division step of dividing the two-dimensional image indicated by the input image data into a plurality of two-dimensional images in the X direction in the work area;
A first setting step of setting a first position coordinate indicating a position in the work area corresponding to a pixel of image data of the two-dimensional divided image obtained by the division;
The first head address of the working area, on the basis of the start coordinates of the X-direction width size of the two-dimensional image segment pixel size of one pixel in the image data of the two-dimensional image segment and the two-dimensional image segment A first address generation step of generating an address in the memory space corresponding to the set first position coordinate;
First transfer for transferring the input image data stored in the first storage means to the second storage means for each of the two-dimensionally divided images according to the address generated in the first address generation step. Stages,
A signal processing stage for performing the signal processing on the image data of the two-dimensionally divided image transferred to the second storage means;
A second setting step of setting a second position coordinate indicating a position in the work area corresponding to a pixel of the image data subjected to the signal processing;
A second head address of the work area, a width size in the X direction of the two-dimensional image indicated by the image data subjected to the signal processing, a pixel size corresponding to one pixel in the image data subjected to the signal processing, and A second address generation step of generating an address in the memory space corresponding to the set second position coordinate based on the leading coordinates of the two-dimensional image indicated by the image data subjected to signal processing;
According to the address generated in the second address generation stage, the image data subjected to the signal processing is transferred as the output image data from the second storage means to the first storage means, and the transfer In this case, the input image data transferred to the second storage means next to the last input image data transferred to the second storage means, and the signal processing is performed and transferred to the first storage means Output image data obtained by shifting the input image data to be transferred next in the X direction in the work area and performing the signal processing on the shifted trace area so that the output image data does not overlap in the work area. A second transfer stage overwriting
An image processing method comprising:

Images