JP2022086187A - Image processing device, image processing device, image processing method, and program - Google Patents

Abstract

To provide an image processing device having excellent processing performance and being able to be configured at low cost.SOLUTION: An image processing device 100 includes transfer means (input DMAC 103) for dividing and transferring original image data representing an original image into image data consisting of a certain number of lines along a main scanning direction, and variable magnification processing means (variable magnification processing unit 1041) for setting the image data transferred from the transfer means as input image data in a processing unit, and executing variable magnification processing for the input image data at a preset magnification to generate variable magnification image data (magnification processing means). The variable magnification processing means includes output setting means for setting the number of output lines to be output from the variable magnification processing means, and from among the variable magnification image data that can be generated using the input image data, image data limited to the number of output lines set by the output setting means is output as the output image data.SELECTED DRAWING: Figure 1

Description

The present invention relates to an image processing apparatus, an image processing method, and a program for performing scaling processing on input image data.

The image processing device provided in a digital copier or the like is provided with a function of performing scaling processing such as reduction / enlargement of original image data. In the scaling processing of the original image data, the data of the entire image is not used as the processing unit, but the image data of multiple lines called bands in which the entire screen is vertically divided into bands is used as the processing unit, and each band is used from the working memory. The data is read out in order and scaled. Further, in some image processing devices, a processing unit that performs processing such as compression processing and quantization processing on the image data that has been scaled is provided after the scaling processing unit.

Examples of the processing unit provided after the variable magnification processing unit (hereinafter referred to as the post-stage processing unit) include a JPEG compression unit and a dither processing unit. In these post-stage processing units, a fixed number of lines (8 or more) 16 lines) There are some that perform processing in units. Therefore, in order to directly input the image data that has been scaled by the scaling processing unit to the post-stage processing unit, the image data output from the scaling processing unit is input from a fixed number of lines according to the processing unit of the post-stage processing. It is necessary to output the image data.

Patent Document 1 discloses a technique for constantly outputting a fixed number of lines of image data from a scaling processing unit. Specifically, the number of lines to be input to the scaling processing unit is changed for each input band so that the number of output lines after scaling becomes constant. According to this, the image data output from the scaling processing unit can be directly input to the subsequent processing without going through the buffer memory.

Japanese Unexamined Patent Publication No. 2006-135565

However, Patent Document 1 requires a circuit configuration or a processing process for changing the number of lines of each input band input to the scaling processing unit, which causes an increase in equipment cost and a decrease in processing performance. be.

Therefore, an object of the present invention is to provide an image processing apparatus and an image processing method which are excellent in processing performance and can be configured at low cost.

The present invention has a transfer means for dividing and transferring the original image data representing the original image into image data consisting of a certain number of lines along the main scanning direction, and the image data transferred from the transfer means as a processing unit. A scaling processing means that can generate scaling image data by performing scaling processing on the input image data at a preset magnification as input image data, and an output line to be output from the scaling processing means. The variable magnification processing means includes an output setting means for setting the number, and the scaling processing means is limited to the number of output lines set by the output setting means among the variable magnification image data that can be generated using the input image data. It is an image processing device characterized by outputting image data as output image data.

According to the present invention, it is possible to provide an image processing apparatus and an image processing method which are excellent in processing performance and can be configured at low cost.

It is a block diagram which shows the whole structure of the image processing apparatus in 1st and 2nd Embodiment. It is a conceptual diagram for demonstrating the operation of the input DMAC. It is a figure which shows the input pixel which is input to a variable magnification processing unit, and the output pixel which has undergone a variable magnification processing. It is a figure which shows the state which the image data of a plurality of bands was scaled by the 1st comparative example. It is a figure which shows the state which the image data of a plurality of bands was scaled by the 2nd comparative example. It is a figure which shows the state which the image data of a plurality of bands was scaled by this embodiment. It is a flowchart which shows the algorithm of the scaling process by this embodiment. It is a block diagram which shows the whole structure of the image processing apparatus in 3rd Embodiment and 3rd comparative example.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

(First Embodiment)
FIG. 1 is a block diagram showing the overall configuration of the image processing apparatus 100 according to the first embodiment of the present invention. In the present embodiment, an image processing device having a function as a copying machine that reads a manuscript image, generates original image data, and outputs the generated original image data to a recording medium will be described as an example.

In FIG. 1, the image processing device 100 has a CPU 101 that comprehensively controls the operation of each part described later. Further, the image processing apparatus 100 includes a scanner 102, an input DMAC 103, a scanner image processing unit 104, a JPEG compressor 105, a JPEG decompressor 106, a printer image processing unit 107, a printer 108, a main memory 130, and a bus 150.

The scanner 102 functions as an image acquisition means for scanning a document to obtain bitmap image data (original image data) composed of RGB multi-valued pixels. However, the image acquisition means is not limited to the scanner, and may be an i / f that can acquire image data from other devices in the order of raster. The bitmap image data acquired by the scanner 102 is temporarily stored in the main memory 130. Further, the main memory 130 is a storage means for temporarily storing the processing results of each part in each part described later, and is usually composed of a DRAM.

The input DMAC (Direct Memory Access Controller) 103 reads RGB bitmap data stored in the main memory 130 in band units from the main memory 130 and supplies the data to the scanner image processing unit 104. Here, the band refers to image data obtained by dividing the data of the entire original image acquired by the scanner 102 into strips in the vertical direction. This band includes a plurality of horizontal lines composed of pixels arranged in the horizontal direction intersecting the vertical direction, and this is a processing unit for image processing described later. In the following description, the horizontal line is also simply referred to as a line. Further, in order to distinguish from this line (horizontal line), a line composed of pixels arranged in the vertical direction in image data is referred to as a vertical line.

The scanner image processing unit 104 performs image processing such as color adjustment, edge enhancement, and scaling on the RGB image data read from the main memory 130 for each predetermined number of bands. In this embodiment, the scaling process is performed for each band. Further, inside the scanner image processing unit 104, a scaling processing unit (magnification processing means) 1041 that performs interpolation scaling processing of RGB image data for each band is provided. The scaling processing unit 1041 performs scaling processing on a band having a fixed number of lines input from the input DMAC 103, and obtains image data (magnification image data) after scaling consisting of a predetermined number of horizontal lines in the subsequent stage. Output to JPEG compressor 105.

The JPEG compressor 105 performs JPEG compression on image data (output band data) having a predetermined number of lines equal to the height of the MCU (Minimum Code Unit). Therefore, it is necessary to output the number of lines of the processed RGB image data output from the scanner image processing unit 104 to the JPEG compressor 105 to be equal to the height of the MCU. The compressed image data (JPEG data) output from the JPEG compressor 105 is stored in the main memory 130. The JPEG decompressor 106 reads out the JPEG data stored in the main memory 130, decompresses the JPEG data, and returns the JPEG data to the RGB data. This RGB data is stored in the main memory 130 again.

The printer image processing unit 107 reads RGB data from the main memory 130, and performs processing such as conversion to the CMYK color space used in the printer 108 and quantization. The printer 108 records an image on a recording medium such as recording paper based on the CMYK data quantized by the printer image processing unit 107. As the printer 108, there is an inkjet printer or the like that ejects ink for recording. However, the type of printer is not particularly limited, and other forms of printers, for example, printers using an electrophotographic method, and the like can also be applied.

FIG. 2 is a diagram showing a transfer process in which the input DMAC 103 in the present embodiment reads out the image data stored in the main memory 130 and transfers the data in pixel units to the scanner image processing unit 104 in the subsequent stage of the input DMAC 103. ..

The main memory 130 secures an area capable of storing RGB image data having an input image width for a plurality of lines (two bands B1 and B2 of the input band described later). This area is referred to as scan buffer 1301. The data output from the scanner 102 for each horizontal line is sequentially stored in the scan buffer.

The input DMAC 103 has an HV conversion memory 1031 inside. The HV conversion memory 1031 has a capacity corresponding to "the number of bytes of the DRAM burst length" x "the height of the input band (the number of horizontal lines of the input band)". The image data corresponding to this capacity is referred to as an HV (Horizontal-Vertical) block here.

The input DMAC 103 reads image data in units of DRAM burst length (32 bytes or the like) from the upper left end of the input band B1 stored in the scan buffer 1301 and writes the image data to the HV conversion memory 1031. At this time, the image data is read from the scan buffer 1301 in the input band B1 for the number of bytes corresponding to the DRAM burst length (one horizontal line) as shown by the broken line arrow in FIG. 2 (a). Read the minute image data). Then, each time the image data corresponding to the DRAM burst length is read, the image data is moved to the next horizontal line (lower horizontal line in the figure) in the band, and the image data of the horizontal line is read. This process is repeated for the height of the input band B1 (for the number of horizontal lines). Then, when the line at the lower end of the input band B1 is read (that is, when one HV block is read), the process shifts to the first horizontal line of the next input band B2, and the input band B2 The image data is sequentially read out in the same manner as the input band B1.

The image data read from the scan buffer 1301 is sequentially stored in the HV conversion memory 1031 as shown by the broken line arrow in FIG. 2 (b). When the data for one HV block is stored in the internal HV conversion memory 1031 of the input DMAC 103, then, as shown by the broken arrow in FIG. 2 (c), the vertical line of the image from the upper left end of the HV block. Image data (pixel data) is read out one pixel at a time in the direction. Then, the read image data is transferred to the scanner image processing unit 104 provided in the next stage. The vertical line referred to here refers to a line consisting of a plurality of pixels arranged along a direction orthogonal to the above-mentioned horizontal line. Further, in the following description, the direction along the vertical line is also referred to as a sub-scanning direction, and the direction along the horizontal line is also referred to as a main scanning direction.

When the reading / transfer of the image data of the first vertical line (the line located at the left end in the figure) in the image data for one band stored in the HV conversion memory 1031 is completed, the first vertical line is completed. Move to the second vertical line adjacent to the right side of. Then, the pixel data is sequentially read and transferred from the pixel data at the upper end of the second line. This process is repeated, and the image data for one HV block, that is, the image data for one band is transferred to the scanner image processing unit 104.

By the above processing, for example, when the image data of one input band B1 shown in FIG. 2A is transferred to the scanner image processing unit 104, the image data (HV) for the input band B2 on the right side is next. (Image data for blocks) is read from the scan buffer 1301. Then, the read data is temporarily stored in the HV conversion memory 1031. Then, the image data for the input band B2 stored in the HV conversion memory 1031 is read out and transferred to the scanner image processing unit 104 according to the progress of the image processing for the previously transferred band B1. This process is repeated up to the lower right corner of the input band.

In this way, the image data constituting the vertical line for each input band is transferred from the input DMAC 103 to the scanner image processing unit 104 as RGB data for each pixel. At this time, in the scanner image processing unit 104, each of the plurality of sub-blocks (not shown except the scaling processing unit) including the scaling processing unit 1041 sequentially processes the input pixel data. That is, the process of transferring the result processed in one subblock to the next subblock is repeated in the pipeline for each pixel.

FIG. 3 is a diagram showing an input / output band in the scaling processing unit 1041 and a relationship between an input pixel and an output pixel in the scaling processing. Here, the image data of the input band consisting of six horizontal lines stored in the HV conversion memory 1031 is input to the scaling processing unit 1041 in the order of vertical lines as shown in FIG. 2 (c), and the image data is changed. The case where the reduction is performed by the double processing unit 1041 is shown. The reduction referred to here is to reduce the number of pixels, that is, to reduce the resolution, so that the size of each pixel can be considered to be large. Hereinafter, the scaling process will be described by taking the reduction process as an example.

FIG. 3A shows the position of the input pixel Pi to the scaling processing unit 1041, and FIG. 3B shows the position of the output pixel Po output from the scaling processing unit 1041. In FIG. 3A, when the image data of 6 pixels (pixels 0 to 5) is input to the scaling processing unit 1041 in the vertical line direction (Y direction), the input position of the input pixel Pi is set to the right. It moves up by 1 pixel and 5 pixels upward, and then 6 to 11 input pixels are sequentially input along the vertical line direction. This operation is repeated until image data for one input band is input. Further, FIG. 3B shows an output band obtained by subjecting the image data of the input band to scaling processing. Here, the output band consisting of four horizontal lines is output with respect to the input band consisting of six horizontal lines shown in FIG. 3A. That is, in FIG. 3B, when the four output pixels 0 to 3 are output along the vertical line direction (Y direction), the output position of the output pixel Po is set to the right by one pixel. It moves up by 3 pixels, and then outputs the 4th to 7th output pixels. This operation is repeated using the input image data for one input band.

Here, a method of calculating the pixel value of the output pixel Po with respect to the pixel value of the input pixel Pi will be described with reference to FIG. 3 (c). In addition, in FIG. 3C, Pi (0) shows the 0th input pixel Pi in FIG. 3A. Similarly, Pi (1), Pi (6), and Pi (7) indicate the input pixels Pi of Nos. 1, 6, and 7, respectively. Further, Po (0) indicates the 0th output pixel Po shown in FIG. 3 (b). In the following description, the same notation is used for other pixels (not shown).

The pixel value of the output pixel Po can be calculated by inputting four input pixels Pi including the output pixel Po. In the examples of FIGS. 3A and 3B, the pixel value of the output pixel Po (0) can be calculated by inputting the 7th input pixel Pi (7). Similarly, the pixel value of the output pixel Po (1) is calculated when the input pixel Pi (8) is input, and the pixel value of the output pixel Po (4) is calculated when the input pixel Pi (13) is input. be able to.

Further, in FIG. 3C, the central points of the four pixels Pi (0), Pi (1), Pi (6), and Pi (7) around the output pixel Po (0) are A and B. , C, D, and let P be the center point of the output pixel Po (0). Further, point E is a point where one-dimensional linear interpolation is performed in the main scanning directions of points A and B, and point F is a point where one-dimensional linear interpolation is performed in the main scanning directions of points C and D. .. In this case, the point P is obtained by performing one-dimensional linear interpolation in the sub-scanning direction of the points E and F.

The position of the point P is represented by coordinates on a plane in which the distance between the points AB is a base (fixed value, for example, 8192) and the distance between the points AC is also a base. Is (0,0), and the coordinates of the point D are (base, base).

The pixel value of the pixel P in this coordinate system is obtained by the following calculation. Let the pixel values of points A, B, C, D, E, F, and P be a, b, c, d, e, f, and p, respectively. Further, the position of the point P in the main scanning direction (X direction) is represented by the distance hphase from the point A when the point AB is the base, and the position of the point P in the sub-scanning direction is the base between the points AC. It is expressed by the distance vphase from the point A at the time of.

Here, the pixel value of the point P is expressed by the following equation.
e = a + (baa) × hphase / base
f = c + (dc) × vphase / base
p = e + (fe) × vface / base
Although the position of each pixel is shown by the coordinates of the center of each pixel in FIG. 3C, the relative position of each pixel is the same even if it is shown by the coordinates of the upper left end of each pixel, as described above. The calculation result is the same. Therefore, in the following description, the upper left position of each pixel (A'with respect to A) is used as the coordinates indicating the position of the pixel.

Further, in the input pixel at the upper left end of the input band, the initial coordinate value can be specified as (hphase_base, vphase_base). When the processing is started with the initial coordinate value set to 0, each phase value is always 0 at the time of reduction of 1 / n (n is an integer) and the interpolation calculation is not performed, resulting in thinning reduction and the image quality of the output image. descend. If the initial coordinate values are specified as (hphase_base, vphase_base), deterioration of image quality can be avoided. However, since the designation of hphase_base is irrelevant to the characteristic processing of this embodiment, the following description will be given with 0. The value of vphase_base is set for each input band based on the coordinate value of the line of the output image generated in the immediately preceding input band in order to match the phase of the output image generated for each input band.

Next, scaling processing for image data of a plurality of bands will be described. In the following description, in order to clarify the characteristics of the scaling process executed in the present embodiment, first, the conventional scaling process executed in the comparative example of the present embodiment will be described, and then the present embodiment will be described. The scaling process executed in the form will be described.

<Scale processing of the first comparative example>
FIG. 4 is a diagram schematically showing scaling processing for a plurality of input bands according to the first comparative example. Here, the variable magnification (magnification) in the sub-scanning direction (Y direction) is given as a relative value magv with respect to the base. For example, if base = 8192 and 80% reduction is to be performed, magv = 8192 / 0.8 = 10240 is set. The example shown in FIG. 4 shows a reduction of 75%, that is, a process when magv = base / 0.75 = 10922 is set. Here, vphase_base = base / 3 of the first band is set. Further, in FIGS. 4 to 6, the numbers attached to the leftmost pixels of the input image and the output image are not the numbers indicating the processing order of each pixel as in FIG. 3, but are horizontal in the entire input image and output image. Shows the line number of the line. In the following description, the line with line number 0 will be referred to as the first line, the line with line number 1 will be referred to as the second line, and the same shall apply hereinafter.

As shown in the upper left corner of FIG. 4, the coordinates of the upper left corner of the input image are (0,0), and the height of the input band is 6 lines. The coordinate values in the sub-scanning direction of the input image start from 0 and are added by base for each pixel. The coordinate values in the sub-scanning direction of the output image start from vphase_base and are added in magv increments. The input in the sub-scanning direction enables an output process for acquiring an output pixel when the coordinate value of the input pixel becomes vphase_base + magv or more. In the example of FIG. 4, by inputting up to the second line (line number 1) of the input image, the first line (line number 0) of the output image can be output. Further, by inputting up to the third line (line number 2) of the input image, the second line (line number 1) of the output image can be output. Similarly, when inputting up to the Lth line (6th line in the figure), the input coordinate value becomes L × base, and the number of output lines M is M = (L × base-vphase_base) / magv (calculation result is rounded down).
It can be obtained by the calculation.

In the example of FIG. 4, M = 4, and the fifth line of the output image cannot be output because the number of input lines is insufficient as shown in the figure. For example, in order to output the 5th line of the output image, it is necessary to input the 7th line (line number 5) of the input image, but the number of input lines in the first band is 6 lines (line number 0). ~ 5 6 lines). Therefore, the number of input lines is insufficient for the input of the first band, and the fifth line cannot be output. Therefore, in this first band, 4 lines are output for 6 lines of input.

Further, in the scaling process of the next input band following the previous input band, as shown in the figure, input is always performed from the line one line back from the previous input band, and vphase_base at the start of processing of the next input band. The value,
(Vphase_base + magv × M) -base × (L-1)
, The same processing as the previous input band is performed. As a result, the next output image can be continuously output to the output image of the previous input band.

In the above processing, the number of input lines of each input band is constant, and the offset address between the input bands (the difference between the read start address of one input band and the next input band, in the example of FIG. 4). (For 5 lines of input image) is also constant. Therefore, the input DMAC 103 can perform continuous operation with a relatively simple address calculation.

However, when the above-mentioned scaling process is repeated, the number of lines of the output image corresponding to the input band is not constant depending on the value of magv. This is because the value of vphase_base changes for each band as the processing of the input band progresses, so the value of the number of output lines M M = (int) (L × base-vbase_base) / magv
Is not constant. In the example of FIG. 4, the number of output lines is 3 in the third band, and the number of output lines is different from that of the first band and the second band.

That is, in the first comparative example, an input band having a fixed number of lines is input to the scaling processing unit 1041, and the number of offset lines for each band is also fixed (here, (number of lines of input band)-(1 line)). The scaling process is performed for each band. As a result, the number of lines of the output band output from the scaling processing unit is not constant depending on the value of the scaling factor magv, and variations occur. Therefore, there arises a problem that the output of the scaling processing unit cannot be directly input for a process that requires input of a certain number of lines after the scaling processing unit 1041, for example, a JPEG compression process. As a method for solving this problem, the second comparative example described below is known.

<Second comparative example>
In the second comparative example, the number of input lines that can output a certain number of output lines is calculated for each input band, and the image data for the number of lines calculated for each input band is read from the main memory and stored in the scaling processing unit. By supplying the data, the problem in the first comparative example described above is solved.

FIG. 5 shows the scaling process executed in the second comparative example. Here, the scaling process is performed by setting Base and magv in the same manner as in FIG. As described above, when the process shown in FIG. 4 is performed, the number of input lines is insufficient in the third band, and the number of output lines is not constant (4 lines in FIG. 3). Therefore, in the second comparative example, the number of input lines in the third band is 7 lines, which is increased by 1 line, instead of the same 6 lines as in the 1st and 2nd bands. That is, 7 lines of line numbers 10 to 15 are input as the third band. As a result, it becomes possible to output the image data of the output line of the line number 11 by the two lines of the line numbers 14 and 15, and it is possible to output the output image of the four lines (output lines of the line numbers 8 to 11). It will be possible.

However, in the second comparative example, the following problems occur. That is, in the second comparative example, it is necessary to make the number of lines of the input band to the scaling processing unit 1041 different for each band. Therefore, the circuit configuration of the input DMAC 103 that supplies the image data of the input band to the scaling processing unit (in the configuration of FIG. 1, the scanner image processing unit 104 including the scaling processing unit 1041) determines the number of lines in each band. It is more complicated than the second comparative example, which is constant. For example, in the HV conversion memory 1031 in the input DMAC 103, it is necessary to change the input / output operation of the image data shown in FIGS. 2 (b) and 2 (c) for each band.

Further, as the number of lines differs for each input band, there arises a problem that the number of offset lines between adjacent input bands also needs to be changed for each input band. In the example of FIG. 5, the number of offset lines set between the first band and the second band and the number of offset lines between the second band and the third band are both five lines. However, the number of offset lines between the third band and the fourth band is 6, and the number of offset lines is changing. In order to perform such processing, it is necessary to calculate the offset address for each input band. However, there are many difficulties in allowing the calculation of the offset address to be performed automatically at the input DMAC 103.

Therefore, if the DMAC 103 itself does not calculate the offset address, but the CPU 101 calculates the offset address for each input band and sets the calculated value in the input DMAC 101 each time, the circuit configuration of the input DMAC 103 can be simplified. .. However, in that case, it takes time for the CPU to calculate and set the offset address while outputting each band. Therefore, the total processing time becomes longer than when the input DMAC 103 performs automatic calculation, which causes a problem that the processing performance of the entire device is affected. Therefore, in the present embodiment, the following scaling processing is executed.

<Variable magnification processing in the first embodiment>
In the present embodiment, the problems of the first comparative example and the second comparative example described above are solved by the configurations listed below.

i) Input A fixed number of input lines are read from the main memory 130 by the input DMAC 103, and an input band composed of the read fixed number of input lines is input to the scaling processing unit 1041 included in the scanner image processing unit 104.

ii) In the scaling processing unit 1041, among the output lines that can be generated based on the input band of a fixed number of input lines, the output lines exceeding the preset number of output lines are output as unnecessary output lines. Do not make it disappear internally. That is, in the existing scaling processing unit such as the first modification and the second modification, when a fixed number of input bands are input, the number of output lines after scaling becomes N or N + 1, and the number of output lines. Will vary. On the other hand, in the present embodiment, even if the number of output lines exceeding the preset number of output lines can be generated, the output lines exceeding the set number of output lines are scaled as extra output lines. It does not output from the part, but disappears internally. As a result, the number of output lines supplied to the subsequent stage of the scaling processing unit is set to a fixed number (here, N).

The value of the number of output lines N is output in the scaling processing unit 1041 in advance according to the image processing unit of the post-stage processing unit that performs predetermined image processing on the image data output from the scaling processing unit 1041. It is fixed to a fixed number of output lines by the setting means. However, it is also possible to configure the image processing device 100 so that an arbitrary number of output lines can be set by an input unit or the like.

iii) The scaling processing unit 1041 counts the number of input lines consumed in the scaling processing, and transmits information indicating the count value to the input DMAC 103. The input DMAC 103 automatically calculates the offset address of the input band to be input next based on the information transmitted from the scaling processing unit 1041, and determines the input line of the next input band.

Here, with reference to FIGS. 6 and 7, the processing executed by the input DMAC 103 and the scaling processing unit 1041 in the present embodiment will be described.

When performing scaling processing at the magnification (magv) set by the input unit (magnification setting means) of the image processing device 100, each band is capable of outputting the same number of output lines M in all input bands. The number of input lines L of the above is calculated in advance by the CPU 101. Since the maximum value of vphase_base is base-1, L can be calculated by the following equation 1.
L = (int) (((base-1 + magv × (M-1)) / base) +1) (Equation 1)
Here, (int) means a truncation process. In the example shown in FIG. 5, L = 7 with respect to M = 4.

The input DMAC 103 always reads image data having the number of lines L for each input band from the main memory 130. As a result, the input DMAC 103 may always supply image data for the L line to the scanner image processing unit 104. Therefore, it is not necessary to change the data reading operation in the HV conversion memory 1031 shown in FIG. 2 for each band as in the second modification. Therefore, according to the present embodiment, the configuration of the input DMAC 103 can be significantly simplified as compared with the case where the operation of the second comparative example is to be realized.

Further, when the data for L lines is always input to the scaling processing unit 1041 and the scaling processing similar to that of the first comparative example is performed, the number of lines in the output band is determined by the value of vphase_base in each input band. Is M or M + 1. Therefore, it is not possible to supply a fixed number of output lines to the subsequent stage.

On the other hand, in the scaling processing unit 1041 of the present embodiment, if the number of output pixels of each vertical line is cut off by M and the output pixel of the M + 1th line can be generated, the output pixel is not output to the subsequent stage. Extinguish. Then, the output coordinates are moved one to the upper right, and the output pixels of the next vertical line are sequentially generated from the top and output. This process is repeated for all the input pixels of the input band. As a result, it is possible to output image data having a constant force line number obtained by subjecting each input band to scaling processing. That is, according to the present embodiment, the scaling processing unit 1041 outputs the image data after the scaling processing, which is always limited to a fixed number of lines, to the subsequent processing unit (JPEG compressor 105 in this example). can do.

Further, in the present embodiment, the input of the next band is not performed from the line returned by one line from the end of the previous input band as in the first and second comparative examples, as shown in FIG. , Return the extinguished output line to the line that can be generated.

Further, the initial phase value vphase_base of the next band is not the value of the disappeared line (horizontal line composed of M + 1st pixel from the top of each vertical line of the output image), but the last output line immediately before (the horizontal line). Let it be the vphase value of the horizontal line composed of the Mth pixel).

By performing the above processing, the output line disappeared in one output band can be supplemented by the next output band. Therefore, it is possible to smoothly connect a plurality of output bands after scaling.

FIG. 6 shows the scaling process when M = 4, magv = base × 4/3, and vphase_base = base / 4. In this case, the number of input lines L for each of the first, second, and third input bands is 7. The number of input lines L is calculated by the CPU 101 performing the calculation of the above equation 1.

When the first input band is input to the scaling processing unit 1041, the number of output lines that can be originally generated is 5, but in the present embodiment, the lowest line of the output band is eliminated and 4 lines are eliminated. Output line is generated and output as the first output band.

Further, the input of the second input band is performed from the input line used when generating the disappeared fifth output line (output line of line number 4). That is, from the 6th line (line number 5), which is the line returned by 2 lines including the 7th line (input line of line number 6) which is the final line of the 1st input band. 2 Input the input band. In this case, the number of offset lines is five. The initial phase value vphase_base of the output band corresponding to the second input band is the phase value vphase of the fifth line that has been extinguished.

Using the flowchart of FIG. 7, the processing contents from performing the scaling processing in the scaling processing unit 1041 for each input band to outputting the output image of the M line will be described in detail. However, since the scanning scaling process is a well-known technique disclosed in Patent Document 1 and the like, detailed description thereof will be omitted in the present embodiment, and only the scaling process in the sub-scanning direction will be described. In the flowchart of FIG. 7, S attached to each process number means a step.

Before executing the process shown in the flowchart, the CPU 101 calculates in advance the number L of input lines of each input band according to the above-mentioned equation 1. The input DMAC 103 always reads out the calculated image data having the number of lines L as one band from the main memory 130, and supplies the image data to the scaling processing unit 1041 in the order of vertical lines as shown in FIG. 2 (c).

In S700, the initial setting is performed. Here, an arbitrary value of 1 or more and less than base is set as the phase value vphase_base of the first input band, and the scaling process of the first input band is started.

In S701, the internal variables of the scaling processing unit 1041 are initialized as follows.
vface = vface_base
in_ctr = out_ctr = 0
Here, vphase is the sub-scanning phase value at the time of pixel output, in_ctr is the count value of the input counter that counts the number of input pixels in the sub-scanning direction, and out_ctr is the count value of the output counter that counts the number of output pixels in the sub-scanning direction. Each is shown. After the initialization process, the input band scaling process is started.

In S702, the count value in_ctr of the input counter is incremented by 1 each time the input pixel is input to the scaling processing unit 1041.

In S703, it is determined whether or not it is possible to generate an output pixel by using the input pixel input to the scaling processing unit 1041 based on the count value in_ctr. This determination is made by determining whether or not the total value (in_ctr × base) of the sub-scanning phase values of the input pixels is equal to or greater than the total value of vphase_base and magv. That is,
in_ctr × base ≧ vphase + magv (Equation 2)
Judgment as to whether or not the relationship of is satisfied. Here, if the relation of the equation 2 is satisfied (in the case of Yes), the process proceeds to S704, and if the relationship is not satisfied (in the case of No), the process returns to S702.

In S704, the value of vphase is updated to vphase + magv.
In S705, the count value out_ctr of the output counter is incremented by 1 each time the output pixel is output from the scaling processing unit 1041.

In S706, the out_ctrth pixel value from the top of the output band is calculated, and the result is output to the next stage. Here, the output value out_val is obtained by the following equation 3.
out_val
= (In_val [in_ctr-1] × vphase
+ In_val [in_ctr] × (base-vphase)) / base (Equation 3)
In addition, in_val [x] represents the pixel value of the x + 1th pixel from the top of the output band.

In S707, the following equation 4 is calculated.
next_vphase = vphase-in_ctr × base (Equation 4)
Here, next_vphase is an initial value of vphase_base in the next band, and is held in the scaling processing unit 1041 as a register until the processing of the next band.

In S708, it is determined whether or not the last output pixel of one vertical line is output. This determination is performed by determining whether or not the count value out_ctr of the output counter has reached the preset number of output lines M. That is,
out_ctr = M
It is performed by determining whether or not it is. As a result of the determination, if out_ctr = M, the process proceeds to S709 (if Yes), and if not, the process returns to S703 (if No). By making this determination, the number of output lines after scaling becomes a constant value M. When the next stage of the scaling processing unit 1041 is a JPEG compressor 105 as shown in FIG. 1, the number of lines M to be output from the scaling processing unit 1041 is usually M = 8 or M = 16. be.

After that, in S709, the following equation 5 is calculated.
offset = in_ctr-1 (Equation 5)
The in_ctr in the equation 5 is the number of input pixels at the time when the Mth output pixel is generated in one vertical line in a certain input band. That is, the offset means the difference (number of lines) between the start addresses of the input band currently being processed and the next input band. Further, in S709, when an input pixel capable of generating an Mth output pixel is input in one vertical line included in a certain input band, the input pixels input after that are ignored.

In step S710, it is determined whether or not the input pixel at the time of generating the Mth output pixel is located at the right end of the input band in one vertical line of the input band. That is, it is determined whether or not all the input pixels for generating the output image data of the M line have been input. Here, if the determination result is Yes, the process proceeds to S712, and if NO, the process proceeds to S711.

In S711, since all the vertical lines of the input band are not input, in each of the input band and the output band, the vertical line to be processed is moved to the right by one line, and then the process proceeds to S701. do. Then, in S704 to S710, the scaling process is performed again from the top pixel of the vertical line.

Further, in S712, after the processing of the following equation 6 is performed, the processing of the input band is terminated.
vface_base = next_vface (Equation 6)
The vphase_base in Equation 6 is the initial phase value vphase in the sub-scanning direction of the next output band.

By performing the above processing, the scaling processing unit 1401 in the present embodiment can always output the output band of the M line subjected to the scaling processing to the input band of the input L line.

In the first and second comparative examples, the number of output lines corresponding to S708 is not determined, and the scaling process is performed using up to the last pixel of each vertical line of the input band. On the other hand, in the present embodiment, when it is determined that the Mth output pixel is output in S708, the input pixel that is not used in the vertical line is ignored, and the next vertical line (right vertical) of the input band is ignored. Shift to processing using line). That is, if all the input pixels of the input band are used, the pixels that should be able to be output in the first output band are extinguished without being output. This disappearing pixel corresponds to the output pixel of the fifth line of the first band in FIG. 6, that is, the pixel of (4).

Further, the initial phase value vphase_base in the next band is not a value calculated by processing the disappearing pixels, but is the next_vphase at the time when the final pixel (Mth pixel) of each vertical line of the output band is output. The value is used (S712). Further, for this band, the input DMAC 103 starts reading from a position advanced by an address corresponding to the number of offset lines from the previous band. Specifically, assuming that the address of the read start position of the previous input band is A and the number of bytes of one input image line is K, the next input band starts reading from the address of (A + offset × K). As a result, the pixels of the horizontal line including the disappearing pixels (pixels of (4) in FIG. 6) can be output by the processing of the next band.

In the above description, the reduction process is taken as an example of the scaling process, but the above process can be similarly applied to the enlargement process and is effective. That is, even in the enlargement process, when a fixed number of input image data is input, the number of lines of the output image data is not constant in the first modification, and variations occur. On the other hand, by applying the above-mentioned processing in the present embodiment, it becomes possible to obtain output image data of a certain number of lines.

(Second Embodiment)
Next, a second embodiment of the present invention will be described. In the above-mentioned first embodiment, it is necessary to change the number of offset lines depending on the input band according to the set value such as magv or the first vphase_base. Therefore, it is necessary to set the address advance amount (offset × K) in the input DMAC 103, but it is difficult to give the input DMAC 103 itself a function of calculating the address advance amount. Therefore, it is considered that the CPU 101 calculates the address advance amount or the read start address value and sets the calculated value in the input DMAC103. However, in this case, a large amount of time is required for the processing of the CPU 101. .. That is, the CPU 101 reads the offset value from the scaling processing unit 1041 after the processing of a certain input band is completed, calculates the address advance amount based on the value, and sets the calculated value in the input DMAC 103. conduct. Therefore, the processing in the CPU 101 takes a lot of time, and the processing time of the entire system may increase accordingly.

In order to avoid such an increase in processing time, the following configuration is adopted in this second embodiment. In FIG. 1, a communication line 1043 (broken line arrow in FIG. 1) from the scaling processing unit 1041 in the scanner image processing unit 104 to the input DMAC 103 is provided as an information transmission means. Then, information related to the number of offset lines obtained by the above-mentioned processing performed in the scaling processing unit 1041 is transmitted via the communication line 1043. The input DMAC 103 calculates an offset address (offset × K [byte]) based on this information, and automatically derives the read start address of the next band using the value.

In this way, by transmitting the information regarding the number of offset lines from the scaling processing unit 1041 to the scaling processing unit 1041, the interrupt response time of the CPU for address calculation and setting and the time required for the calculation of the CPU 101 are ignored. become able to. Therefore, the scanner image processing unit 104 can continuously process one page without stopping for each band.

(Third Embodiment)
Next, a third embodiment of the present invention will be described. In each of the above embodiments, the JPEG compressor 105 is configured to directly input the image data of a fixed number of output lines subjected to the stepless scaling process by the scaling processing unit 1041 included in the scanner image processing unit 104. However, the scaling process itself may be performed in any processing flow as long as it is in the RGB multi-valued state. Therefore, as in the above embodiment, the scaling process is not limited to the scaling processing unit 1041 included in the scanner image processing unit 104 before the JPEG compression, and the scaling is performed in the printer image processing unit 107 after the JPEG decompression. A processing flow for processing is also conceivable.

For example, in the quantization processing required in the final stage of the printer image processing unit 107, a configuration may be adopted in which dither processing is performed in tile units such as 4 × 4 pixels. In this case, the input to the dithering process must be a multiple of 4 lines. Therefore, it is effective to apply the scaling process shown in the above embodiment even in the case of a processing flow in which the stepless scaling process is performed immediately before the dither process. In this embodiment, an example is shown in which the scaling process shown in the first embodiment is performed in the printer image processing unit.

Hereinafter, the image processing apparatus according to the present embodiment will be described with reference to FIG. 8 in comparison with the conventional image processing apparatus.

FIG. 8A is a block diagram showing a conventional image processing apparatus, and FIG. 8B is a block diagram showing an image processing apparatus according to the present embodiment. In the image processing apparatus shown in FIG. 8A, the printer image processing unit 107 is provided with a scaling processing unit 107b that performs the same scaling processing as in the first comparative example described above. On the other hand, in the image processing apparatus of the present embodiment, the printer image processing unit 107 is provided with a scaling processing unit 107b that performs the same scaling processing as that of the first embodiment.

In the conventional image processing apparatus of FIG. 8A, the input DMAC107a of the printer image processing unit 107 after JPEG decompression reads image data from the main memory 130 with a fixed number of lines and a fixed feed amount, and the image data is read from the main memory 130 in the printer image processing unit 107. The scaling processing unit 107b performs the scaling processing. In this case, the number of output lines of the scaling processing unit 107b is variable for each output band. Therefore, the output band output from the scaling processing unit cannot be directly input to the dither processing unit whose processing unit is 4 × 4 tile image data. Therefore, it is necessary to temporarily store the image data in the in-memory 130 for adjusting the number of lines, and then read out the image data of the number of lines that can be processed by the dither processing unit. In this case, the output DMAC107c is required after the scaling processing unit, and the input DMAC107d is required before the dither processing unit 107e, which consumes hardware resources and memory bandwidth.

On the other hand, in the present embodiment shown in FIG. 8A, the scaling processing unit 107b in the printer image processing unit 107 performs the same scaling processing as in the first embodiment to obtain image data having a constant number of output lines. (Output band) can be output. That is, in the present embodiment, after the number of input lines of the input DMAC107a is fixed, among the output lines that can be generated by the scaling processing unit 107b, all the output lines other than the preset number of lines are eliminated. This makes it possible to output a fixed number of image data (output bands) that have undergone scaling processing by the scaling processing unit 107b. Therefore, the scaled image data can be directly supplied to the dither processing unit 107e, which eliminates the need for hardware resources such as DMAC 107c and 107d for adjusting the number of lines and reduces the memory bandwidth. It will also be possible. Further, also in the present embodiment, it is possible to send information on the number of offset lines for each band from the scaling processing unit 107b to the input DMAC107a as in the second embodiment. According to this, since the input DMAC107a can be automatically operated for each band, it is possible to improve the processing time.

(Other embodiments)
In the above embodiment, an example is shown in which the CPU 101 realizes a function as an input setting means for calculating and setting the number of lines of input image data to be input to the scaling processing unit, but the present invention is not limited thereto. The number of lines of image data to be input to the variable magnification processing unit can also be set by the user from an input device or the like provided in the image processing device. However, with the number of input lines set by the user, there is a possibility that the number of lines to be output from the scaling processing unit may not be obtained. You may do so. In addition, when the user sets an excessive number of input lines compared to the number of lines to be output from the scaling processing unit, a notification is given to prompt the user to reduce the number of input lines or change (increase) the scaling factor. You may.

Further, the present invention supplies a program that realizes one or more functions of the above embodiment to a system or an apparatus via a network or a storage medium, and one or more processors in the computer of the system or the apparatus read and execute the program. It can also be realized by the processing to be performed. It can also be realized by a circuit (for example, ASIC) that realizes one or more functions.

100 Image processing device 101 CPU
103 Input DMAC (transfer means)
1041 scaling processing unit (magnification processing means)

Claims (11)

A transfer means for dividing the original image data representing the original image into image data consisting of a certain number of lines along the main scanning direction and transferring the data.
The image data transferred from the transfer means is used as input image data for a processing unit, and the input image data is subjected to scaling processing at a preset magnification to generate scaling image data. Means and
An output setting means for setting the number of output lines to be output from the scaling processing means is provided.
The scaling processing means is characterized in that, of the variable magnification image data that can be generated using the input image data, image data limited to the number of output lines set by the output setting means is output as output image data. Image processing device. An input setting means for setting the number of lines of the input image data to be transferred to the scaling processing means by the transfer means, and an input setting means.
A magnification setting means for setting the magnification of the variable magnification process and
The image processing apparatus according to claim 1, further comprising. The image processing apparatus according to claim 2, wherein the input setting means sets the number of lines of the input image data by using the magnification and the number of output lines. When the scaling processing means outputs the output image data of the number of output lines set by the output setting means according to the input image data of a predetermined processing unit, the output image data is said to be the same. A storage means for storing the phase value in the sub-scanning direction intersecting the main scanning direction is provided.
The scaling processing means outputs the phase value stored in the storage means corresponding to the input image data of the next processing unit in the scaling processing of the input image data of the next processing unit of the predetermined processing unit. The image processing apparatus according to any one of claims 1 to 3, wherein the image processing apparatus is used as an initial phase value of image data. An information transmitting means for transmitting information related to the amount of the input image data used for generating the output image data consisting of the number of output lines set by the output setting means to the transfer means is provided.
The transfer means derives an address for reading the input image data of the processing unit from the memory storing the original image data based on the information, and transfers the input image data of the derived address to the scaling processing means. The image processing apparatus according to any one of claims 1 to 4, wherein the image processing apparatus is to be transferred. The image processing apparatus according to claim 5, wherein the transfer means includes an input DMAC (Direct Memory Access Controller) provided between the scaling processing means and the memory. The output setting means sets the number of lines corresponding to the image processing unit in the post-stage processing means that performs predetermined image processing.
The image processing apparatus according to claim 1, wherein the scaling processing means directly inputs the output image data to the post-stage processing means. The image processing apparatus according to claim 7, wherein the post-stage processing means is a JPEG compression means that performs compression processing on the output image data output from the scaling processing means. The image processing apparatus according to claim 7, wherein the post-stage processing means is a dither processing means for quantizing the output image data output from the scaling processing means. A transfer process in which the original image data representing the original image is divided into image data consisting of a certain number of lines along the main scanning direction and transferred.
The image data transferred by the transfer step is used as input image data for a processing unit, and the input image data is subjected to scaling processing at a preset magnification to generate scaling image data. Process and
It is provided with an output setting step of setting the number of output lines to be output by the scaling processing step.
The variable magnification processing step is characterized in that, of the variable magnification image data that can be generated using the input image data, image data limited to the number of output lines set by the output setting step is output as output image data. Image processing method. Computer,
A transfer means for dividing the original image data representing the original image into image data consisting of a certain number of lines along the main scanning direction and transferring the data.
The image data transferred from the transfer means is used as input image data for a processing unit, and the input image data is subjected to scaling processing at a preset magnification to generate scaling image data. Means and
An output setting means for setting the number of output lines to be output from the variable magnification processing means, and an output setting means.
Among the variable magnification image data that can be generated by the variable magnification processing means using the input image data, the limiting means for outputting the image data limited to the number of output lines set by the output setting means as the output image data, and the limiting means.
A program that is characterized by its function.

Images