JPH11205576A - Image processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To fast and inexpensively perform rotation, scaling, parallel translation processing of an image by using a small amount of capacity of memory. SOLUTION: A data transfer controlling part 30 removes data after conversion procession from image data stored in line buffer memory 20 and successively writes input image data to the memory 20. An address calculating part 80 performs raster scan of output buffer memory 40, calculates the address of the memory 20 that corresponds to the address of the memory 40 and preserves it in address buffer memory 50. An image data converting part 60 writes image data read from the memory 20 based on the address of the memory 50 to the memory 40. An image outputting part 70 takes out image data after conversion which is stored in memory 40 and resets the memory 40.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an image processing apparatus capable of executing affine transformation functions such as rotation, enlargement, reduction, and translation of an image with a simple configuration.

[0002]

2. Description of the Related Art Generally, image rotation processing can be realized by performing a matrix operation represented by the following equation.

[0003]

(Equation 1) This is called an affine transformation. When affine transformation is performed on an image processing apparatus, generally, raster scanning is performed on an output image page memory that stores image data after affine transformation processing, and an address of an input image page memory corresponding to an address of the output image page memory is stored. Then, the image data read from the input image page memory is written to the output image page memory. In this case, since the calculation is performed based on the output image page memory, pixel omission does not occur.

The address of the image memory is a set of two integers on two-dimensional coordinates as shown in equation (1) (x,
y). From the address (x, y) of the output image page memory, the address (U, V) of the input image page memory is calculated by equation (1). In equation (1), since U and V are not always integers, pixel data at the nearest point of (U, V) in the input image page memory is used as data to be written to the output image page memory. Of course, as a conventional example in which interpolation may be performed using pixel values at a plurality of input pixel positions, see Japanese Patent Application Laid-Open No. 61-161.
The technology described in Japanese Patent No. 576 will be described. In the conventional example, the image is processed locally, that is, in units of blocks, thereby speeding up the rotation process. Hereinafter, a conventional example will be described with reference to FIG. In FIG. 4, words and the like are changed to conform to the description of the present invention, but there is essentially no difference from the description in the publication of the conventional example. 4, reference numeral 10 denotes an image input unit, 29 denotes an inverse transformation coordinate calculation unit,
8 is a page memory reading unit, 49 is an input image page memory, 52 is a conversion ROM, 61 is a local data conversion unit, 7
8 is a page memory writing unit, 79 is an output image page memory, 80 is an image output unit, 200 is input image data,
28, input image address information; 229, inverse transformed coordinates;
30 is local data, 231 is conversion information of a conversion ROM, 2
40 is converted local data, 269 is output image address information, and 270 is output image data.

[0005] The image input unit 10 stores the input image data 200 in the input image page memory 49. The inverse transformed coordinate calculation unit 29 calculates coordinates in the input image obtained by inversely transforming the center pixel of the local data in the output image, and calculates the inverse transformed coordinates 22
9 is output. The page memory read unit 48 sends the input image address 228 of the local data centered on the inverse transformation coordinates 229 to the input image page memory 49. The input image page memory 49 stores the input image data 200, and stores the local data 230 according to the input image address 228.
To the local data converter 61. Based on the conversion information 231 in the conversion ROM 52, the local data conversion unit 51 performs conversion processing such as rotation, enlargement, and reduction on the read local data 230, and outputs the converted local data 240 to the output image page memory 79. Page memory writing unit 7
8 is an output image address 26 corresponding to the inverse transformed coordinates 229
9 to the output image page memory 79. The output image page memory 79 stores the converted local data 240,
Output image data 270 according to output image address 269
To the image output unit 80. The image output unit 80 outputs the output image data 270 to the outside.

[0006] The rotation processing in the above conventional example is performed by a conversion R
It is characterized in that it is performed by two processes: a local process based on the OM 52 and a global process for obtaining the address of the local data of the rotation destination. According to this conventional example, high-speed rotation processing is possible. That is, since the former local processing is simplified, the speed can be increased by H / W or the like, and the wide area processing can reduce the number of operations itself. However, the conventional example is premised on having two page memories, and since the page memories are generally expensive, there is a problem that the cost of the entire apparatus increases.

[0007]

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and uses a small amount of memory without having a page memory to rotate, scale, and translate these images at high speed and at low cost. The purpose is to enable processing.

[0008]

According to the present invention, in order to achieve the above object, an image input device for inputting an image and a plurality of lines of input image data in a raster scanning order are temporarily stored in an image processing apparatus. A line buffer memory for storing, data transfer control means for transferring first image data from the input image to the line buffer memory, and performing a predetermined conversion on the first image data stored in the line buffer memory Output buffer memory for storing the second image data generated as a result, address calculation means for calculating the address of the line buffer memory while raster-scanning the output buffer memory, and address buffer memory for storing the calculated address And outputting the image data stored in the line buffer memory from the address buffer. Based on the address, so that provided the image data converting means for writing to the output buffer memory, and an image output means for outputting the image data stored in the output buffer memory.

In this configuration, the data in the line buffer memory can be transferred to the output buffer memory using the address generated by the address calculation means, and a predetermined conversion can be performed. Moreover, the address needs to be calculated only once, and the amount of calculation does not increase.

[0010]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of the image processing apparatus according to the present invention. In FIG. 1, 10 is an image input unit, 20 is a line buffer memory, 30 is a data transfer control unit, 40 is an output buffer memory, 50 is an address buffer memory, 60 is an image data conversion unit, 70 is an image output unit, and 80 is Address calculation unit.

The image input unit 10 sends the input image data 1 to the line buffer memory 20 line by line in the raster scanning order. The line buffer memory 20 stores image data for n lines. For example, in FIG.
5 shows a line buffer memory for storing image data for six lines.

The data transfer control section 30 removes the data after the conversion processing from the image data stored in the line buffer memory 20, sequentially writes the input image data in the line buffer memory 20, and waits for the conversion processing. First, k lines of image data are deleted from the head of the line buffer memory 20, and the remaining image data is moved to the head of the line buffer memory 20. Next, image data for k lines is read from the input image data and written to the back of the line buffer memory 20. Here, the value of k is calculated based on the rotation angle and the size of the output buffer memory.
For example, for 16 lines of image data shown in FIG.
First, the image data for eight lines from the beginning is deleted, and the remaining eight lines are deleted.
The image data for the line is moved to the head of the line buffer memory, and then the image data for eight lines is read into the line buffer memory 20 from the input image data.

The output buffer memory 40 stores m lines of converted image data. The address calculator 80 performs a raster scan of the output buffer memory 40 to obtain an address of the line buffer memory 20 corresponding to the address of the output buffer memory 40. Here, assuming that the length of one line of image data is w, m × w calculations are required to calculate all the addresses of the output buffer memory 40.
The address buffer memory 50 is the output buffer memory 40
M × w addresses are stored in raster order. The image data conversion unit 60 writes the image data read from the line buffer memory 20 to the output buffer memory 40 based on the address of the address buffer memory 50.

The image output unit 70 includes an output buffer memory 40
Is written in the output image, the image data cached in the output buffer memory is cleared, and the converted image data is sequentially obtained from the line buffer memory 20.

FIG. 3 illustrates an operation example of this embodiment. Here, it is assumed that the input image 300 is rotated to become the output image 400. First, the image data in the area 301 in FIG. The input image area 301 is the output image area 402
Corresponding to The number of lines in the area 301 of the input image is n, and the number of lines in the area 301 of the output image is m. At this time, a chain line window 311 is stored in the line buffer memory 20. It is assumed that the input image is newly written every k lines. The values of n and k are set in the output image area 301 (output buffer memory 40).
Is determined depending on the size and the rotation angle. Corresponding reference numerals are given to other input image areas and output image areas.

The image data of the area indicated by the scatter points in the area 301 of the input image is written to the area 401 (output buffer memory 40) of the output image. The image data written in this way is further transferred to the image output unit 70, and then new image data can be written to the output buffer memory 40.

Thereafter, since the region indicated by k in FIG. 3A is not used for the subsequent rotation processing, the image data for k lines is removed from the line buffer memory 20. Then, the image data for the remaining (nk) lines is shifted by k lines. Next, new image data for k lines is transferred to the line buffer memory 20. By this transfer, FIG.
As shown in (b), the next input image area 302 is held in the line buffer memory 20 (window 312). Since the scattered portion of the input image area 302 corresponds to the area 402 of the output image, it is newly written in the area 402 (output buffer memory 40) of the output image. Then, the image data is further transferred to the image output unit 70.

Similarly, the areas of the input image are sequentially processed, and the areas 403 and 404 of the output image are generated and sent to the image output unit 70 in sequence.

The entire image is processed by repeatedly using the line buffer memory 20 and the output buffer memory 40. The line buffer memory 20 and the output buffer memory 20 may have a small capacity, and the configuration can be simplified. Further, the address required for writing data from the line buffer memory 20 to the output buffer memory 40 is the same in the repetitive processing, so that it is only necessary to calculate once.

In the above-described embodiment, the line buffer memory 20 and the output buffer memory 40 are actually mounted. However, the line buffer memory 20 and the output buffer memory 40 are logically mapped to predetermined storage means. You may. In this case, it is necessary to further convert the address of the address calculation unit 50 described above.

In the above-described embodiment, the rotation processing has been described. However, the rotation processing can be widely applied to the affine transformation, and the clipping processing, the color conversion, and the like may be performed simultaneously with the affine transformation.

[0022]

As described above, according to the present invention,
When the image data is cached by the line buffer memory and the output buffer memory is scanned in the raster mode, the input image data corresponding to the converted data is obtained from the cache data, and the rotation processing is performed by the output buffer memory having a small capacity. it can. The cache data is stored in the line buffer memory for a certain period, and image quality can be guaranteed. Further, by using the address buffer memory, the address of the output buffer memory and the address of the line buffer memory can be associated with one address calculation, and the address calculation time can be reduced.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an embodiment of the present invention.

FIG. 2 is a diagram illustrating data transfer in a line buffer memory according to an embodiment of the present invention.

FIG. 3 is a diagram showing an operation example of the embodiment of the present invention.

FIG. 4 is a diagram showing a configuration of a conventional example.

[Explanation of symbols]

 Reference Signs List 1 input image data 10 image input unit 20 line buffer memory 30 data transfer control unit 40 output buffer memory 50 address calculation unit 60 image data conversion unit 70 image output unit 80 address buffer memory

Claims (5)

[Claims] An image input unit for inputting an image; a line buffer memory for temporarily storing input image data for a plurality of lines in raster scanning order; and data for transferring first image data from the input image to the line buffer memory. Transfer control means, an output buffer memory for storing second image data generated by performing a predetermined conversion on the first image data stored in the line buffer memory, and raster scanning the output buffer memory Address calculation means for calculating the address of the line buffer memory, an address buffer memory for storing the calculated address, and converting the image data stored in the line buffer memory into an address output from the address buffer. Image data conversion to write to the output buffer memory based on Means for outputting image data stored in the output buffer memory. 2. The image processing apparatus according to claim 1, wherein the predetermined conversion includes at least a rotation process. 3. The data transfer control means includes means for moving image data in the line buffer memory, and means for sequentially writing k lines of image data from the input image into the line buffer memory, 3. The image processing apparatus according to claim 2, wherein the value of k is calculated based on a rotation angle and a size of the output buffer memory. 4. The image output means writes image data stored in the output buffer memory to an output image,
4. The image processing apparatus according to claim 1, further comprising means for resetting an output buffer memory. 5. An image processing apparatus for generating an output image by performing a predetermined conversion on an input image, comprising: an input buffer memory for temporarily storing data of a partial image of the input image; Means for sequentially transferring image data to the input buffer memory; an output buffer memory for temporarily storing data of the partial image of the output image; and sequentially outputting data of the partial image of the output image from the output buffer memory. Means for generating correspondence data for associating the address of the output buffer with the address of the input buffer memory according to the content of the predetermined conversion; and the input buffer memory based on the correspondence data. The data of the partial image of the output image temporarily stored in the Means for writing to the output buffer memory.