JPH05108815A - Picture processor - Google Patents

Abstract

(57) [Summary] [Structure] When converting an original image, a pixel of interest of the original image to be referred to by each pixel after conversion is obtained based on a predetermined enlargement / reduction ratio, and the pixel of interest and A two-dimensional local memory composed of pixels in the vicinity thereof is formed, and
The contents of the three-dimensional local memory are read in parallel, a predetermined operation is performed, and a converted image is generated from the pixel of interest and its neighboring pixels. [Effects] (1) High-speed, flexible, and high-fidelity image conversion (enlargement / reduction) becomes possible. (2) It is possible to display a plurality of CRT screens having different formats on one CRT by newly generating images. Therefore, the cost of the CRT and the installation space can be reduced.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing method.

More specifically, the present invention relates to an image processing method suitable for editing / converting image information in a computer application device and an image processing device.

[0003]

2. Description of the Related Art In recent years, computer application devices and image processing devices often handle not only code information but also image information. In particular, about 80% of the information captured by humans is said to be image (visual) information. Not only in image processing devices but also in computer-applied devices, improvements in the man-machine interface for CRT display and automatic data entry were started. As a result, the need for images is increasing.

An example of such a computer application device and an image processing device is shown in FIG. In this figure, an NTSC display 3 is connected to the image processing apparatus 2 via a cable 4, and an input TV camera 5 is also connected. A high resolution display 6 is connected to the computer application apparatus 1 via a cable 7, and a low resolution display 8 is connected to the computer application apparatus 1 via a cable 9.

At first glance, it may be considered that one display can handle such a plurality of displays, but it cannot be completed by one because the image display format and the cable of each CRT are different.

[0006]

As described above, in the conventional device, (1) a plurality of CRTs and cables are required, which is expensive.

(2) An occupied space for a plurality of displays is required.

Due to the above problems, a suitable image conversion processing technique has been desired.

Further, the conventional image conversion (enlargement / reduction) processing has the following problems.

Here, referring to FIG. 5, the enlargement processing of a pattern having an inclination will be described. In this figure, the original pattern (A) is enlarged to obtain (B) and (C). Here, (B) is a case where the enlargement ratio is 2.3, and D ₁
The white circle of D and the white circle of D ₂ are like this (B). However,
The white circle of D _{1 and} the white circle of D ₃ are supposed to match as a standard.

In FIG. 5B, x = m + 1, y =
The point D ₅ of n is 0 because 1 × (1 / enlargement ratio = 1 / 2.3) = 0.43 is closer to 0 than 1 and the data of X = M and Y = N of the original pattern is referred to. Becomes 0.

Further, at the point D ₄ of x = m + 2, y = n + 2, 2 / 2.3 = 0.87 is closer to 1 than 0, so X = M +
1, it becomes 0 by referring to the data of the point of Y = N + 1.

Further, the point D _{6 at} x = m + 2, y = n is X =
It becomes X by referring to the data at the point of M + 1 and Y = N. Similarly, a white circle of D _{8 and} a white circle of D ₉ are obtained.

As shown in the example of FIG. 5 described above, when the enlargement / reduction processing of the original pattern is performed, the fidelity to the original pattern generally tends to decrease, and the fidelity also changes depending on the enlargement / reduction ratio. easy.

In such enlargement / reduction processing, the image information to be calculated is enormous (resolution vertical 51
Assuming that 2 × horizontal 512 and 8 bits of color / gray scale are 512 × 512 × 3 bytes (a large amount of information), there is a problem that a great deal of processing time is required.

Further, the ratio required in the image conversion processing is not only a simple integral multiple, for example, 256 pixels should be 650 pixels (that is, enlargement ratio = 650/256 = 2.54).

Therefore, in view of the above points, the object of the present invention is to
An object of the present invention is to provide an image processing method that enables the items (a) and (b) listed below.

(B) The original image can be converted (enlarged / reduced) faithfully, at high speed, and flexibly (conversion freely).

(B) To enable favorable editing of a plurality of images having different formats. That is,
Multiple CRT screens with different formats (computer application device and image processing device screens) are combined into one CRT
Do it on screen. Put multiple screens in one CRT screen.

[0020]

According to the present invention, when converting an original image, a pixel of interest of the original image to be referred to by each converted pixel is obtained based on a predetermined enlargement / reduction ratio, A two-dimensional local memory including the pixel of interest and pixels in the vicinity thereof is formed, the contents of the two-dimensional local memory are read in parallel and a predetermined operation is performed, and a converted image is generated from the pixel of interest and pixels in the vicinity thereof. To do.

Further, by using the present invention, it is possible to temporarily store "a plurality of CRT images having different formats" in an image memory and newly generate and display one CRT image. ..

Further, by using the present invention, "plurality of image data" are temporarily stored in the image memory, and the position where the upper left coordinates of each of the plural images occupy in one screen is the horizontal address of the frame image memory. It is also possible to edit the image by setting the counter and the vertical address counter and generating and displaying a plurality of screens in one screen.

[0023]

According to the present invention, the pixel referred to at the time of image conversion is the pixel of interest and the pixels in the vicinity thereof, so that the conversion process with high fidelity to the original image becomes possible.

Further, by implementing a hardware as a computing means for performing a predetermined computation, a huge amount of image data can be processed at high speed.

Furthermore, since the image conversion ratio can be set independently in the vertical and horizontal directions and can be set to an arbitrary value, flexibility is enhanced.

[0026]

EXAMPLES Examples of the present invention will be described in detail below.

FIG. 1 is a block diagram showing an embodiment of the present invention. In this figure, 1 is a frame image memory, 2
Is a horizontal address counter, 3 is a vertical address counter,
4 is an address counter data table, 5 to 7 are shift registers, 8 is data of x _i-1 , y _j-1 , and 9 is x _i ,
data of y _j−1 , 10 is data of x _{i + 1} , y _j−1 , 11
Is data of x _i-1 and y _j , 12 is data of x _i and y _j ,
13 is data of x _{i + 1} , y _j , 14 is data of x _i-1 , y _{j + 1} , 15 is data of x _i , y _{j + 1} , 16 is x _{i + 1} ,
y _{j + 1} data, 17 is an arithmetic circuit, 18 is an arithmetic circuit data set section, and 19 is a CPU data bus.

Next, the operation of this embodiment will be described.

Original images of different formats are temporarily stored in the frame image memory 1. Here, the image conversion (enlargement / reduction) is performed as shown in FIG.

First, the points of the image after conversion are set to P ₂ (x _r , y
_s ). The image data of this point (x _r , y _s ) is the point P
_{It is obtained} by referring to ₁ (x _i , y _j ). This x _i is x _r
X (1 / a), y _j are y _s × (1 / b) (where a and b are the enlargement / reduction ratios in the horizontal and vertical directions), and are rounded to the integer coordinate values. is there.

The coordinates (X _i , Y _j ) of the original image to be referred to for each converted point are obtained and stored in the address counter data table 4 shown in FIG. For this, it is calculated and stored in advance by using a numerical processor.

Next, the formation of the "two-dimensional local memory" will be described. First, the data in the counter data table 4 shown in FIG. 1 is sequentially read out, and the horizontal address counter 2 and the vertical address counter 3 have coordinates x _i to be referred to,
y _j is designated, the data is read and input to the shift register 5. The shift registers 5, 6 and 7 have one row each, and their parallel outputs are 8, 9, 10 and 11, 1 in FIG.
2, 13 and 14, 15, 16. The parallel output thus obtained is the "two-dimensional local memory" shown in FIG.
To form.

In the two-dimensional local memory shown in FIG. 3, the data of the coordinates around the coordinate (x _i , y _j ) to be focused is output in parallel. This parallel output is output in real time and sequentially over the entire screen. This parallel output is input to the arithmetic circuit 17 shown in FIG. 1 and a predetermined logical operation is executed. The parameter data for calculation is preset in the parameter data setting section 18 via the CPU data bus 19.

Thus, the CPU data bus 19 outputs the image data of the coordinates (x _i , y _j ) to be focused on and the pixels in the vicinity thereof. Thus, the coordinates (x _i ,
Since not only the data of one point of y _j ) but the data in the vicinity thereof are also referred to, the fidelity of the image data after the enlargement / reduction processing becomes good.

Although the processing of FIG. 1 can be performed by software, it has a hardware configuration.
Processing time will be shorter. That is, in software processing, parallel processing of the "two-dimensional local memory" as shown in FIG. 3 cannot be performed and serial processing is performed, so that a long processing time is required.

The enlargement / reduction ratios a and b shown in FIG. 2 may be arbitrary values and can be set separately in the horizontal and vertical directions. Therefore, flexible enlargement / reduction is possible.

From the above, the conversion (enlargement / reduction) of images with high speed, flexibility and high fidelity becomes possible.

The above method is to temporarily store the original image and generate a new image. Therefore, it is possible to newly generate an image even with "data of a plurality of CRT images having different formats" and display the image on one CRT.

Further, according to the embodiment of the present invention, it is possible to perform an "edit" process for generating a plurality of screens in one screen. In this case, it suffices to shift the values of the horizontal counter and the vertical counter by the amount (Δx, Δy) of the upper left coordinates of the plurality of screens.

In image processing, it is often the case that the "image pattern to be inspected" and the "standard image pattern" are compared and collated. The present invention can also be used for As a result, it is possible to perform the normalization processing / collation processing at higher speed.

[0041]

[Effects of the Invention] By implementing the present invention, the following particular effects can be obtained.

(1) High-speed, flexible, high-fidelity image conversion (enlargement / reduction) is possible.

(2) It is possible to display a plurality of CRT screens of different formats on one CRT by newly generating images. Therefore, the cost of the CRT and the installation space can be reduced.

(3) A plurality of screens can be generated in the CRT screen, and "editing of images" is possible.

(4) The collation process between image patterns can be performed at high speed.

[Brief description of drawings]

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

FIG. 2 is an explanatory diagram showing image conversion (enlargement / reduction) in the present embodiment.

FIG. 3 is an arbitrary point (x _i , y) on the screen in the present embodiment.
FIG. 3 is a diagram showing a two-dimensional local memory centering on _j ).

FIG. 4 is a block diagram showing an example of a conventionally known image processing system.

FIG. 5 is a diagram showing a conventional technique for enlarging a tilted pattern.

[Explanation of symbols]

1 frame image memory 2 horizontal address counter 3 vertical address counter 4 address counter data table 5 to 7 shift register 8 x _i-1 , y _j-1 data 9 x _i , y _j-1 data 10 x _{i + 1} , Y _j-1 data 11 x _i-1 , y _j data 12 x _i , y _j data 13 x _{i + 1} , y _j data 14 x _i-1 , y _{j + 1} data 15 x _i , Y _{j + 1} data 16 x _{i + 1} , y _{j + 1} data 17 arithmetic circuit 18 arithmetic circuit data set unit 19 CPU data bus

Claims (1)

[Claims] 1. When converting an original image, a pixel of interest of the original image to be referred to by each pixel after conversion is obtained based on a predetermined enlargement / reduction ratio, and the pixel of interest and its neighboring pixels. Image processing, wherein a two-dimensional local memory is formed, the contents of the two-dimensional local memory are read in parallel, a predetermined operation is performed, and a converted image is generated from the pixel of interest and pixels in the vicinity thereof. Method.