JPH0743769B2 - Digital image processor - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a digital image processing device, and more particularly to a digital image convolution (superimposition integration) process to enhance or suppress a component in a specific frequency band. Then, the present invention relates to an improvement of means for constructing a filter function in a two-dimensional matrix when forming a clear image.

(Conventional Technology) Convolution (superposition integration) processing is one of the grayscale image processing methods which is one of digital image processing technologies. This is a method of performing a product-sum operation in parallel at the same time by using the pixel data of the central pixel to be processed and the pixel data in the vicinity thereof.By changing the weighting coefficient, it is used for processing such as differentiation, averaging, and emphasis. It is effective for line segment extraction, noise removal, and edge enhancement. A digital image processing device for performing such convolution processing is composed of a pre-processing unit, a convolution processing unit, a display unit, a control unit, and an image display TV monitor, and the pre-processing unit and convolution processing unit and The display unit sends and receives image data via the video bus, and it also communicates with the PC in the control unit using GPIB (General Purpose In
terface Bus: general-purpose interface bus).

The analog video signal sent as image data is converted into a digital video signal in the pre-processing unit, subjected to image processing, and sent to the display unit. In the convolution processing unit, the digital video signal is subjected to a matrix operation using a filter function composed of a two-dimensional matrix in the convolution processing unit to emphasize or suppress a component in a specific frequency band, and then the digital video signal. Is converted into an analog video signal and sent to an image display TV monitor, where it is displayed as a clear image.

In order to create a filter function composed of a conventional two-dimensional matrix in such a device, first, the keyboard of the control unit is operated to select a preset matrix size for the control TV monitor of the control unit. Project. For this matrix size, an odd number is usually selected from a matrix of 3 × 3 to 15 × 15. In order to perform convolution processing in the matrix displayed on the control TV monitor, it is performed by inputting a numerical value as a filter function using the numeric keys of the keyboard of the control unit.
By performing the above operation, the image on the image display TV monitor is displayed with being emphasized or suppressed. This process is completed when the central pixel in the matrix size scans one screen.

(Problems to be Solved by the Invention) However, performing the convolution processing by such a method requires an operation of inputting a filter function to each element in the matrix, and thus the matrix size is the minimum. Even in such a case, it is necessary to input 9 points in total in 3 × 3. If the matrix size becomes large, it takes time to input that much, and the matrix size cannot be selected too large.

Further, when the desired image cannot be obtained, the input operation of the filter function is repeated many times by trial and error, but this has a problem that it takes more input time.

In view of the above problems in the conventional technique, an object of the present invention is to perform an input operation of a filter function for each element of the matrix one by one, and instead of performing an input operation for each element of the matrix, an appropriate predetermined pattern for a part of all elements of the matrix. By inputting a filter function with, and copying one block of this input element (one element is the smallest in the case) into another portion of other non-input elements according to a predetermined rule,
In some cases, the first input portion and the first copy portion are combined and copied to another portion, so that a means for inputting the filter function into the matrix and a time-saving digital image can be saved. It is to provide a processing device.

(Means for Solving the Problem) The present invention has the following means configuration in order to achieve the above object.

That is, the digital image processing apparatus according to the present invention is a digital image processing apparatus that performs convolution processing with a filter function formed of a two-dimensional matrix for each pixel of digital image data to enhance or suppress a specific image component. , Means for generating a plurality of matrix sizes and means for setting an arbitrary function in the matrix,
Means for copying one or more of the set functions into the matrix according to a preset copy mode.

(Operation) The operation of the digital image processing apparatus of the present invention having the above means configuration will be described below.

The device of the present invention has means for generating a plurality of matrix sizes, and a matrix of a desired size can be set by operation.

A filter function can be input to each element of the matrix thus set by means of setting a function. This filter function input can be applied to the appropriate part of all elements of the matrix. Thus
When the filter function is input to a part of the matrix,
By means of copying, it is possible to copy to a blank portion in the matrix.

This copying is performed in a predetermined mode, for example, when the axis in the row direction passing through the center of the matrix is the X axis and the axis in the column direction is the Y axis, copying is performed symmetrically with respect to the X axis (X axis symmetrical mode), Copying is performed in a mode of symmetrically copying with respect to the Y-axis (Y-axis symmetrical mode), a mode of symmetrically copying with respect to both the X-axis and the Y-axis (XY-axis symmetrical mode), and other modes.

In this way, when the filter function is input to a part of the matrix of the desired size in an appropriate pattern, the filter function is then repeated for all the elements of the matrix by repeating the copying to the non-input part of the matrix in various copy modes. Will be input. Also, when copying,
Not only can the first input part be repeatedly input to another different part, but the first input part and the first copied part can be combined and copied to another part. By repeating the same operation, The time and effort are reduced.

(Example) Hereinafter, an example of a digital image processing apparatus of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of an embodiment of the digital image processing apparatus of the present invention.

This device is composed of a front-end processing unit 1, a convolution processing unit 6, a display unit 2, an image display TV monitor 3 and a control unit 4. The control unit 4 further includes a matrix size setting unit 8, a filter function input unit 9, a copying unit 10, a keyboard 7 and a control TV monitor 5.

The analog video signal sent as image data is converted into a digital video signal in the preprocessing unit 1 and various image processing is performed.

However, even in this image, the processing of green such as contours is incomplete. Therefore, the convolution processing unit 6 is inserted in front of the display unit 2 to perform the convolution processing,
Processes a blurred image into a Supe image. For that purpose control TV
The matrix size is displayed on the monitor 5, the copy mode is selected, and the matrix function is input.

First, the matrix size is input by the keyboard 7 of the control unit 4, and the matrix of the desired size is determined by the matrix size setting means 8 and displayed on the control TV monitor 5.

Next, one of the copy modes is selected with the keyboard 7. The matrix function is input to the predetermined area in the two-dimensional matrix displayed on the control TV monitor 5 by the filter function input means 9 by operating the numeric keys of the keyboard 7. Then, according to the copy mode selected in the above, the matrix function is copied by the copying means 10 to another uninput portion in the matrix, and the matrix function is completed.

This matrix function is sent to the convolution processing unit 6, and the convolution processing unit 6 performs the convolution processing using the matrix function, and the contour-corrected image passes through the display unit 2 and the image display TV monitor. It is projected on 3. In this case, if the desired contour correction of the image cannot be obtained, the above-described procedure is repeated to perform the desired contour correction.

Next, a method for copying a matrix function by inputting several characters of the matrix function according to a copy mode selected in advance in the convolution processing of the present apparatus will be described. First, as shown in FIGS. 2 (a) to (c),
A matrix size that is divided into an equal number of X and Y and an odd number is created and stored in advance, and a matrix of a size corresponding to the state of the image to be processed is selected from this matrix size.

Then, the copy mode is designated, and the function to be copied (copied) in the copy mode is input from the keyboard 7 to a predetermined position in the selected matrix.

An example of the copy mode in the present invention is as follows.

(1) XY axis symmetry mode (Fig. 3 (a)) The upper left 1/4 part is symmetrically copied to the other part. FIG. 4 (a) shows a concrete example of matrix size 5 × 5. Assuming that the upper left corner of the matrix is address 0, the keyboard 7 is used to (0,0), (0,1), (0,2), (1,0),
Arbitrary filter functions are input to 9 addresses of (1,1), (1,2), (2,0), (2,1) and (2,2). Then, with the X axis of (0,2), (1,2) and (2,2) as the axis of symmetry, (0,0) becomes (0,4) and (0,1) becomes (0,3). To (1,0)
To (1,4), (1,1) to (1,3), (2,0) to (2,4)
, (2,1) is copied to (2,3), and then (2,0), (2,3)
1), (2,2), (2,3) and (2,4) with the Y axis as the axis of symmetry, (0,0) becomes (4,0) and (1,0) becomes (3, 0) to (0,
1) to (4,1), (1,1) to (3,1), (0,2) to (4,1)
2), (1,2) to (3,2), (0,3) to (4,3),
By copying (1,3) into (3,3), (0,4) into (4,4), and (1,4) into (3,4), the filter function into the matrix Numerical value input ends. XY in Figure 5
9 shows a flowchart of a copying operation in an axially symmetric mode.

(2) X-axis symmetrical mode (Fig. 3 (b)) The upper half is copied to the lower side. FIG. 4B shows a concrete example of matrix size 3 × 3. Use the keyboard 7 to enter any filter function at 6 addresses (0,0), (0,1), (1,0), (1,1), (2,2) and (2,1). . Then, with the X axis of (0,1), (1,1) and (2,1) as the axis of symmetry, (0,0) becomes (0,2) and (1,0) becomes (1,2).
By copying (2,0) to (2,2),
Numerical input of the filter function into the matrix ends. FIG. 6 shows a flow chart of the copying operation in the X-axis symmetrical mode.

(3) Y-axis symmetrical mode (FIG. 3 (C)) The left half is copied to the right side. FIG. 4 (C) shows a concrete example of matrix size 3 × 3. Use the keyboard 7 to enter any filter function at 6 addresses (0,0), (0,1), (0,2), (1,0), (1,1) and (1,2). . Then, with the Y axis of (1,0), (1,1) and (1,2) as the axis of symmetry, (0,0) becomes (2,0) and (0,1) becomes (2,1).
By copying (0,2) to (2,2) respectively,
Numerical input of the filter function into the matrix ends. FIG. 7 shows a flowchart of the copying operation in the Y-axis symmetrical mode.

(4) Circuit mode (Fig. 3 (d)) The upper half data on the Y-axis located in the center of the matrix is copied one round. FIG. 4 (d) shows a concrete example of matrix size 5 × 5. With keyboard 7 (2,
Input an arbitrary filter function to each of the addresses 0), (2,1) and (2,2). And (2,0) starts in (0,0) in two directions of X and Y and ends in (4,4) in two directions of X and Y, that is, the outermost green of the matrix is all (2,0). The data input to is copied so that the data of (2,1) is also copied in the second round. Since the data of (2,2) is the center, the data as it is becomes the filter function. FIG. 8 shows a flow chart of the copying operation in the circuit mode.

(Effects of the Invention) As described above, in the device of the present invention, the filter function input to a partial area in the matrix is copied to another non-input area so that the filter function is input to the entire matrix. Therefore, there is no need to perform an input operation by keyboard operation for all the elements of the matrix, and there is an advantage that the time required for inputting the convolution processing data can be significantly shortened.

This effect becomes more remarkable as the matrix size increases. Also, in the past, when setting a function in a matrix, it was time-consuming to create a large matrix because all of the matrix was entered, but with this device, the input time can be shortened, and therefore the matrix size can also be reduced. It has the advantage that it can be made larger.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of an apparatus according to an embodiment of the present invention,
FIG. 2 is a diagram showing an example of matrix size, and FIG.
Is a 3 × 3 size, FIG. 5 (b) is a 5 × 5 size, and FIG. (C) is a 11 × 11 size. FIG. 3 is a schematic diagram of a matrix copy mode. Axisymmetric mode, Figure (b) is X axis symmetrical mode, Figure (c) is Y axis symmetrical mode, Figure (d)
FIG. 4 is a diagram showing a circuit mode, FIG. 4 is a matrix showing a concrete example of a copy mode, FIG. 4A is a XY axis symmetrical mode diagram, FIG. 4B is an X axis symmetrical mode, and FIG. Axisymmetric mode, FIG. 5D is a diagram showing a matrix in circuit mode, FIG. 5 is a flowchart showing a copying operation in the XY axis symmetric mode, FIG. 6 is a flowchart showing a copying operation in the X axis symmetric mode, and FIG. Is a flow chart showing the copying operation in the Y-axis symmetrical mode, and FIG. 8 is a flow chart showing the copying operation in the circuit mode. 1 ... Pre-processing unit, 2 ... Display unit, 3 ... Image display TV monitor, 4 ... Control unit, 5 ... Control TV monitor, 6 ... Convolution processing unit, 7 ... Keyboard, 8 ...... Matrix size setting means, 9 ... filter function input means, 10 ... copying means.

Claims (1)

[Claims] 1. A digital image processing apparatus for enhancing or suppressing a specific image component by performing a convolution process with a filter function formed of a two-dimensional matrix for each pixel of digital image data. A means for generating, a means for setting an arbitrary function in the matrix, and a means for copying one or a block of the set functions in the matrix according to a preset copy mode. Digital image processing device.