JPH06333032A - Image rotation procesing method and device - Google Patents

Abstract

PURPOSE:To keep the high quality of the linear edges of an original picture even when the digital multilevel image data including a pattern having the sharp linear edges is rotated at an optional angle. CONSTITUTION:A rotated image address generator 5 calculates the coordinates of an original picture corresponding to the picture element coordinates of a rotated image. An original image point segment length generator 6 calculates the coordinates of two points near the picture element whose segment corresponding to the main scanning line of the rotated image passes through the coordinates set on the original picture and crosses the lattice of the original picture. An interpolation processor 8 calculates by interpolation the density of two picture elements near lattice points of those two points together with the distance between both picture elements based on the threshold value conditions given from a threshold value generator 9. Thus the density of each intersecting point is calculated. Furthermore the density of the original picture coordinates is interpolated under the same threshold value conditions and based on the interpolation density of those two points and the distance from the original picture coordinates.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image rotation processing method and apparatus, and more particularly to an image rotation processing method and apparatus for rotating a digital gradation image at an arbitrary angle.

[0002]

2. Description of the Related Art Image rotation processing is a technique called affine transformation in the image processing technical field, and various methods have been developed. Affine transformation is a method for describing rotation of a two-dimensional coordinate system by polar coordinate transformation.

By the way, the position coordinates of the pixels of the digital gradation image data stored in the digital computer are on the grid points of the integer value coordinates of the two-dimensional coordinate system. Affine-transformed coordinates of this original image are generally not mapped on the grid points of the rotating coordinate system. Therefore, the density value of the pixel on the grid point of the rotating coordinate system is obtained by performing interpolation processing from the density values of the corresponding original image pixel and its neighboring pixels.

[0004]

When digital gradation image data including a pattern of linear edges having a steep density difference is subjected to Affin conversion, the linear edges of the rotated image are stairs and are visually difficult to see. Therefore, there is a problem that the image quality is extremely deteriorated.

Therefore, a main object of the present invention is to preserve the linear edge of the original image with high quality even if the digital gradation image data including a picture having a steep linear edge is rotated at an arbitrary angle. It is an object of the present invention to provide an image rotation processing method and apparatus capable of performing the above.

[0006]

The invention according to claim 1 is
An image rotation processing method for image rotation processing of digital gradation image data at an arbitrary angle, wherein a step of obtaining coordinates of an original image corresponding to pixel coordinates of the rotation image and a line segment corresponding to a main scanning line of the rotation image A step of obtaining coordinates of two pixel neighborhoods passing through the coordinates on the original image and intersecting the lattice of the original image, and density values of two neighboring pixels on each of the two lattice points and threshold conditions of their distances. And the step of interpolating the density value of each intersection, and the step of interpolating the density value of the original image coordinate under the same threshold value condition from the calculated interpolated density value of the two points and the distance of the original image coordinate.

According to a second aspect of the present invention, there is provided an image rotation processing device for performing image rotation processing of digital gradation image data at an arbitrary angle, the image rotation angle input means for inputting an image rotation angle, and the image rotation processing. In response to the input of the image rotation angle from the angle input means, original image coordinate generation means for generating a signal indicating the coordinates of the original image corresponding to the pixel coordinates of the rotated image, and a line segment corresponding to the main scanning line of the rotated image. Is a coordinate calculation means for obtaining coordinates of two points in the vicinity of a pixel which passes through the coordinates on the original image based on the signal generated by the original image coordinate generating means and intersects with the grid of the original image, and First density interpolating means for interpolating the density value of each intersection according to the density value of two pixels and each threshold value condition of the distance, the interpolated density value of the two obtained points, and the distance of the original image coordinates From the same threshold The matter, and comprises a second concentration interpolation means for interpolating the density values of the original coordinates.

[0008]

When the image of the digital gradation image is rotated, the coordinates on the original image corresponding to the pixel coordinates of the rotated image are generally not mapped on the grid points of the original image. Therefore, in the image rotation processing method and apparatus according to the present invention, in order to store the rotation information of the image, the main scanning line of the rotated image passing through the corresponding coordinates on the original image and the horizontal or vertical grid of the original image and the rotation angle θ.
Then, the neighborhood coordinates of two points intersecting with each other are obtained. Then, it is determined whether the density value of each intersection coordinate is a steep edge from the distance between the coordinates of two points on the grid point near each intersection coordinate and each density value, and the edge is saved. Threshold processing for interpolation. Further, it is determined whether or not the edge is a steep edge in the same manner from the density value of the threshold value interpolation of these two points and the distance between the original image corresponding coordinate and the intersecting coordinate, and by the threshold value interpolation performing the edge preservation. ,
Obtain the interpolated density value of the coordinates corresponding to the original image.

[0009]

First, the principle of the present invention will be described before the description of the embodiments.

FIG. 2 is a schematic diagram of monochrome / digital gradation image data. In FIG. 2, the horizontal direction h of the XY coordinate system is the X axis, and the vertical direction v is the Y axis. The image is composed of X ₀ pixels P in the horizontal direction and Y ₀ pixels in the vertical direction,
Each pixel is located on a grid point which is an integer position coordinate of the XY coordinate system.

FIG. 3 is a schematic diagram when the pixels of the rotated image obtained by rotating the digital gradation image data of the original image by the image rotation angle θ and the main scanning line passing through the point are superimposed on the original image. In FIG. 3, the point A (α1, β1) is the coordinate when the coordinate of the pixel P1 point in the rotated image is mapped to the original image. The rotation image main scanning line of the pixel P1 is
2 near point A (α1, β1) that intersects the vertical grid of the original image
The points are B point (x1, β2) and C point (x2, β3), respectively.
And Further, the neighboring pixels on the grid points of B point and C point are B1 point (x1, y1), B2 point (X1, y2),
Let C1 point (x2, y2) and C2 point (x2, y3).

However, it should be noted that the Y coordinates of the points B and C differ depending on the coordinates of the point A. In addition, the length of the line segment created by these points and its relational expression are given in (1)
Equation (2) is shown below. In the expressions (1) and (2), x1, x2, y1, y2, y3 are integers,
There is a relationship of x1 <x2, y1 <y2 <y3.

[0013]

[Equation 1]

In the above description of FIG. 3, the image rotation angle θ is −1.
When 80 ° ≦ θ ≦ 180 °, −45 ° ≦ θ ≦ 45
This is the case of °, θ ≦ −135 °, θ ≧ 135 °. For image rotation angles other than this, two points B (α2, y1) and C (α3, y2) near point A where the main scanning line of the rotated image passing point A intersects the horizontal grid of the original image are obtained. , Neighboring pixels on each grid point B1 point (x2, y1), B2 point (x3, y1), C1 point (x3, y2), C2 point (x
4, y2) and line segment lengths ΔX1, ΔX2, ΔX3.
Similarly, ΔX4, ΔY1, and ΔY2 are obtained.

In FIG. 4, the image rotation angle θ is -135 ° <
It is a schematic diagram showing the coordinate of each point in the case of θ <−45 ° and 45 ° <θ <135 °. The line segment lengths formed by these points and their relational expressions are shown in the following equations (3) and (4), respectively. Here, x2, x3, x4, y1, y2 are integers, and have a relationship of x2 <x3 <x4, y1 <y2.

[0016]

[Equation 2]

It should be noted that points B and C are A1 and A2, respectively.
The reason that the main scanning line of the rotated image and the grid intersect each other instead of the point is to store the rotation information of the image. Assuming that the coordinates of a pixel P1 point on the rotated image is (x0, y0) and the image rotation angle is θ, the coordinates of the point A on the original image corresponding to the P1 point are expressed by the equation (5).

[0018]

[Equation 3]

Further, the image rotation angle is θ ≦ −135 °, −4
When 5 ≦ θ ≦ 45 ° and θ ≧ 135 °, the Y coordinates of the B point and the C point are obtained by the equation (6). However, at other image rotation angles, point B (α2, y1) and point C (α
3, y2) The respective X coordinates are represented by the equation (7).

[0020]

[Equation 4]

Under the above preparations, points A, B, C, B
When the density values of 1 point, B2 point, C1 point and C2 point are a, b, c, b1, b2, c1 and c2 respectively, B
Interpolate the density values at points C, A. The image rotation angle θ is θ ≦ −135 °, −45 ° ≦ θ ≦ 45 °, θ ≧ 135 °
Table 1 shows each interpolation formula in the case of. Table 2 shows interpolation formulas for image rotation angles other than these. Here, TN is a threshold value for detecting a steep edge point, a density difference parameter between two pixels such as 100 and 200, and TD is a threshold for storing steep edge information. Is a value, a line segment length parameter smaller than 0.5 such as 0.3 or 0.4, and γ is a constant for storing steep edge information, such as 0.7 or 0.9. It is a value smaller than 1.0.

[0022]

[Table 1]

[0023]

[Table 2]

Based on the above-mentioned principle, one embodiment of the present invention will be specifically described below.

FIG. 1 is a schematic block diagram of an embodiment of the present invention. In FIG. 1, the keyboard 3 is for inputting the rotation angle of the image, and the input rotation angle is stored in the rotation angle memory 4. On the other hand, the image memory 2 stores the data of the original image. The controller 11 controls the rotated image address generator 5, the original image dotted line segment length generator 6, the original image data read controller 7, the interpolation processor 8 and the rotated image data recording controller 10. That is, when the rotation angle is stored in the rotation angle memory 4, the controller 11 causes the rotation image address generator 5 to output the address (x ₀ ,
y ₀ ) is generated and given to the original image dotted line segment length generator 6. The original image dotted line segment length generator 6 determines each pixel address (x
₀ , y ₀ ) is input, the A shown in FIGS. 3 and 4 is obtained in accordance with the equations (1) to (7) described above according to the value of the image rotation angle θ. Point, B point, B1 point, B
The coordinates of 2 points, C point, C1 point, and C2 point are obtained, and the line segment lengths ΔX1 to ΔX4 and ΔY1 to ΔY4 are obtained. Then, the original image dotted line segment length generator 6 sends the obtained address information of each image to the original image data read controller 7 and gives the line segment length information to the interpolation processor 8. It should be noted that the interpolation processor 8 includes the above-mentioned parameter T from the threshold generator 9.
H, TD and constant γ are given. Interpolation processor 8
Calculates the density value of the coordinate of the point A corresponding to the pixel P1 point of the rotated image from the density value of each pixel, the address information, and various kinds of threshold information based on Table 1 or Table 2 described above. Then, the obtained interpolation density value is stored in the rotated image address (x ₀ , y ₀ ) of the image memory 2 via the rotated image data recording controller 10.

FIG. 5 is a flow chart for explaining the operation of the embodiment of the present invention, and FIG. 6 is a flow chart for explaining the density value interpolation operation shown in FIG.

Next, referring to FIGS. 1, 5 and 6,
A specific operation of the embodiment of the present invention will be described.
When the rotation angle is input from the keyboard 1, the rotation angle information is stored in the rotation angle memory 4. In the step SP1 shown in FIG. 5 (abbreviated as SP in the drawing) SP1, the controller 11 has rotation angles of 0 ≦ −135 °, −45 ° ≦ θ ≦ 45 °,
It is determined whether or not θ ≧ 135 °. When the controller 11 determines that the rotation angle is within the range, in step SP2, the rotation image address generator 5 sets the pixel address to y = 0, and then in step S2.
Set x = 0 at P3. The address information of y = 0 and x = 0 is given to the original image dotted line segment length generator 6. The original image dotted line segment length generator 6 detects the point A shown in FIG. 3 in step SP4. Then, in step SP5, it is determined whether or not the point A is located in the original image. When it is determined in step SP5 that point A is located in the original image, address signals of all pixels of the rotated image are sequentially generated and given to the original image dotted line segment length generator 6.

The original image dotted line segment length generator 6 executes step SP1.
At point 0, point B and point C are detected, and their respective coordinates are obtained. At step SP11, point B1, point B2, point C1
The coordinates of points C2 and C2 are detected, and the line segment lengths ΔX1 to ΔX4 and ΔY1 to ΔY4 are obtained. These pieces of information are given to the interpolation processor 8. In step SP12, the interpolation processor 8 outputs the points B1, B2, C1 and C2.
From each density value b1, b2, c1, c2 of the point B
The density values b and c at points C and C are interpolated, and at step SP13, the density values at point A are interpolated. The interpolated density value thus obtained is recorded in the image memory 2 via the rotation image data recording controller 10, and the CRT display 12
Is displayed in. Further, the controller 11 updates the x address at steps SP6 and SP7 until x <X _o , and then repeats the operations at steps SP8 and SP9 until the y address becomes y <Y _0. Processing is performed on all pixels of.

Next, the operation of interpolating the density value at point B will be described with reference to FIG. The threshold value generator 9 gives the interpolation processor 8 a density difference parameter TN, a line segment length parameter TD, and a constant γ as threshold values. To obtain the interpolated density value b at point B, the interpolation processor 8 finds the difference between the density value b1 at point B1 and the density value b2 at point B2 in step SP21, and the absolute value of the difference is the density difference parameter TN. Is greater than or equal to. | B1-b2
If | is smaller than or equal to TN, interpolation is performed according to the interpolation formula of b = b1 × ΔY2 + b2 × ΔY1, which is the formula (1) at the point B in Table 1.

If │b1-b2│> TN, it is determined in step SP23 whether the difference between b1 and b2 is larger than TN. If b1-b2> TN, B1 is determined in step SP24. It is determined whether the line segment length ΔY1 of the point is smaller than or equal to the line segment length parameter TD. If the line segment length is ΔY1 ≦ TD, step S
In P25, interpolation is performed according to the interpolation formula of b = b1 * (1- [gamma] + [Delta] Y2 * [gamma]) + b2 * [Delta] Y1 * [gamma], which is the expression (2) at point B in Table 1.

At step SP24, ΔY1> TD
If so, in step SP26, the interpolation is performed according to the interpolation equation of b = (b1 + b2) / 2, which is the equation (3) of the point B in Table 1.

At step SP23, b1-
If it is determined that b2> TN, step SP27
It is discriminated whether or not ΔY2 ≦ TD in step S1, and if so, it is the formula (4) at point B in Table 1, according to the interpolation formula of b = b1 × ΔY2 × γ + b2 × (1-γ + ΔY1 × γ) Interpolate.

At step SP27 described above, ΔY2
If it is determined that> TD, in step SP29, interpolation is performed according to the interpolation equation of b = (b1 + b2) / 2, which is the equation (5) of point B in Table 1. By such processing, the interpolated density value b at the point B can be obtained.

Interpolation density values c and a at points C and A
In order to obtain, the above-mentioned parameters B1, b2, Δ at point B
Instead of Y1 and ΔY2, parameters c1, c2 and Δ at point C
Y3, ΔY4 and A point parameters a1, a2, X
It suffices to input 1, X2 and perform the interpolation process according to the flowchart of FIG.

Further, as shown in FIG. 4, -135 ° <θ
In the case of <−45 °, 45 ° <θ <135 °, the formulas (1) to (5) in Table 2 are replaced with the formulas (1) to (5) in Table 1 above. The interpolation process may be performed using an expression.

In the above description, the interpolation process for monochrome / digital gradation image data has been described.
If the same processing is performed for each plate for digital gradation image data such as RGB colors and printing YMCK colors, rotation processing can be executed for various color / digital gradation image data.

[0037]

As described above, according to the present invention, the neighboring coordinates of two points at which the main scanning line of the rotated image passing through the corresponding coordinates on the original image intersects the horizontal or vertical grid of the original image at a rotation angle θ. Then, the density value of each intersection coordinate is interpolated by threshold processing from the distance between the coordinates of two points on the grid point near each intersection coordinate and each density value, and the threshold value of these two points is further calculated. Interpolation with the same threshold value processing from the interpolated density value and the distance between the original image-corresponding coordinates and the intersection coordinates, and the interpolated density value of the original image-corresponding coordinates is obtained to rotate the digital gradation image with high quality at any angle. Can be done.

[Brief description of drawings]

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

FIG. 2 is a schematic diagram showing an example of monochrome / digital gradation image data.

[Fig. 3] Image rotation of the original digital gradation image data for each θ
FIG. 6 is a schematic diagram when a pixel of a rotated image rotated by only a point and a main scanning line passing through the point are superimposed on an original image.

FIG. 4 is -135 ° <θ <-45 °, 45 ° <θ <13.
It is a schematic diagram when the pixel of the rotation image in the case of 5 degrees and the main scanning line which passes the point are superimposed on the original image.

FIG. 5 is a flowchart for explaining a specific operation of the embodiment of the present invention.

6 is a flowchart showing a specific operation of the interpolation processing shown in FIG.

[Explanation of symbols]

 2 image memory 3 keyboard 4 rotation angle memory 5 rotation image address generator 6 original image dotted line segment length generator 7 original image data read controller 8 interpolation processor 9 threshold generator 10 rotational image data recording controller 11 controller 12 CRT display

Claims (2)

[Claims] 1. An image rotation processing method for image rotation processing of digital gradation image data at an arbitrary angle, the method comprising: obtaining coordinates of an original image corresponding to pixel coordinates of the rotation image; Determining the coordinates of the two points near the pixel where the corresponding line segment passes through the coordinates on the original image and intersects the grid of the original image; and the density values of the two neighboring pixels on each of the two grid points. The step of interpolating the density value of each intersection according to each threshold value condition of the distance, and the original image coordinates based on the same threshold value condition from the calculated interpolated density value of the two points and the distance of the original image coordinates. An image rotation processing method comprising the step of interpolating the density value of 2. An image rotation processing device for performing image rotation processing on digital gradation image data at an arbitrary angle, comprising image rotation angle input means for inputting an image rotation angle, and image rotation from the image rotation angle input means. Original image coordinate generating means for generating a signal indicating the coordinates of the original image corresponding to the pixel coordinates of the rotated image in response to the input of the angle, and the line segment corresponding to the main scanning line of the rotated image is the original image coordinate generating means. Coordinate calculation means for obtaining the coordinates of two points near the pixel which pass through the coordinates on the original image based on the signal generated by and intersect with the grid of the original image, and two pixels near the pixel on each of the two points The first density interpolating means for interpolating the density values at the respective intersections according to the density value and the threshold value condition of the distance, and the two interpolated density values obtained by the first density interpolating means. The same threshold condition from a distance of the original coordinates, with a second concentration interpolation means for interpolating the density value of the raw image coordinates, image rotation processing apparatus.