JPS60251474A - Picture data processor - Google Patents

Abstract

PURPOSE:To attain the rotational processing of a picture by adding a displaced variable held by a displaced variable holding means to an increased variable held by an increased variable holding means and successively finding out a grating point corresponding to a point processing at present from the added result. CONSTITUTION:If it is defined that a point string obtained by rotating a point string pn by a angle theta is qn, a part of the point string is qn-1-qn+2. At that time, the displaced variable holding part 1 holds displacements RX, RY from a grating point at the point qn. The increased variable holding part 2 holds increments DELTAX, DELTAY between two points qn and qn+1 in the X and Y axial directions. A reference grating length holding part 3 holds reference grating length l. A grating point position holding part 4 stores qn-1, qn, an+1, and qn+2 as rounded status like Kn-1, Kn, Kn-1, and Kn+2. A processing position storage part 7 stores that the current point pn is in processing when the moved position of the point pn is being found out by an operation means 5. An original information storage part 6 sores the original picture divided like a grating.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to an image data processing device that performs image rotation processing.

Conventional configurations and their problems In recent years, with the spread of microcomputers, the field of application of image information processing by computers is rapidly expanding.

Hereinafter, an image processing method will be described in which image information existing in a certain first area is rotated by an angle θ and transferred to a second area.0 As shown in FIG. ) by an angle θ to move to area 12 so that point A corresponds to point B. It is only necessary to move the lattice point p1 to the corresponding position in the area 12 by tilting it by the angle θ.
When the 0 image ゛, which results in the work of transferring the density of p6 to points q1 to q5, is divided into grid-like pixels and processed, pixels exist only at the grid points. Since the points 91 to q6 are not lattice points, they are generally approximated by a point r-T obtained by rounding the coordinate values of each point.

5 Now, the coordinates of points pn and qn are (xpn, Yp), respectively.

When (x9n, Yqn), the following relationship exists. However, xb and Yb are constants for parallel movement. Conventionally, in order to find the coordinates of each point q1 to q5, equation (1) was calculated individually, which required four multiplications for each point, that is, a total of 20 times.

In general, multiplication requires more complicated hardware and more calculation time than addition and subtraction, so it has the disadvantage that a huge amount of calculation time is required when the original image to be rotated has a large number of pixels.

OBJECTS OF THE INVENTION In view of the above drawbacks, the present invention provides an image data processing device that can process an image consisting of a large number of pixels divided into a lattice pattern at high speed with simple calculations when rotating an image. It is something.

Structure of the Invention The present invention provides a displacement amount holding means for holding the displacement amount of a position from a point processed immediately before to a grid point when performing image rotation processing, and an arbitrary initial value is given to the displacement amount holding means. an initial value setting means, an increment holding means for holding an increment between each point to be processed, a lattice point position holding means for storing a position of a lattice point corresponding to a point on which processing is performed, and the displacement amount. By adding the displacement amount held by the holding means and the incremental amount held by the incremental holding means, a lattice point corresponding to the new coordinate point of the point to be rotated is found, and the lattice point The above object is achieved by providing calculation means for notifying the position holding means.

DESCRIPTION OF EMBODIMENTS An embodiment of the present invention will be described below with reference to FIGS. 3 and 4.

FIG. 3 is a block diagram of an image data processing device according to an embodiment of the present invention.

Furthermore, Fig. 4 shows a part of the dot sequence qn, which is the result of rotating the dot sequence pn of the original image by an angle θ. Each point qn
This shows the positions from −1 to qn+2.

In FIG. 3, reference numeral 1 denotes a displacement holding section which holds the displacements Rx and RY from the grid point at point q8.

Reference numeral 1a denotes an initial value setting means for setting an arbitrary initial value to the displacement amount holding section 1. Reference numeral 2 denotes an increment holding unit that holds increments ΔX and JY in the X-axis direction and Y-axis direction between point q6 and point qs+1, respectively.

Reference numeral 3 denotes a reference grating length holding section that holds l, which is the reference length of the grating. Reference numeral 4 denotes a lattice point position holding unit, which temporarily stores which lattice point Ks will be when the point qs determined by the calculation of the calculation means 5 is rounded, as will be described later. For example, point qn-1 is Kn-1, qn is Kn, qn+1 is Kn+
1. qn+2 is stored in a rounded state as Kn+2. That is, the amount of displacement is rounded down and held as a value. 6
is an original image information storage unit that stores original image information such as density divided into a grid pattern.

Reference numeral 7 denotes a processing position storage unit which stores information that the current point pn is being processed when the calculating means 5 is determining the destination of the point pn.

The operation of the above configuration will be explained below with reference to the flowchart shown in FIG.

In the following explanation, after calculating the moving position of point qn, the moving position of point qn+1 is calculated. Then, an addition operation is performed, and the contents of the processing pixel memory 7 are increased by 1 from N corresponding to point pn+1 to N+ corresponding to point pn+1.
1, and the point to be processed by the calculation means 6 in the future is point pn+1.
Remember that.

(Next, the calculation means 6 adds the displacement Rx from the grid line Kxn in the X-axis direction held by the displacement amount holding unit 1 and the increment ΔX in the X-axis direction held by the increment holding unit 2. .

(]) The calculation means 5 uses lX held by the reference lattice length holding unit 3 and the calculation result (R
, lc+ΔX), and (Rx7■)≧ll
I judge that.

Note that if (Rx+ΔK)<lx, proceed directly to the processing block (→).

) The calculation means 6 subtracts lx from (R + ΔX).
Then, the value is sent out to the displacement amount holding means 1 again as an R process, and the displacement amount holding means 1 is stored and updated as a new displacement amount in the X-axis direction.

(e) The calculation means 5 performs an addition operation on the grid point position holding unit 4, and obtains the value obtained by adding lx to the grid line Kxn of the X axis stored in the grid point position holding unit 4, that is, Kx(n+1
) is stored and updated in the grid point position holding means 4 as a new grid line.

←) Next, the calculation means 5 adds the displacement Ry from the grid line Kyn in the y-axis direction held by the displacement amount holding unit 1 and the increment ΔY in the y-axis direction held by the increment holding unit 2. .

(8 calculation means 6 is held by reference grid length holding section 3.
A and the calculation result Ry+l performed in the processing block (c)
Comparing the magnitude with Y, it is determined that (Ra+ΔY)≧la.

Note that if (Ra+n)-4y, proceed directly to the processing block (field).

(→ The calculation means 5 subtracts lA from (Ry+ΔY).

Then, the value is sent again as Ry to the displacement amount holding means 1, and the displacement amount holding means 1 is updated as a new displacement amount in the Y-axis direction.

(1) The calculation means 6 performs an addition operation on the lattice point position holding unit 4, and calculates the value obtained by adding 4y to the grid line Kyn stored in the lattice point position holding unit 4, that is, a(n+1).
is used as a new grid line to update the memory.

(Takumi The contents of the newly set grid point position holding unit 4 (Kx
(n+1) tKy(n+1) is sent to the original image information holding section 6. The original image information holding section 6 is a processing pixel storage section 7
With reference to the contents of , the moving position of point pn+1 is (Kx(n
+1)tKy(n+1)), and transfers various information at point pn+1 to the position (Kx(n+1)1Ky(n+1)) as processing information.

As described above, according to this embodiment, by repeating the above operations, it is possible to sequentially determine the movement position only by addition/subtraction and size judgment, and each pixel in area 11 shown in FIG.
It can be rotated to 2.

That is, originally qn-1 to qn+2 are each pixel P, n-1 to pn+2 of the original image parallel to the X axis, as shown in FIG.
Since these points are the result of tilting the column of
The interval between is also constant. Therefore, as shown in FIG. 4, the difference ΔX, , fY between the coordinate values of each point is constant, and the following holds true.

Therefore, using the divine ceremony, the coordinates of pixel qn (xqn, Yq
If the coordinates of point qn+1 are (x9n+1.Y
The value of qn+1) can be determined from Y = y + ΔY qn+1qn 0. Therefore, based on the coordinates of the first point q1, points q2 and subsequent points can be sequentially determined using the formula.

In the embodiment, a reference grid length holding unit 3 is provided to determine the movement position using the constants of the reference grid lengths IIx and Ita, but if the reference grid lengths 1x and lA are both the length of one pixel, the processing block ), (The calculation means 6 may be set to subtract 1 for males, and the reference lattice length holding unit 3 may be omitted.0 With reference to FIG. 6 below, A more specific configuration of the image data processing device will be explained.

In FIG. 6, 601 is a storage device that stores images divided into grids. 602 is a bus that connects the storage device 601 and a processor which is a collection of components to be described later, and 603 is an internal bus of the processor. Reference numeral 604 denotes an instruction register, and the stored instructions are interpreted by a control unit 605 to control each of the pro riffs 606 to 616, which will be described later.

eoe is an arithmetic unit that executes necessary additions, subtractions, size comparisons, etc. according to the algorithm explained using FIG. 60
Reference numeral 7 denotes an original pixel pointer, and information (for example, density information) possessed by the pixel specified by this pointer is held in the original image data register 608. 609 is a pointer for grid lines in the X-axis direction (hereinafter referred to as xP), which holds the number 8 for grid lines perpendicular to the X-axis. 610 is a residual register (hereinafter referred to as RX) in the X-axis direction, which holds, for example, the displacement R of point qn shown in FIG. 3 from the grid point. 61
0a is an RX initial value setting means for giving an arbitrary initial value to the X-axis direction residual register 610. 611
is an increment register in the X-axis direction (hereinafter referred to as Dx), which holds the values of ΔX1, that is, It and cos θ in FIG. 612°613.613a and 614 are respectively grid line number pointers (yp) in the Y-axis direction;
The residual register (RY), RY initial value setting means, and increment register (DY) have the same functions as the registers 609 to 611, respectively. 615 is an interval register which holds the interval between grid lines, that is, the value l in FIG.

The operation of the above configuration will be explained below.

First, when an image rotation command is stored in the command register 604, each part of each block eo6 to 615 operates as follows.

(1) First, Rx initial value setting means 610a sets R
R by the RY initial value setting means 613a provided in the x610.
An arbitrary initial value is set for each Y613.

(io Next, increase the content of the original picture pointer 607 by 1, output it to the bus 602 as pixel address information, read the content of that address from the storage device 601,
The data is stored in the original image data register 608. (Corresponds to block (-1) in Figure 6) (11) Next, the contents of Rx610 and px611 are transferred to the arithmetic unit 606 via the internal bus 603, the two are added, and the result is stored in Rx610 again. .

(Corresponds to block (+:!l) in Figure 4) (iQ Rx6
The contents of 10 and the contents of the -interval register port 16 are input to the arithmetic unit 6.
06 and the contents of Rx610 are the interval register 6.
If it is smaller than the content value of 15, go to (V) described later, Rx
If the content value of 610 is larger, the process moves to (1) described later. (Corresponding to block 0 in FIG. 4) (y) Subtract the contents of the interval register 615 from the contents of Rx610 and store the result in Rx610.

Thereafter, the contents of XP 609 are incremented by 1 by arithmetic unit 606. (Corresponds to blocks (=) and (e) in Figure 4) (vi) Regarding the Y-axis direction (iiD, (iv) and (
0 (block (υ, (
-fi, corresponds to (su)) (vi◇Stores the contents of the original Cheetah register row 08 to the address in the storage device 601 corresponding to the position specified by XP609 and YP812. (Block of FIG. As described above, according to the above configuration, there is no need for multiplication, and it is possible to determine the density of the lattice points after rotation by only adding and subtracting decimal times and comparing the sizes. Image data processing can be performed at a higher speed than in the conventional method in which all equations are calculated for each point.

Next, the operation when two conditions described below are given to the configuration shown in FIG. 6 will be described.

(First condition) Rx610. Assume that the maximum value that RY613 can hold is equal to the grid spacing it x r ly. (Therefore, RX
Initial value setting means 610a, RY initial value setting means 613a
Therefore, the initial value to be set is also within this range (0) In other words, IJLId, lx, 1 = 1, Rx610°RY613 is configured to hold only values below the decimal point.

(Second condition) Addresses of lattice points are assigned as shown in FIG. In other words, the addresses of points on the same horizontal line increase as they go to the right, and the next address from the rightmost end of the horizontal line corresponds to the leftmost end of the horizontal line one line below, and the number of pixels in the width of the image space is W.
Suppose that there are

Above 11th "K, 1: Shif, f, Rx610.RY
When Rx613 becomes larger than the grid spacing lx, ly, a carry occurs during addition, and at the same time Rx610. R
The value obtained by subtracting the maximum value that Y613 can hold becomes the addition result. Therefore, the block (+:?l
The block (1v) shown in FIG. 7 is also a block (to).

i) The block (viD) is replaced by 0. This condition allows the interval register 616 shown in the figure to be omitted.

On the other hand, according to the second condition above, it takes
Since the two values 609 and YP612 can be filled with only the -dimensional address A, the block shown in Figure 5 ((e) means to move forward by 1 in the X-axis direction. is replaced by the block (1v) shown in Figure 7.Also, since the block (go) means to advance by 1 in the Y-axis direction, the address should be made smaller by exactly the width of the image space. If possible, it can be replaced with a block (viD).

Furthermore, according to the flowchart shown in Fig. 5, when two adjacent points on the original image are rotated and correspond to the same point by truncating the RX610, the value of the point on the right side of the original image is the final value that remains. However, as shown in block 011 of FIG. 7, by calculating and storing the density p of the pixel of the original image and the value of the point where it should be stored, a result that is more faithful to the original image can be obtained. It can also be used as a finale.

Although p has been described as representing the density of each pixel of the original image, it goes without saying that information other than density may also be used.

Next, initial value setting will be explained in more detail.

For example, RX610. The case where the initial value of RY613 is set to 0 will be explained with reference to FIG. 9.

FIG. 9 shows the positions of some points -091 to q5 when the result of rotating the point sequence p1 to pn of the original image by an angle θ1 is q1 to qn.

Now, as shown in Figure 9, if the initial value is 0, the starting point of the straight line Q obtained as a result of the rotation process is 0 and ql
and q2 are rounded to 12, q3 is rounded to 3, q4 is rounded to 4, and q5 is rounded to 5.

That is, RX610. By setting the initial value=0 to RY613 from the initial value setting means 610a and 613a, the result of the rotation process always shows the truncated lattice points.

Next is RX610. The case where the initial value of RY613 is 1/2 will be explained with reference to FIG. 10.

Figure 1o is a dot sequence p1 to pn of the original image, similar to the rotation in Figure 9.
When the results of rotating the angle θ1 are q1 to qn, the positions of the points 91 to q5 are shown.

Now, as shown in FIG. 10, if the initial value is 1/2, the starting point of the straight line Q obtained as a result of the rotation process is Qo. Then, ql is rounded to 1, q2 is rounded to 2, q3 is rounded to 3, Sq4 is rounded to 4, and q5 is stored in K6.

That is, RX610. For RY613, the initial value -l is set by the initial value setting means 6'10a and 613a, respectively.
By setting /2, the result of the rotation process in FIG. 9 is always rounding down, whereas in the case of FIG.

By changing the initial value settings as described above, different processing results can be obtained. A result like this ′ can be obtained by dividing an image into several rotations, for example, and then finally stitching together the results of the image processing to reproduce the image. By changing this, it is often possible to seamlessly connect images that have been processed separately. Of course, initial value setting is also effective for other processes.

Effects of the Invention As described above, the present invention calculates the amount of displacement from the lattice point at the immediately previous processing point held by the displacement amount holding means, and the incremental amount between each processing point held by the incremental amount holding means. Add and
By configuring the system to sequentially find grid points corresponding to the point currently being processed from the addition results, it is possible to rotate an image by processing an image consisting of a large number of pixels with simple calculations and at high speed. , Also, by the initial value setting means,
By being able to set an arbitrary initial value to the displacement amount holding means, more delicate processing can be performed, which has great value.

[Brief explanation of drawings]

1 and 2 are diagrams showing image rotation, FIG. 3 is a block diagram of an image data processing device according to an embodiment of the present invention, FIG. 4 is a diagram showing image processing of the device, and FIG. 6 is a block diagram of the image data processing device according to the second embodiment of the present invention, and FIG. 7 is a flowchart of the same device.
FIG. 8 is a flowchart of the image data processing device in the third embodiment, and FIG. 8 is a diagram showing address allocation in the device, and FIGS. 9 and 10 are diagrams showing the concept of rotation processing based on initial value setting. 1... Displacement amount holding section, 1a... Initial value setting means, 2... Increment holding section, 4... Grid point position holding section, 5... ...Calculation means. Name of agent: Patent attorney Toshio Nakao and 1 other person No. 1
Figure L-! -□”'F・I) No. 1 p+2 No. 6 Fig. 7 Fig. 8 w2 Se, -11111 Fig. 9 Reno 11

Claims (1)

[Claims] Displacement position holding means for holding the displacement of the position from the point processed immediately before to the grid point when performing image rotation processing; and initial value setting means for giving an arbitrary initial value to the displacement amount holding means; an increment holding means for holding an increment between each point where processing is performed; a grid point position holding means for storing the position of a grid point corresponding to the point where the processing is performed; and the displacement holding means holds. calculating means for determining a lattice point corresponding to a new coordinate point of the point to be rotated by adding the displacement amount and the increment amount held by the increment holding means, and notifying the lattice point position holding means; An image data processing device comprising: