JPS6042988B2 - Input figure rotational distortion correction processing method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention provides a method for correcting rotational distortion of an input figure, particularly in an image processing apparatus that processes an input figure stored in an image memory using an input means such as a facsimile. This invention relates to a processing method for correcting rotational distortion of input figures, which takes into account that stored figures are subject to rotational distortion and corrects the rotational distortion through relatively simple processing.

Although this is not necessarily limited to facsimile input means, when a given input figure is scanned and stored in an image memory, the figure in the image memory is generally free from rotational distortion. I can't.

To correct such rotational distortion, please refer to Figures 1 and 2.
As will be described later with reference to the figure, since there is a certain relationship between the original input figure and the figure undergoing rotational distortion,
It is sufficient to perform the correction for each mesh on the figure. If you perform this kind of correction for each mesh, the result will be "
In general, sin θ and cos
It is necessary to prepare the value of θ, and it is necessary to measure the angle θ correctly. The present invention aims to solve the above-mentioned problems, and the input figure rotational distortion correction processing method of the present invention is an image processing method that stores an input figure on an image memory and reads and processes the stored image memory. In the apparatus, a predetermined number of at least three
For each of the three points, the points on the rotationally distorted figure stored in the image memory corresponding to the three points are extracted as first coordinate position information, and each point on the input figure is divided into grids. The coordinate position information of at least one point of each of the lattices on the figure subjected to the rotational distortion corresponding to at least one point of the lattice is determined as second coordinate position information based on the first coordinate position information. The figure information of the mxn mesh belonging to one lattice on the input figure is extracted as the figure information of the Mxn mesh that does not have rotation based on the second coordinate position information. It is a feature.

This will be explained below with reference to the drawings. Figures 1A and B are explanatory diagrams for explaining rotational distortion correction according to the present invention that takes into account the rotation of figures, Figures 2A and B are explanatory diagrams for explaining calculation processing used in the present invention, and Figure 3 is an explanatory diagram for explaining the calculation process used in the present invention The configuration of one embodiment is shown.

When an input figure 2 having a figure 1 is stored in an image memory as shown in FIG. 1A, it is generally stored in a rotationally distorted state as shown in FIG. 1B.

That is, it is stored as a figure 3 that has undergone rotational distortion. Now, assuming that both figures 2 and 3 coincide at the illustration origin (0,0), the point (X, y) on the input figure 2 is converted to the point (x'' , y''), there is the following relationship.

Therefore, it is sufficient to perform the correction assuming that the point (X, y) on the input figure 2 is located at the point (x'', y'') on the image memory. However, performing the above correction for each mesh on the figure requires an extremely large amount of processing.
For this reason, in the case of the present invention, the following correction is limited to the correction that satisfies the allowable range. That is, the input figure 2 is divided into grids 4 made up of Mxn meshes, and the coordinate position on the image memory of the point (XO, yO) at the upper left corner of each grid, for example, is determined. That is, the correspondence between the illustrated (XO, yO) and (XO'', YO'') is determined based on equation (1). And the Mx
It is assumed that the n-mesh lattice 5 is an Mxn mesh lattice 52 shown by dotted lines in FIG. 1B. In other words, regarding figure 3 on the image memory,
The upper left corner point (XO, yO) of the lattice 5 of the input figure 2 has the coordinates (
If it is found that the coordinate point (XO'', YO'') on the figure 3 is ″)
The M x n mesh grid 5'', which is obtained by extracting m meshes in the horizontal direction and n meshes in the vertical direction, is read and processed. By adopting this method of dividing into each grid, It is easy to understand that the amount of correction processing is 1/m-n.However, by adopting such a simplified method, how much error is included in each mesh within the grid 5''? need to be evaluated.
When the size of the grid 5゛ is chosen as 10 x 10 mesh,
Assuming that an error of up to 2 meshes is allowed for the true mesh point on input figure 2, the calculated rotation angle θ is approximately 11
It is sufficient if it is 0 or less. Further, assuming that an error of one mesh at most is allowed, it is sufficient that the rotation angle is 5 degrees or less.
Conversely, if about 200 rotations actually occur, and if an error of up to 2 meshes is allowed, then the size of the grid 5'' should be selected to be 5 x 5 meshes. Even when performing error correction using the simplified method described with reference to FIGS. 1A and B above, it is complicated to measure the rotation angle 0, and measurement errors are introduced. easy.

For this purpose, the following process is taken. Figure 2 A and B show, for example, four reference points (090)) (XW
9O), (09YW)) (XW9Y7) are given, and this shows that any point (X, Y) becomes a point (X'', Y'') on figure 3 that has undergone rotational distortion. .

The relationship is expressed by the above equation (1).

Here, as shown, Xw, Y, v, X, YL, ΔX,
Considering ΔY,
)~(0?YW))(XW9Yw) is known.

Then, display the figure 3 on a display, for example, and input the coordinates of the point corresponding to the reference point using a light pen or the like to obtain (0, 0), (Xa, Ya), (Xb, Yb),
If extracted as (Xc, Yc), it is given as a known value.

FIG. 3 shows the configuration of an embodiment of the present invention.

In the figure, 6 is an image input device, 7 is an image memory, 8 is an image memory address control unit, 9 is a display, 10 is a light pen, 11 is a reference point table memory, 12 is an address correction circuit, and 13 is an address・Table memory,
14 represents a control section. The input figure 2 shown in FIG.
is stored in At this time, rotational distortion occurs as shown in figure 3 shown in FIG. 1B. When the graphic 3 having the distortion is read out from the image memory 7 using the same address information as the storage address and displayed on the display 9, the graphic 3 shown in FIG. 1B is displayed. At this time, using the displayed figure on the display 9, the points (0,0), (Xa, Ya), (Xb, Yb) shown in FIG.
), (Xc, Yc). By inputting, the coordinates of each point are stored on table memory 1.

Assuming that Mxn mesh grid 4 is specified as shown in FIG. and stored in the address table memory 13.Next, in obtaining the corrected figure (corrected video),
It is done as follows.

That is, one coordinate value (x'', 1y'') on the address table memory 13 is supplied to the address control unit 8,
The address control unit 8 controls the coordinate value (x″) on the image memory 7.
m pieces in the X-axis direction from the address position corresponding to
Control is performed to read n pieces of mesh information in the Y-axis direction.
Then, the information of the read Mxn mesh is treated as information of a grid having coordinate values (X, y) in the input figure 2 shown in FIG. 1 and processed. When the reading for one grid is completed, the address control unit 8 receives the next coordinate value from the address table memory 13 and performs the same reading as described above. These controls are performed by the control section 14. As described above, according to the present invention, the process for correcting rotational distortion is greatly simplified, and there is no need to directly measure the rotation angle θ, thereby simplifying the operator's operations.

[Brief explanation of drawings]

Figures 1A and B are explanatory diagrams for explaining rotational distortion correction according to the present invention that takes into account the rotation of figures, Figures 2A and B are explanatory diagrams for explaining the calculation process used in the present invention, and Figure 3 is an explanatory diagram for explaining the rotational distortion correction according to the present invention that takes into account the rotation of figures. 1 shows the configuration of an embodiment of the invention. In the figure, 2 is an input figure, 3 is a figure subjected to rotational distortion, 4 and 5 are lattices, 5'' is a pseudo lattice, 6 is an image input device,
7 is an image memory, 8 is an address control unit, 9 is a display, 10 is a light pen, 11 is a reference point table, 12 is an address correction circuit, and 13 is an address table.

Claims (1)

[Claims] 1. In an image processing device that stores an input figure in an image memory and reads out and processes the stored image memory, for at least three predetermined points on the input figure, information is stored in the image memory corresponding to the three points. A point on the stored figure subjected to rotational distortion is extracted as first coordinate position information, and the figure subjected to rotational distortion corresponds to at least one point of each grid obtained by dividing the input figure into lattices. The coordinate position information of at least one point on each of the above-mentioned grids is determined as second coordinate position information based on the first coordinate position information, and
The input graphic is characterized in that graphic information of an m×n mesh belonging to one grid is extracted as graphic information of an m×n mesh having no rotation based on the second coordinate position information. Rotational distortion correction processing method.