JPH03198167A - Method and device for deforming two-dimensional bit map image - Google Patents

Abstract

PURPOSE:To deform the two-dimensional bit map image so that a path is put thereto artificially by combining the processing for executing variable power of picture element density in accordance with a shape of a deformed image with respect to a two-dimen sional bit map original image, and the processing for deforming the shape of the original image to the shape of the deformed image designated arbitrarily. CONSTITUTION:Variable power of picture element density in one direction of a bit map original image being a rectangular square shape is executed so as to be propot- tional to the ratio of a pair of opposite side lengths (Q1Q2 and Q3Q4) of a bit map image of a quadrangle to be deformed, and also, variable power of picture element density in the remaining one direction of the original image is executed so as to be proportional to the ratio of the remaining pair of opposite side lengths (Q1Q4 and Q2Q3) of the image to be deformed, by which the processing for giving a depth sense is executed. Simultaneously, by a pseudo affine deformation processing, the deformation processing extending from the rectangular quadrangle of the original image to a deformed arbitrary square shape is executed. In such a way, the two-dimensional bit map image can be deformed so that a path is put thereto.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an image transformation method and apparatus for changing the shape of a two-dimensional bitmap image.

(Prior art) Various types of transformation functions can be considered for two-dimensional bitmap images, including enlargement/reduction and rotation, depending on their use and purpose, and many methods have been proposed to realize these functions. has been done. One of its functions is the ability to add perspective to two-dimensional plane images. This is necessary when transforming a flat image so that it takes on the shape of a three-dimensional image when viewed from various viewpoints, or when pasting or transforming a two-dimensional image so that it can be combined with a three-dimensional object. .

Quadratic projective transformation is known as a general method for realizing this function, and the image processing subroutine package rsPIDE developed by the National Institute of Electronics Technology, Agency of Industrial Science and Technology.
That software has also been created in RJ.

Although quadratic projective transformation does not use three-dimensional information in its calculation formula, it is calculated based on the arrangement of images in three-dimensional space, so it can express accurate perspective.

However, in addition to this accurate calculation method, there has been a demand for pseudo-perspective calculations using only two-dimensional information.

(Problem H to be Solved by the Invention) The present invention has been proposed in response to the above requirements, and includes an image transformation method and method for pseudo-parsing a two-dimensional bitmap image by calculation using only two-dimensional information; The object is to provide an apparatus for carrying out the method.

(Means for Solving the Problems) The present invention provides a method for scaling a set of coordinates of a rectangular bitmap image to be transformed and a pixel density in one direction of a rectangular bitmap original image corresponding thereto. In addition, by combining the remaining one set of coordinates of the image to be transformed, the process of scaling the pixel density in the remaining one direction of the original image to correspond to the remaining set of coordinates of the image to be transformed, and the pseudo affine transformation process, This is a two-dimensional bitmap image transformation method that transforms the image as if it were given a perspective.

be.

Furthermore, when the apparatus of the present invention for implementing this method performs the following transformations based on the coordinates of each of the four vertices of the original image and the transformed image of the rectangular bitmap to be transformed, a first processing means for calculating a formula representing the correspondence between the coordinates of each pixel in the transformed image and the corresponding coordinates on the original image; , and the pixel density of the bitmap image between the opposite sides is scaled to correspond to the other set of coordinates of the original image after transformation, and the pixel density of the bitmap image between the opposite sides is changed to correspond to that. changing the scale; applying a pseudo-affine transformation to the ratio of the lengths of the other pair of opposite sides; and calculating the correspondence obtained by the first processing means from each coordinate in the area defined by the four vertices of the transformed image. Based on the formula representing the original image, calculate the pixel value of the original image at the corresponding coordinate on the original image calculated by the first processing means and the second processing means from each corresponding coordinate on the original image, and third processing means for determining pixel values at the coordinates of the image area after deformation.

(Operation) Bitmap image of the rectangle to be transformed (Fig. 1 (
A pair of opposite sides of C)) (e.g. sides Q1Q1 and Q = Q a
) in one direction (e.g., X
direction), and the remaining 11 pairs of opposite sides (for example, sides QIQ and Q*) of the image to be transformed.
A process that gives a sense of depth by scaling the pixel density in the remaining direction (for example, the X direction) of the original image in proportion to the length of (
Processing for b)) is performed. At the same time, a pseudo-affine transformation process is performed to transform the rectangle of the original image into an arbitrary rectangular shape after transformation (processing from FIG. 1(b) to FIG. 1(C)). The method of the present invention, by combining these treatments,
It is possible to transform a two-dimensional bitmap image to make it appear more perspective.

Further, in the image deformation device of the present invention, the first processing means is capable of determining the correspondence between the coordinates of each pixel of the deformed image and the corresponding coordinates on the original image when the deformation process is performed using the above deformation method. (inverse mapping function formula of deformed coordinate function) is
Based on the coordinates of each of the four vertices of the original image and the transformed image,
demand.

The second processing means calculates the locus d on the original image corresponding to each coordinate within the area defined by the four vertices of the transformed image using the inverse mapping function equation of the transformed coordinate function.

The third processing means calculates pixel values of the original image at the corresponding coordinates on the original image calculated by the second processing means. For example, the pixel value is calculated from the pixel value of a pixel in the original image near the corresponding coordinates using interpolation or the like. The determined pixel value is stored in an image memory or the like as a pixel value at the coordinates of the transformed image area.

(Example) First, an example of the method of the present invention will be described with reference to the drawings.

Figure 2 shows an example of images before and after the transformation process, and Figure 2(a)
is the original rectangular image before transformation, and FIG. 2(b) is the final image after transformation with perspective. In Figure 2, 1
1, 12, 13.14 are the four corner vertices P+, P* of the original image
, P-, P4, and 19.20 and 21.22 are the four corner vertices Q+, Qs, Qa, and Q4 of the transformed image, which correspond to the vertices P, , P2+P3+P4 of the original image. Second
The method of performing the transformation as shown in the figure and its calculation will be explained.

First, a rectangular original image is processed to give a sense of depth.

Figure 1(a) shows the image before this processing, that is, the second image.
This is the original image in FIG. 1(a), and FIG. 1(b) is the image after giving a sense of depth. 11.12.13.14 in Figure 1
are the four corner vertices PP2 of the original image. PI1 P4 and 1
5.18, 17.18 are each vertex pHP2+ P of the original image
The four corner vertices of the image after depth transformation processing corresponding to 3+P4 are R++R21Rs+R4.

In the deformed image of FIG. 2, point Q 19 can be regarded as the closest point to the viewpoint in three-dimensional space, and point Q 21 can be regarded as the farthest point. If the distance from the viewpoint differs, the pixel density will differ between them.

Therefore, by performing a process that changes the pixel density between the side close to the viewpoint and the side far from the viewpoint, a pseudo depth effect is given. Although the pixel density is different between the point QI side and the point QI side, the same thing is true between the side Q1Q and the QsQ4 side, and between the Q,Q side and the QtQ- side. I can say it. Utilizing this feature, the image is scaled while continuously changing the scale ratio in each of the horizontal and vertical directions. Looking at the horizontal direction, the horizontal pixel sizes on sides Q, Q, and above remain the same size, and the sizes in the same direction on sides Q, Q, and above are side Q r
IQ words compared to Q x Q4 l / I Q IQs
It is processed so that it becomes I times. In the vertical direction, as in the horizontal direction, the size of the vertical pixel on the side Q s Q s is l Q * Qs I / compared to the size on the side Q 1Q 4.
Process so that I QI Q41 times. The calculation formulas for these two conversion processes are shown in formulas (1) and (2), respectively.

In the formula, xy is the horizontal and vertical coordinates before the depth transformation process, and X and Y are the coordinates after the process. P.I.

Each point p, . . . ~Q++Qt+ ~R3, R- has coordinate values in the horizontal and vertical directions, respectively. In actual calculations, the vertex at the lower left corner of the original image is arranged so as to be the origin position on the xy plane coordinates.

Length IR of two sides of a rectangular image after processing to give a sense of depth
, R, l, IR, Ri are given by equations (3) and (4).

Next, the rectangular image given a sense of depth is transformed into an arbitrary non-rectangular square. Figure 1 (C) is the finally obtained parsed image, where 19°20.21.22 is the four corner vertices Q++Qt*03+Q4 corresponding to each vertex R11RatR3,R, of the image before transformation. be. This transformation uses a pseudo-affine transformation in which pixels within the image are arranged at linear positions. The calculation formulas for the conversion process here are shown in formulas (5) and (6), respectively.

In the formula, x+y, X, and Y are the vertical coordinates of the horizontal plane before and after arbitrary rectangular transformation processing, respectively.

As described above, deformation is performed by combining two processes: conversion to give a sense of depth and conversion to an arbitrary rectangle. In actual processing, the former process is used to obtain an intermediate image, and then the latter process is performed on that image (the former and latter processes are performed successively on a pixel-by-pixel basis).

FIG. 3 shows an embodiment of a hardware configuration for executing the method of the embodiment of the present invention detailed above.

In the figure, 31 is a computer, 32 is a central processing unit, 33 is an image transformation processing section, 34 is an image memory, 35 is an input device, 36 is a display, 37 is a scanner, 38 is a printer, and 39 is a bus. The input unit @35 is for giving commands etc. to the computer. The display 36 outputs and displays images. The scanner 37 is for inputting images, and the printer 38 is for outputting images. The bus 39 plays the role of exchanging image data and control information. The image memory 34 stores image data, and is a central processing unit ffi! Reference numeral 32 controls the entire image processing apparatus and performs general calculations.

The image deformation processing section 33 performs processing for image deformation, which is a feature of the present invention, and as shown in FIG. 41, an inverse mapping calculation unit 42 that calculates an inverse mapping function of the deformation function formula from the input initial value, and an inverse mapping calculation unit 42 that calculates the coordinate values on the original image from the coordinate values of the image area after transformation traced using the inverse mapping function. An inverse mapping coordinate calculation unit 43 that calculates, an interpolation processing unit 44 that calculates new pixel values on the transformed image by interpolation from the calculated coordinate values, and a pixel input/output control unit that writes and reads pixel values. It consists of 45. 46 is an image memory in which image data before and after transformation is stored.

Now, the overall processing flow from reading out pixel values to writing them through coordinate calculations and the like will be explained. FIG. 5 shows the processing flow of the calculation procedure.

(2) The transformation initial value input unit 41 inputs the four corner vertices P1 of the image before and after transformation, which are the initial values. P1. P,,P.

and Q, Q8. Q3. Input the coordinate values of Q4.

(2) In the inverse mapping function calculation section 42, an inverse mapping function of the deformed coordinate function is calculated from the coordinate values input as initial values. The deformed coordinate function is a series combination of calculation formulas for two processes: transformation to give a sense of depth and transformation to an arbitrary rectangle, that is, the above-mentioned formulas (1) to (6).

(2) The inverse mapping coordinate calculation unit 43 first starts tracing the region of the deformed side image.

(2) Using the traced coordinate values of the transformed image area, calculate the coordinate values on the original image using an inverse mapping function.

(2) The interpolation processing section 44 first reads out a plurality of pixel values in the vicinity of the coordinate values obtained by the process of step (2) of the inverse mapping coordinate calculation section 43.

(2) The interpolation processing unit 44 then performs interpolation processing based on the read pixel values to calculate new pixel values.

An appropriate interpolation method may be selected depending on the required image quality and processing speed. For example, four-point linear interpolation using pixel values at four points around the calculated coordinate value is given as an example.

■The obtained new pixel value is sent via the pixel input/output control unit 45 to -
It is written to the coordinate position of the traced post-deformed image area in the image memory 46.

(2) Determine whether tracing within the image area after deformation has been completed. If Yes, the process ends; if No, the process returns to step (2).

(Effects of the Invention) As described in detail above, according to the present invention, processing that gives a sense of depth by scaling the pixel density of a two-dimensional bitmap original image according to the shape of the transformed image, and In combination with the process of transforming the shape of the image into the shape of an arbitrarily specified transformed image, an image with a pseudo-perspective added to the two-dimensional bitmap original image can be easily obtained by calculation using only two-dimensional information. be able to.

[Brief explanation of drawings]

FIG. 1 is a diagram showing a deformation process in an example of image deformation according to the present invention, in which (a) is an image before deformation, and (b) is an image before deformation.
) shows an example of an image after transformation processing to give a sense of depth, and (c) of the same figure shows an example of an image with added perspective after further performing pseudo-affine transformation processing. FIG. 2 is an example of an image before and after a transformation process that adds perspective to a plane image. FIG. 3 is a diagram showing an example of a hardware configuration for implementing the method of the present invention. FIG. 4 is a block diagram of an image transformation processing unit that performs calculations for image transformation. FIG. 5 is a diagram showing the processing flow of the overall calculation procedure in which pixel values are read, coordinate calculations, etc., and then written. 41... Deformation initial value input section, 42... Inverse mapping function calculation section, 43... Mass mapping coordinate calculation section, 44... Interpolation processing section, 45... Pixel input/output control section, 46 ...image memory.

Claims (2)

[Claims] (1) The pixel density in one direction of the original rectangular bitmap image is scaled in proportion to the length of a pair of opposite sides of the bitmap image, which is the rectangular one to be transformed, and The image is transformed by a combination of a process of scaling the pixel density in one direction of the remaining original image in proportion to the length of the remaining pair of opposite sides of the image to be transformed, and a pseudo-affine transformation process. A method for transforming a two-dimensional bitmap image, which is characterized by transforming it as if it were given perspective. (2) Based on the coordinates of each of the four vertices of the original image and the transformed image of the rectangular bitmap to be transformed,
1] [2] [3] A first processing means for calculating a formula representing the correspondence between the coordinates of each pixel of the transformed image and the corresponding coordinates on the original image when the transformations are performed; 1] The pixel density of the bitmap image between opposite sides is scaled in proportion to the length ratio of a pair of opposite sides of the original image of the rectangle to be transformed. [2] The original image of the rectangle to be transformed is [3] Perform pseudo-affine transformation on the scaled images of [1] and [2], which scales the pixel density of the bitmap image between opposite sides so as to be proportional to the length ratio of the other pair of opposite sides. a second process of calculating the corresponding coordinates on the original image from each coordinate in the area determined by the four vertices of the transformed image based on the formula representing the correspondence determined by the first processing means; and a third processing means that obtains a pixel value of the original image at the corresponding coordinates on the original image calculated by the second processing means, and sets it as a pixel value at the coordinates of the image area after deformation. A device for transforming two-dimensional bitmap images.