JPH04205275A - Synthesis of image and formation of character font - Google Patents

Abstract

PURPOSE:To allow the exact adjustment of the size of a superposing image by previously setting three-dimensional position information on the object appearing in a background image and automatically calculating the magnification ratio and reduction ratio of the superposing image from the two-dimensional position in the background image. CONSTITUTION:The superposing image is automatically magnified or reduced to the size suitable for the superposing position thereof and is synthesized by superposition when the superposition place (superposing position) is assigned by using the three-dimensional information applied in the background image. The synthesized images superposed with the superposing image in the state of having a natural perspective feel on the background image is automatically obtd. in this way without making laborious operation. The labor for assigning the magnification ratio and reduction ratio at every time of changing the position of the superposed image, changing its layout, etc., is eliminated.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to an image synthesis method for superimposing images used in design simulation, etc., and as a special case, a method for synthesizing character fonts of arbitrary size and shape. Concerning how to generate character fonts.

[Conventional technology]

Conventionally, 1i-fold compositing processing of digital images has basically been performed in the following steps from +l) to (3). prepare. For example, the background imager is a U@ quantity, and the superimposed image is an image of a car.

(2) Separate a certain area in the superimposed image (This area is called a mask) 0 For example, only the automatic 1 part is separated. For this purpose, there are methods such as using one-color information, and a method in which the operator specifies the outline pixel by pixel.

(3) In order to overlap the cut out area in the superimposed image on the background image, perform the following operations (a) to (C1).

(al) Specify the overlapping position to specify the i-row, j-column component of the background image (i, j), the i-row, j-column component of the superimposed image to 8, j), Assume that the image is a color image and is represented by the three components of red, green, and blue (for example, the three components of A (i, j) are each represented by Ar (i.

j), , Ag (i, j), Ab (i, j)).

A common method for specifying the superimposed position is to specify where one pixel of the superimposed image is located in the background image. That is, B(11,3I) becomes A
As overlapping with (io, JO), l,,j+,jc
++ Specify jo.

(bl Specify the scaling ratio If you simply superimpose the superimposed image on the background image, the size will not match in most cases.Therefore, adjust the superimposed image by scaling it.Specifically, the superimposed image If it is multiplied by a and overlapped, the result is as follows.

C(i 0+i, jo+ j)= However, since x and y are not integer values, B' (x, y) is obtained by interpolating the values of pixels around the superimposed image.

A typical example is bilinear interpolation, which is expressed by the following formula.

B'(x, y) - (fx1+ 1-xl((y", + 1-ylB(nee, xl, [yj 'r ten(1-xj + 1-xHy
−(y', IB(CX), [y3 −? l)+(x−
[xi)(ry')+I-y)B([xl+l
,Iy”)+(x-jxl)(y-Cy',>(B
(jxl + I, rye, "H... (21 'x] is a Gaussian symbol, which harms integers not exceeding X.

tta Adjusting the tone of the cut out area The background image and the superimposed image are shadowed under different solar and lighting conditions, so if they were combined as they were, the tone would be different and it would look strange. Then, add the superimposed image or background image! 11 Conversion to adjust color tone 0 Normally, the three components of red, green, and blue are each expressed in 256 gradations from 0 to 255, with a total of 16.7 million colors. In the g*w4 conversion, each of the three components red, green, and blue is converted using a conversion function whose domain and value range are from 0 to 255. The conversion functions are represented by Fr, Fg, and Fb, and the pixel after conversion is represented by C'.

Cr'('IJ)- Fr(Cr' (i, j)) Inside the mask 'C
r'(i,j) Outside the mask...(The same applies to green and blue.

The above-mentioned position designation, scaling ratio designation, and two-tone adjustment are usually performed interactively.

Here, the method of expressing the mask will be explained.

Conventionally used mask expressions include bitmap expression and outline expression, and there is also a special bitmap expression called alpha mask. This is written by Toshihiro Anzai in his book titled ``Introduction to Digital Painting for Graphic Designers (3)''.
PIXEL 嚘84J pp, 181-184 (hereinafter “
Reference IJ).

Bitmap representation represents a mask as a binary image.

That is, typically, the pixels inside the mask are 1 and the pixels outside the mask are 0. Alpha mask is a multivalued or real number version of this. , typically from 0 to 1
When expressed as a real number α between 1. Let this be α(i,
j) Then, C(i,j)=(1-α)A(i,j)
+αR(f,j)...(4) A composite image is calculated. For example, if α is set to 0.5, both the background image and the superimposed image will appear to overlap. So 1. It is used when superimposing transparent objects, or when making the outline of the object visible to reduce the unnaturalness caused by superimposition. In outline representation,
The contour is represented by a series of dots, and the inside of the slope created by smoothly connecting the series of dots with, for example, a curved line is defined as a mask and used.

Furthermore, there are cases where it is desired to place the object to be superimposed behind something in the background image. This is the case, for example, when the background image is an image of a street corner with a streetlight, and it is desired to create a composite image in which a car cut out from the superimposed image is placed behind the streetlight.

In order to do this, it is assumed that masks for cars and streetlights have been created respectively. There are usually two methods. In the first method, the superimposition operation is performed twice, and first the car is superimposed on the background image. At this stage, the street lights are below the cars. Next, the streetlights are superimposed and synthesized at their original positions. At this stage the car is behind a street light. The second method is to define an order relationship in the depth direction for each mask area in the image and overlap them in one image. The parts of the background image other than the streetlights are placed at the bottom, the cars next, and the streetlights at the top, starting from the bottom.

Here, if the background image is made to be solid white, the superimposed image is made to be solid black, and the mask represents a character pattern, image synthesis becomes the same problem as printing a character font.

Using a bitmap mask will result in a bitmap font, and using an outline mask will result in an outline font. Outline fonts have the characteristic that they do not produce jagged edges (also known as jaggies) even when enlarged or reduced. Regarding this, edited by Maeyoumori, published by Maruzen, 1
- Workstation J, p, [43 to ρ, +98
, (s'A90) <hereinafter referred to as F'-Reference 2J).

(Problem to be solved by the invention) The prior art disclosed in Document 1 has the following problems.

The first problem is to adjust the size of the superimposed image to 1Eft so that the size corresponds to the distance # (perspective) in the depth direction of the superimposed image that touches the background image (accurate sense of distance).
This is difficult, and a picture can still be created even if the size of the superimposed images in the combined image is inaccurate. Further, each time the layout is changed, it takes a lot of effort to specify all the positions and scaling ratios of one gradation.

The second problem is that when images are superimposed and combined, they are scaled to adjust the size, but when the bitmap mask is scaled up, jaggies appear in the outline, and when it is scaled down, it becomes thin like hair. The thickness of the buttocks and pads may become unnatural,
It may disappear.

The third problem is that it takes a lot of effort to widen and narrow the mask. In addition, the mask may be widened or narrowed by swelling or shrinking processing in which the mask is expanded or shaved outward two pixels at a time.

For example, it is better to vary the amount of contour correction depending on the location of the superimposed image, such as modifying the contour of a person's skin to be more inward than it actually is, and modifying the contour of clothing to be more outward than it actually is. You may be able to get it. However, the conventional technique has a problem in that the contours become unnatural because all the contours are changed by the same amount.

The fourth problem is that when an ordering relationship in the depth direction is defined for the mask area, this does not correspond to the actual distance, so the scaling ratio and ordering relationship do not correspond, and it becomes a bit choppy. There was a problem.

The fifth problem is that when there are subtle shadings such as shadows in the superimposed image, semitransparent compositing through a window cannot be performed at full 7
It is said that unnatural i (+@ world will come to life in Amasu)
: etc. Also, in the conventional technology related to the t-character tsuonto ↓ in Reference 2, there was a problem that when the outline font was expanded to 4 large fonts, the cotton would look thin. #L Bold designation, which simply makes the character font thicker, cuts out all the strokes uniformly, making it impossible to display the delicate 7-onoto shape, so this problem has not been resolved.

Therefore, an object of the present invention is to provide an image synthesis method that solves the problems No. 1 to No. When changing the layout, such as changing the position of the superimposed image, the position, scaling ratio, and vkJ must be adjusted one by one. To save you the trouble of specifying. i2, in the case of superimposition compositing, when adjusting the superimposed image to an appropriate size by scaling it up or down, the bit magnification force itself is equivalent to the required scaling rate (scaling rate necessary to obtain the appropriate size). By keeping the bitmap mask small, jaggies can be prevented from occurring in the outline when the bitmap mask is enlarged, and the thickness of thin parts can be prevented from becoming unnatural or disappearing when the bitmap mask is being reduced. Thirdly, the mask should be easily swollen and contracted to eliminate unnatural contours.Fourthly, the order relationship between masks and the scaling ratio should be Make sure that they correspond to each other so that there are no contradictions between them.Fifth, if there are subtle shadings such as shadows in the superimposed image, the shading should be able to be expressed without being unnatural in the composite image.

Another object of the present invention is to prevent the line thickness from becoming unnaturally thin or disappearing compared to the character size when the character font size is enlarged or reduced. To provide a method for generating a character font by which a character font having a thickness suitable for the characters can be easily obtained.

[Means to solve the problem]

In order to achieve the above-mentioned first object, the image synthesis method of the present invention uses objects (topography) appearing in the background image.

The three-dimensional position information regarding the object (object, etc.) and the position information of the viewpoint are set in advance, and the size of the superimposed image is also set, and the (two-dimensional) superimposed position in the background image is specified here. Then, the two-dimensional position in this background image and the above settings are determined (so that a superimposed image of a size that matches the actual depth direction, that is, the position of the background object at this superimposed position in the depth direction) is obtained. The enlargement/reduction ratio is automatically calculated by using the relationship between the enlargement/reduction ratio of the superimposed image and the size of the superimposed image. Conversely, if the scaling ratio is specified, (of the superimposed image whose size is determined by it)
It is configured to limit the positions that can be superimposed.

Furthermore, by specifying the atmospheric light absorption spectrum (absorption rate characteristics of each color of red, green, and blue) (and as mentioned above,
By specifying the position and scaling ratio and, as a result, specifying the depth of superimposition in the background image, it is configured to calculate the color change due to atmospheric absorption (at that depth position) and adjust the gradation. do.

In order to achieve the second object, the image synthesis method of the present invention uses a plurality of bitmap masks or outline masks of different sizes and selectively uses them. Here, each bitmap mask represents a cutout target of a superimposed image that has been enlarged or reduced at a different magnification.

For example, when using multiple bitmap masks,
A bitmap mask created from a superimposed image with a magnification closest to the enlargement/reduction ratio specified at the time of synthesis is enlarged/reduced to a specified value, and synthesis processing is performed.

When using an outline mask and a bitmap mask, the bitmap mask is created for the reduced superimposed image. During compositing processing, a bitmap mask is used if the specified reduction rate is large, and an outline mask is used otherwise. Here, "the reduction ratio is large" means that the image as a result of scaling becomes small. "Thick" refers to a case that includes both (a case where there is a large change in the size after enlargement or reduction compared to the original size).

In order to achieve the third objective, the image synthesis method of the present invention is such that a cross section (cross section cut in the thickness direction) cut along a plane passing through an arbitrary position and parallel to a certain direction always has topological geometry. The present invention is characterized in that a tube-shaped solid body having the same surface area is prepared, and a mask that specifies a part of the superimposed image is defined by a cutting plane that passes through a predetermined position and is parallel to the plane.

In the present invention, this mask is referred to as a surface mask. The area of the cross section (thickness of the tubular solid) increases as the cutting position of this tubular solid changes in order.
It is possible to configure the mask so that it monotonically increases or decreases, and to change and adjust the size of the mask by moving the plane vertically in parallel (by changing the cutting position).

, In order to achieve the above fourth objective, the image synthesis method of the present invention divides the f-view image and the superimposed image into small regions, and provides not only information indicating a mere order relationship between masks for these small regions. , is characterized by having information (depth information) corresponding to the depth of each mask so that the order relationship and scaling ratio do not conflict.

In order to achieve the fifth object, the image compositing method of the present invention calculates the image density after superimposition by using the product of the image density of the weighted image and the image density of the background image at the same location (corresponding location). It is characterized in that it is a value divided by a constant that does not depend on the image density (for example, the image density of a relatively bright area in the superimposed screen; the reference density).

In order to achieve the above object, the character font generation method of the present invention prepares a tubular solid whose cut surface size and thickness of the tubular solid change depending on the cutting position, and A character font is defined by the size (thickness) and shape of a cutting plane generated by cutting the font at a certain position in the thickness direction. In the present invention, this character font is called a surface font.

This tubular solid can be defined by the positions of three-dimensional points and a curved surface that interpolates between the stiff points. For example, two bases, an upper base and a lower base, and a curved surface that interpolates between the two bases. There is a lack of definition. The cutting plane of the tube-shaped solid is such that when its cutting (positioning) changes, the character size and the thickness of the strokes ('Ja of each stroke) change.

In other words, depending on the cutting position, the character size is 1. The system is configured to generate a character font whose stroke thickness is determined at the same time as appropriate for the character size.

As a typical structure of a tubular solid, the contour between the upper base surface and the lower base surface can be configured in the shape of a cone with an outline font of different points.

[For production]

Based on the above configuration (the operation will be explained).

According to the means for achieving the first object, three-dimensional position information 11 (three-dimensional structure) and viewpoint information are set in the image, and the depth in the direction perpendicular to the screen is set at the position of each pixel in the image. can be calculated.

That is, the X-axis is set in the horizontal direction on the screen, the y-axis is set in the vertical direction, and the Z-axis is set in the depth direction (depth direction) perpendicular to the screen. The unit of length is defined as the pitch of pixels of an image on the xy plane. The viewpoint is Vp (x
o, )'o, Zo), image display surface (
P (x, ).

yr, o), and the expression representing the surface S in the image is y−f
Let it be (x, z). Here, the viewpoint Vp is the X-y plane (
It is assumed that the surface S is located in front of the plane P) and the surface S is located behind the x-y plane. The point where the straight line l passing through p and Vp intersects with S is Q (Xz, 3'z.

z2), then Y+ +t (Y+ 3'o)=f (x+
+t (x+ -xo)+-tzo) -
(6) This can generally be solved numerically, but in particular, if the surface S is a plane, r (x, z,) = ax + b
, when expressed as knee C, [2 is written as follows.

L= laX, fC-V, )/ Dr -y e a (Xt
Xo) I-b lo)
-+71Fili The depth from the plane of fear (x-y plane) to the surface in the image is defined by PQ, so n, is d
(x.

yr): When written as x, it is as follows.

d (x, y)= The distance from the viewpoint is Vp S, which can be expressed as e(x, y
), it becomes the following yo・).

e (x, y) = Through these calculations, assuming that the object to be superimposed and synthesized is on the surface of the image, the depth d is calculated relative to the location on the image.
is decided. Also, by solving this problem in reverse, if the depth d is determined, the positions 11P (x+, Y+, O) on the image can be limited. Specifically, d, xo, yo, f
Assuming that (x, y) is given, use equation (6) and equation (8) to calculate X, and y.

Obtain the equation of In this way, the surface S is the viewpoint Vp
The image will be projected onto the x-y plane (display surface) as an image viewed from. When the surface S is a plane, the equation becomes as follows.

” ()'+ -'10 a(Xt Xo)
+ b 20 ) ”= (a xl +c Y+
)2rcXI X. )”” (yr )'o
)”+Zo”) ...0ωAlso, even if objects are the same length, the farther away they are, the smaller they appear; therefore, an object with a length of p in the x-y plane F will look smaller in depth. At location d, it appears shrunken as shown below.

p'-pe(x,y)y'(+1(x,y)+e
(x,y)l-.QB, that is, once the location of the image to be superimposed and synthesized is specified, the scaling ratio of how much it should be scaled is determined. Conversely, if the size of the object to be superimposed is specified, t! The places where it can be synthesized are limited.

Furthermore, if the atmospheric color absorption rate is specified, 1
．． How much the color changes is determined by depth. That is, if we simply let γr, γg, and γb be the transmission ratios of red, green, and blue components of light when they travel by unit length, then at first Ar
(X, V), Ag (x, V).

Each color component that was Ab (x, y) becomes the next Ar' (x, y), Ag' (x, y) after passing through the depth d.
, Ab'(x, y).

Ar'(x, y)=rr' Ar(x, y)
)Ag'(x,y)=rg'Ag(x,y)
f-CJ2Ab'(x, y)=rb' Ab
(x, y) 'Reverse 22, color Ar' after this absorption,
Depth d is determined by specifying Ag' and Ab'. Also,
If the relationship between the superimposition position and the depth is also used, the superposition position and the color after absorption can also be related. Furthermore, γr5, γg, and γb can be specified to be more typical of the weather, such as a rainy day and a sunny day. If you memorize this in advance,
You can change the color tone in the same way by specifying only the weather without specifying the color absorption rate value.

According to the means for achieving the second object, when using a plurality of bitmap masks of different sizes,
Since the magnification or reduction ratio for enlarging or contracting the mask becomes smaller, the degree of jaggies becomes smaller. Furthermore, when using an outline mask and a bitmap mask, the outline mask generally gives good results except when reducing a thin outline pattern, and the bitmap mask is used when the reduction ratio is large. By using the two properly, there will be less jaggies.

According to the surface mask, which is a means to achieve the third objective, the size of each cross section of the tube can be arbitrarily defined, so for example, a very clear outline can be bulged or shaved. The mask is deformed by inflating or shaving areas where the outline is unclear.
This can be done by simply specifying that the plane on which the tube will be cut be lowered.

According to the means for achieving the fourth object, since the order relationship between masks and the scaling ratio are calculated using the same depth information, there is no 1- which contradicts them.

According to the means for achieving the fifth objective of item 1, the shadow is created by faithfully imitating the fact that a shadow is generated when there is an obstruction between a light source and an object, reducing the amount of incident light. Since an amount proportional to the image temperature of a portion is discarded to the background image, a natural composite image can be obtained.

Furthermore, according to the surface font, which is a means for solving the problems associated with the zero character font, the thickness of each stroke of an outline font realized by each cross section of the tube is changed. As a result, the thickness of the strokes can be changed just by changing the plane on which the tube is cut, so even large point fonts can be made into appropriate fonts. Based on this font, it is also possible to set bold, italic, etc.

〔Example〕

Embodiments of the present invention will be described below with reference to the drawings.

Since the hardware configuration used in each embodiment of the present invention is common, the configuration will first be explained with reference to the hardware configuration diagram in FIG.

In FIG. 6, the hardware configuration is computer 6.
01, memory 603 for storing digitized images;
A CRT 602 for displaying color images or characters, an image input/output device 605 for inputting digital images such as photographs, natural paintings, two-person drawings, etc., and a keyboard 604 for printing images. It consists of a mouse 606 with buttons 607 attached.

Here, to explain the function of the mouse 606, the control program on the computer 601 controls the mouse 606.
A small arrow is displayed on the CRT 602 according to the plane position indicated by 6, and by moving the mouse 606,
The operator can freely move this arrow on the screen of the CRT 602. In addition, a button 607 on the mouse 606
When the button is pressed, the position information at the end of the arrow on the screen on the CRT 602 is sent to the computer 601, and can be used by software running on the computer 601. Below, this operation is
We will use the term "pick". Also, C
Regarding the quality of the image displayed on the RT602,
000x1000, or about 1 million pixels per screen,
Since a total of about 16.7 million colors with 256 gradations for each of red, green, and blue is common, the present invention also uses image quality of this level. However, it is not limited to this.

The image input/output device 605 is also compatible with this.

Next, in the second embodiment of the present invention, an image synthesis method according to
Figure 1, '! Figures 2, 3, 7, and 8.

' This will be explained with reference to Figures 9 and 10.

FIG. 7 (fal and mountain) respectively indicate a background image and a superimposed image applied to the first embodiment.

The image synthesis method of the embodiment of $1 is to combine the car 703 in the superimposed image of Fig. 7(h) with the road 702 extending from the foreground to the upper left in the cyanide image of Fig. 7(al). When a location is specified, a composite image is created and displayed by automatically calculating the scaling factor and color absorption due to atmospheric absorption.The processing steps in the first embodiment are as follows.
It is shown in Fig. 1, Fig. 2, and Fig. 3.

1 to 3 are process flow diagrams common to each embodiment.6 In FIG. 1, in step 101, back 1! ! !
Input and display the i-image. In this example, the background image is a FJA view with a road in the $7 diagram 181. In step 102, the superimposed image is input and displayed. In this example, the superimposed image is the seventh
This is an image of the car 703 in Figure (BL).In both the background image 1 and the superimposed image, the mask of the one-boat fishing processing operator 111 (2) described in the prior art has already been cut out, and that information is also input at the same time. In this example, it is assumed that the tree 701 is distinguished from other parts in the background image and a mask is created, and a mask of the car 703 exists in the superimposed image.In step 103, the background image is Enter the three-dimensional information of the surface of the terrain and objects that exist in the background image, the viewpoint information, the unit length, and the color absorption rate by the atmosphere.Here, regarding the three-dimensional structure of the surface in the background image, the shooting conditions, the topography, If the information is known, this is used, but if it is not known, it is estimated and used.

In this example, using the coordinate system described in the operation, the roadway 70
2 is considered to be on a nearly horizontal plane, so the surface equation is y=0. Calculate by estimating the viewpoint as (500, 500, -500). In step 104, the image superimposition position, enlargement/reduction ratio, etc. are determined. This will be explained in detail in FIG. Step 105 calculates a composite image. At this time, the color tone is adjusted by atmospheric absorption. This will also be explained in detail with reference to FIG. In step 106, the composite image is displayed and, if necessary, outputted to a file. Next, details of step 104 will be explained with reference to FIG. Step 201
Now, specify 1 in the background image to superimpose the car.

In this example, we draw a cursor and designate it as the center of the car. In step 202, as described in [Operation], the depth at which the car is placed is calculated using the position of the plane, the viewpoint position, and the center position of the car. In step 203, as mentioned in the operation, the formula o
Calculate the scaling factor using n. In step 204, the outline of the vehicle to be superimposed, which has been determined in the processing up to now, is superimposed on the background image as a vA vehicle. Figure 8 (al is 0 and 2 representing this situation, Figure 8 ♂ "a" is the superimposed value'
Figure 8 1b) shows the composite image. Contour 801 and contour 802 represent two cases in different locations (near and far) in the same picture. If it is specified nearby, the composite image will look like the car 803 in FIG. 8(b). In step 205,
This branch is related to whether or not to change the display position.

If it is to be changed, return to step 201. If not changed, the center of gravity of the car determined by the above processing.

According to the scaling ratio, where is the car heavy? Compute a mask to be used at the time of synthesis that indicates whether Next, details of step 105 will be explained with reference to FIG. Step 30
1, the color absorption rate is calculated from the depth of the car using equation (2). It would be convenient to use a table for this purpose. From step 302, a loop is executed for each pixel of the composite image.

In step 302, it is determined whether the designated pixel is inside the mask. If the designated pixel is inside, the process advances to step 303, and the pixel value of the corresponding point in the superimposed image is set as the pixel value of the composite image.

Equations (1) and (2) are used for this. Next step 3
Color conversion is performed according to the table defined in 01 (step 304). Equation (3) is used for this. If the determination in step 302 is No, the image density of the corresponding point (same point) in the background image is set as the image density of the composite image (
Step 305).

In this embodiment, additional processing and changes described below can be performed. Regarding masks, the problems described in the conventional technology may occur, so we will develop a means for achieving the second purpose and the third purpose described in the means for solving the problem. A high-quality composite image with less jaggies can be obtained.

That is, when applying a method that uses multiple bitmap masks as in the first method, the superimposed image is enlarged or reduced at different magnifications in advance, and masks are created for each. Step 2037: Based on the calculated scaling ratio, a mask that is close to the eel is selected, and a mask at the time of synthesis of the step 206 in FIG. 2 is calculated. When using a bitmap mask and an outline mask as in the second method, a bitmap mask is created for a small image, and an outline mask is also prepared.

In this embodiment, a bitmap mask is used up to about twice the scale created by the bitmap mask, and an outline mask is used beyond that. A means for achieving the third objective uses a surf mask. FIG. 9 shows an example of a surface mask, which is a type of outline mask.A surface mask 901 expresses, for example, a shape with a protrusion on the right side of a rectangle (this is used for convenience). However, if the outer shape of a car is used, it can be used as is in this example).

The mask 90 is smaller when it is closer to the top like the cutting plane 902.
3, and when it is close to the bottom like the cutting plane 904, it becomes a large mask 905.

Further, although this embodiment does not use any background structure, it can be used. That is, FIG. 7 (al
In the background image, the tree 701 and the others are distinguished, and a tree mask exists. Since three-dimensional information is set on the ground surface including the road 702, the depth can also be set for this tree 701. During superimposition synthesis, not only can the car 703 be scaled up or down in accordance with the specified depth, but also the car can be automatically placed behind the tree according to its depth. Regarding this, FIG.
'b), on the other hand, is the case where the car 703 is placed in front of the tree 701.

In addition, this example does not handle the case where the object to be superimposed has a shadow.There is a shadow drawn from the automatic room to the ground, and when the shadow is also composited, it is difficult to express subtle midtones. As mentioned in [Operation], the image density of the composite image after superimposition is calculated by dividing the product of the image density of the superimposed image at the same location and the blue firefly image by a constant that does not depend on the location. do.

: Figure 11 (al to (dl) is an explanatory diagram of an example in which the superimposition target has a shadow;
The figure (b) shows the foreground image, the figure 4dl shows the composite image,
FIG. 11(C) shows the image density of the superimposed image.
)'s triangular object lll and its shadow 112 in the same figure)'s back! When superimposing a 1m image, the back 1m is proportional to the shade of the shadow.
By darkening one image, a composite image with natural shadows can be obtained. Therefore, in the superimposed image, the image srX of the line segment M−N changes as shown in FIG. 11 (c+), and the image density at the position of the shadow on the line is B (i.

j), and the image density of the same point in the background image is A(i.

j), and assuming that the reference brightness (brightness at a relatively bright location) in the superimposed image is B, the image density C(i, j) at the same position in the composite image is given by the following equation.

Next, the processing of the image synthesis method according to the second embodiment of the present invention will be explained using the processing flow diagrams of FIGS. 1, 3, and 4. In this embodiment, when the size of the car in the superimposed image is specified, the positions where the images can be superimposed are automatically limited, and the superimposed position is selected from among these to perform the superimposition synthesis process. In the process flow diagram of FIG. 1, steps other than step 104 are the same as in the first embodiment. Details of step 104 will be explained with reference to FIG. In step 401, the enlargement/reduction ratio of the superimposed image is specified. Step 401
In the formula +61. (81 or use your head to calculate the superimposition possible position and display it on the background image.

In step 403, superimposition possible positions are designated from among them. In step 404, the depth is calculated in the same way as step 202 in FIG. Steps 405 to 407 are similar to steps 204 to 206 in FIG.

Next, an image synthesis method according to a third embodiment of the present invention will be explained using process flow diagrams shown in FIGS. 1, 3, and 5. In this embodiment, when the degree of color absorption by the atmosphere of a superimposed image is specified, positions where superimposition is possible are automatically limited, and a superimposition location is selected from among these to perform superimposition synthesis processing.

Details of step 104 will be explained using FIG. 5. In step 501, a color change due to color absorption is specified. Since equation (2) is a simple calculation once the depth is determined, the image density of the superimposed image multiplied by a constant is displayed and an appropriate one is selected.

In step 502, depth is calculated. Step 501
Since it is the constant T determined by d raised to the d power, the depth can be determined by dividing the logarithm of the constant by the logarithm of T. Since the depth has now been determined, the second embodiment and subsequent interpolation processing will be performed. Steps 502 to 508
is the same as step 402 to step 407 in Figure 4]
It is.

Next, a method for generating a character font, which is another embodiment of the present invention, will be explained. Explained by conventional technology.

As in, make the superimposed image one color? -7 If the shape of the mask is a character, the image synthesis method becomes a method for generating a character font. Assuming that the cross section of the surface mask 901 in FIG. 9 represents characters, it is possible to move the plane up and down.
! : generates character fonts with different stroke thicknesses. For characters with large points, choose thicker strokes (the size of each line of the character), and for characters with smaller points, choose thinner strokes.

In the second means for achieving the objective, a bitmap mask is created in advance for an image with a large reduction ratio (the resulting mask size is small), an outline mask is created in advance for an image with a normal magnification, When synthesizing images, a bitmap mask is used when the reduction ratio is large, and an outline mask is used in other cases. Therefore, although there is essentially no difference in the ease with which lines disappear between both masks, when compositing images, the normal magnification ratio It is easier to erase the image when using a reduced outline mask created in advance for the image. Therefore, by creating outline masks in advance for multiple different magnifications, and using the outline mask created for relatively small images (images with large reduction ratios), thin lines can be prevented from disappearing. be able to.

〔Effect of the invention〕

As explained in detail above, according to the present invention, when specifying a superimposition location (superimposition position) by using three-dimensional information added to a background image, the superimposition image is adjusted to a size suitable for the superimposition position. Since the superimposed image is automatically enlarged or reduced, a composite image is automatically obtained in which the superimposed image is superimposed on the background image with a natural sense of perspective, without any troublesome operations. be able to. Moreover, at the same time, a composite image can be obtained in which the relationship between the superimposition position and the color tone is consistent with each other. 2. Using ni number of bitma-tub masks. Select a mask with a bit pine with a scaling factor that is close to the specified scaling factor. 2 or bitma
By properly using the edge mask and outline mask, it is possible to prevent jaggies from appearing in the outline of the superimposed image and from thin lines disappearing. By using a surface mask, swelling and shrinking of the outline can be easily performed. There is no contradiction between the mask order (front-back relationship) and the scaling ratio, and the boundaries of the composite image are clean, making it possible to generate a composite image that expresses subtle warmth like a shadow. .

Additionally, it is possible to generate an outline font that allows the stroke thickness to change naturally when the character size changes.

[Brief explanation of the drawing]

Fig. 1 is a flowchart of the processing of the first to '#43 embodiments of the present invention, Figs. 2 to 5 are flowcharts showing details of the processing of the embodiment of Fig. 1, and Fig. 6 is a flowchart of the processing of the embodiment of the present invention. 7 is a configuration diagram of a background image and a superimposed image for explaining an embodiment of the present invention, and FIG. 8 is a diagram of a weight position according to an embodiment of the present invention. An explanatory diagram of the specification method, Figure 9 is an explanatory diagram of the outline mask, and Figure 10 is an explanatory diagram of the outline mask.
FIG. 11 is an explanatory diagram of an image synthesis method according to an embodiment of the present invention, and FIG. 11 is an explanatory diagram of an image composition method for images with shadows according to an embodiment of the present invention. 601...Computer, 602...CRT, 60
3...Memory, 604...Keyboard, 605...
・Image input/output bag f, 606...Mouse, 607...
Mouse button, 701...tree, 702...roadway. 1, 2, 3, 4, 5, 6, 7 (b) 8, 9, 10, 11 (o) (C)

Claims (1)

[Claims] 1. A composite image is created by superimposing a part of another digitally represented image or a part of a background image on a digitally represented background image as a superimposed image. In a computer system, position information and viewpoint position information representing the three-dimensional position of the surface of the object represented in the background image are set in advance, and the superimposed image is compared to the two-dimensional position in the background image on which the superimposed image is superimposed. By using this relationship with the scaling ratio of the superimposed image, when the superimposition position in the background image is specified, the scaling ratio is automatically set, and when the scaling ratio is specified, the background image An image compositing method characterized by limiting superimposition positions within the image. 2. In a computer system that creates a composite image by superimposing a part of another digitally represented image or a part of a background image on a digitally represented background image as a superimposed image, Information representing the three-dimensional position of the surface of the object represented in the background image, viewpoint position information, and the absorption spectrum of light by the atmosphere are set in advance, and the depth in the background image and color changes due to atmospheric absorption are calculated. By using the relationship, when the superimposed depth in the background image is specified, the degree of color change is automatically set,
An image synthesis method characterized by limiting the depth of superimposition in a background image when the degree of color change due to atmospheric absorption is specified. 3. In a computer system that creates a composite image by superimposing a part of another digitally represented image or a part of a background image on a digitally represented background image as a superimposed image, Information representing the three-dimensional position of the surface of the object represented in the background image, viewpoint position information, and the absorption spectrum of light by the atmosphere are set in advance, and the two-dimensional position in the background image on which the superimposed image is superimposed is determined in advance. By using the relationship between the scaling ratio of this superimposed image and the relationship between the depth in the background image and the color change due to atmospheric absorption,
When the superimposition position in the background image is specified, the scaling ratio, depth, and degree of color change due to atmospheric absorption are automatically set, and when the degree of color change due to atmospheric absorption is specified. An image synthesis method characterized by limiting the depth of a background image to be superimposed, the scaling ratio, and the superimposition position in the background image. 4. The image synthesis method according to claim 1, wherein the information representing the three-dimensional position of the surface of the object is information representing a plane of the object. 5. In a computer system that creates a composite image by superimposing a part of another digitally represented image or a part of a background image on a digitally represented background image as a superimposed image, Prepare multiple images with the same scene but different magnifications, cut out and prepare a bitmap representation mask to represent a part of the area from each image with different magnifications, and then use the specified scaling factor. An image synthesis method characterized by performing superimposition synthesis by enlarging or reducing a mask with a magnification close to , to a specified value. 6. In a computer system that creates a composite image by superimposing a part of another digitally represented image or a part of the background image on a digitally represented background image as a superimposed image, it is relatively easy to use. Prepare a bitmap mask created using an image with a large reduction ratio and an outline mask that expresses a part of the image in a vectorial manner with closed curved boundaries. An image synthesis method characterized by performing superimposition synthesis by enlarging/reducing based on a mask, and performing superimposition synthesis by enlarging/reducing based on an outline mask in other cases. 7. In a computer system that creates a composite image by superimposing a part of another digitally represented image or a part of a background image on a digitally represented background image as a superimposed image, an arbitrary Prepare a tube-shaped solid whose cross section is always topologically equal when cut by a plane parallel to a predetermined direction, and pass through a predetermined position and apply a mask that specifies a part of the superimposed image. An image synthesis method characterized by defining a cutting plane parallel to a plane. 8. The tube-shaped solid body has a cross-sectional area that monotonically increases or decreases as the arbitrary positions are sequentially changed, and the size of the mask can be changed and adjusted by moving the plane in parallel. The image synthesis method according to claim 7, characterized in that: 9. In a computer system that creates a composite image by superimposing a part of another digitally represented image or a part of a background image on a digitally represented background image as a superimposed image, Both the image and the superimposed image are divided into small regions, and when superimposing, depth information is given to each small region, and based on this depth information, the nearer the small region is, the more it overlaps. Featured image synthesis method. 10. In a computer system that creates a composite image by superimposing a part of another digitally represented image or a part of a background image on a digitally represented background image as a superimposed image, An image synthesis method characterized in that the image density of a subsequent synthesized image is set to a value obtained by dividing the product of the image density of a superimposed image at a corresponding location and the image density of a background image by a constant independent of location. 11. In a computer system that creates a composite image by superimposing a part of another digitally represented image or a part of a background image on a digitally represented background image as a superimposed image, the background image Information representing the three-dimensional position of the surface of the object represented in the image, viewpoint position information, and light absorption spectra corresponding to each weather condition due to the atmosphere are set in advance, and by specifying the weather, the three-dimensional An image composition method characterized by applying a change in color tone according to a target position on a composite image. 12. Prepare a tube-shaped solid whose cut surface size changes depending on the cutting position, and cut this tube-shaped solid in the thickness direction at a certain position. Depending on the size and shape of the cutting plane, A character font generation method characterized by defining a character font. 13. The tubular three-dimensional cutting plane is characterized by being configured to generate a character font whose character size is determined according to its cutting position and whose strokes are determined by a thickness suitable for this character size. 13. The character font generation method according to claim 12. 14. The character font generation method according to claim 12, wherein the tube-shaped solid has a conical shape with an outline between an upper base surface and a lower base surface as outline fonts at different points.