JPS62140172A - Image synthesizing method - Google Patents

Abstract

PURPOSE:To obtain an image having no unnaturalness at the joint part of images by controlling variably the output rate between the output data of both images in response to their positions within the overlapped part. CONSTITUTION:The picture elements of image data A and B stored in both memories 12 and 13 are read out successively by address counters 10 and 11 and then synthesized. A selector 14 selects and outputs one of both outputs of memories 12 and 13 by a signal SEL given from a controller 15. The controller 15 positions the images at coordinates (i) and (j) in terms of the overlapped part of images and defines the picture element read out first as '101' and the picture element read out last as '102' respectively. The signal SEL is set at a single bit and equal to '1' in an odd address and then to '0' in an even address. Then the selection output of the image close to a position within the overlapped part is set larger than the selection output of the image distant from said position. This eliminates the unnaturalness of an image.

Description

[Detailed Description of the Invention] Field> The present invention relates to an image synthesis method for synthesizing two images,
In particular, the present invention relates to a method for processing joints between images.

<Prior Art> Conventionally, when two images are combined, outputted, and recorded, no special processing has been performed on the boundary of the joining portion of the images. For example, when compositing a car or a person against a landscape background, the boundaries between the composites often appeared to be visually emphasized, creating a sense of discomfort.

(Objective) The present invention aims to provide an image synthesis method that eliminates the drawbacks of the above-mentioned conventional example and eliminates the sense of discomfort at the joint.

<Example> An embodiment of the present invention will be described below with reference to the drawings.

Now, consider the junction between image A and image B as shown in FIG. At this time, an overlapping portion 1c of images A and H is provided at this junction. Furthermore, for the sake of explanation, it is assumed here that the systematic dither method is used for output.

Also, the dither matrices of both images are set to have a size of 4×4. 2a to 2d in FIG. 2 are examples of outputs by the dither method. Now, consider a group 3 in which the matrices 2a to 2d, as surrounded by the outline in FIG. 2, form one set in the overlapping portion 1c. Further, the inside of the overlapping portion 1c is divided into three parts, for example, as shown in 4a to 4C in FIG. 3.

The portion 5 in FIG. 3 corresponds to the portion 1a in FIG. 1. Since only image A is output here, the entire matrix outputs and records the image data of the corresponding position of image A as shown in FIG. 4(a).

Next, in the part 4a of FIG. 3, the ratio of the image data of image A and the image data of image B is 3:1.
As shown in figure (b), image data at the corresponding position of image B is output for one of the four matrices in the group, and image data at the corresponding position of image A is output for the other three matrices. and record it. In this way, the ratio of the image data of image A to the image data of image B is similarly set to 2:2 for 4b and 1:3 for 4C with respect to FIGS. 4b and 4c, and
), (d) to achieve this.

A portion 6 in FIG. 3 corresponds to the IC portion in FIG. 1, and contrary to 5, the image data of image B is output as shown in FIG. 4(e).

An embodiment that realizes this will be described with reference to FIGS. 5 to 10. FIG. 5 is a block diagram of an embodiment of the present invention. Image data A and B are combined by the address counter 10 for image A and the address counter 11 for image B, respectively, to read out the pixel data of the 0 image data stored in the image A memory 122 and the image B memory 13, respectively. , are performed in pixel order of the output image. The outputs of these image A and B memories 12 and 13 are input to the selector 14, selected by the SEL signal of the controller 15, and outputs pixel data of either one. The SEL signal is a 1-bit signal, and the selection circuit outputs the image A memory output when the SEL signal is 1, and outputs the image B memory when the SEL signal is 0. This output is sent to the printer 1 via the dither processing circuit 16.
7 is output.

The operation of the controller 15 will now be described.

Now, as shown in FIG. 6, the image position is determined using the coordinates i and j for the image overlap portion shown in FIG. 3. It is assumed that x+j is read from the memory and output in the direction of the arrow. In other words, for ■, first read from js, read up to je, move to the next i, and similarly read j
Let's read out. Also, the first pixel read out in the overlapping part is pixel 101, that is, (is, js),
The last pixel to be read out is pixel 102, that is, (ie,
je).

Next, before describing the method for determining the SEL signal, the subroutine to be used will be described.
bl) is a subroutine for selecting output from the same image data for all pixels from jS to je for i. Next, subroutine 2 (sub2) performs j for i.
This is a subroutine for alternately selecting outputs from two image data from S to je. In particular, this subroutine 2 outputs the SEL signal 1 if the address is an odd number, and outputs 0 if the address is an even number. In other words, the data of image A is output to odd addresses, and the data of image B is output to even addresses as composite image data. Finally, subroutine 3
(sub3) is the opposite of subroutine 2, in j, image B data is output to odd addresses and image A data is output to even addresses as composite image data.For three subroutines of 0 or more, see Figure 7. - Fig. 9 shows respective flowcharts.

Finally, the main routine for determining the SEL signal is as shown in the flowchart of FIG.

First, when the overlapping portion is divided as shown in FIG. 6, the number n of rows of i in the divided portion is determined.

Regarding the non-overlapping part (iris) of image A in FIG.
The EL signal is set to 1, and only the data of image A is output from the 1 selector 14.

When i becomes is, start outputting the overlapped part. 1st
Next, the output is determined for the part (is≦i<i s +n) in FIG. 4a. This outputs both image A and image B data alternately to row i by subroutine 2, and
Only the data of image A is output in one line by subroutine 1. In other words (i.

j), (i+1, D, (i, j+1), (i+
1, j+1), the structure is as shown in Figure 4(b).

Second, the part of FIG. 6 4b (i s+n≦i<is+2
Determine the output for n). Here, image A1 image B
The proportions occupied by the image data are the same. Therefore, a checkered pattern based on the image data of images A2 and B can be realized by alternately executing subroutine 2 and subroutine 3. In other words, FIG. 4(c) can be realized.

Thirdly, the part in FIG. 6 40 (i s+2n≦i≦ie)
Determine the output for. Here, the proportions occupied by the image data of the image A1 and the image B are reversed from that of the portion shown in FIG. 4A. Therefore, FIG. 4(d) is realized by outputting only the data of image B to line i by subroutine 1, and alternately outputting both data of image A and image B by subroutine 2 to line i+1. i becomes ie, the ie-th row determines the SEL signal by the same method, and (is.

When the same output is completed for je), the output determination of the pixel data of the overlapping portion is completed.

Finally, the SEL signal is set to 0 for the non-overlapping portion (ie<i) of image B in FIG. 6, and only the data of image B is output from the selector 14.

In this way, by variably controlling the output ratio of the output data of both images in accordance with the position within the two overlapping parts, an image that does not give any strange feeling at the joint can be obtained. In particular, it is possible to obtain a more natural image.

In the above example, the number of matrices constituting a group was set to 4, and three-stage grading was performed, but smoother multi-stage grading can be obtained by increasing the number of matrices. Furthermore, by increasing the width of the overlapping portion as shown in FIG. 1c, it is possible to perform fade-outs and fade-ins in a wide area of the image. Moreover, the same thing can be achieved not only by dithering but also by other methods if it is possible to configure one group from a plurality of output units.

Furthermore, it is possible to use the present invention not only as a recording method but also as an indicator such as a display.

Further, since the output of subroutine 3 of the controller 15 in the embodiment is the negation of the output of subroutine 2, it is of course possible to configure using that.

Effect> As explained above, according to the unreleased image synthesis method,
Since the image is synthesized by blurring the boundaries, it has the effect of enabling visually consistent image synthesis and grading-like expression.

[Brief explanation of drawings]

Fig. 1 is an explanatory diagram of a joining part of images, Fig. 2 is an explanatory diagram of a group of matrices, Fig. 3 is an explanatory diagram of an overlapping part of images, Fig. 4 is an explanatory diagram of the distribution of image data in a matrix,
FIG. 5 is a block diagram of the embodiment, FIG. 6 is a diagram showing pixel coordinates of an overlapping portion of images, and FIGS. 7 to 10 are control flowcharts of the controller in the embodiment. In the figure, IC is an overlapping part of the image, 3 is a group of dither matrices, 4a to 4C are divided areas within the overlapping part, 10.11 is an address counter, 14 is a selector, 15
1 represents a controller, and 17 represents a printer. '! 512]

Claims (1)

[Claims] (1) When combining and outputting two images, an overlapping part is created at the boundary, and the output data from each image is distributed within the overlapping part at a ratio according to the position of the overlapping part. An image synthesis method that blurs the image.