JP2878614B2 - Image synthesis method and apparatus - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image synthesizing method and apparatus for synthesizing a plurality of images on a computer, handling and displaying the images as one continuous image.

[0002]

2. Description of the Related Art Conventionally, a technique for synthesizing a plurality of images, especially a plurality of maps, is intended for a map composed of vector data, or for correcting a map based on meridians and parallels of a raster data map. There is. As a technique for synthesizing a map composed of vector data, for example, there is a connection processing method for boundary vector data disclosed in Japanese Patent Application Laid-Open No. 63-206874. Japanese Patent Laid-Open No. 60-175182 discloses a map synthesizing apparatus that performs correction based on meridian latitude lines.

[0003]

However, the technique using vector data has a problem that it cannot be applied at all to raster data format map data read by an image scanner.

In correction using meridian latitude lines of raster data, meridians (latitude lines) are matched, and data between meridians (latitude lines) uses the same correction coefficient as that of the meridian (latitude lines). For this reason, the same correction is performed for a plurality of different error targets, and there is a problem that accurate correction cannot be performed.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and an object of the present invention is to accurately image images on a plurality of adjacent drawings without causing a positional shift on a connection boundary thereof. It is an object of the present invention to provide an image synthesizing method and apparatus capable of connecting and obtaining one seamless image.

Another object of the present invention is to provide an image synthesizing method and apparatus capable of obtaining a seamless image at high speed.

[0007]

In order to achieve the above-mentioned object, the present invention provides an image processing apparatus which is read by image input means.
Sets of points to be connected on adjacent images
After that, the point on the image where the image is to be corrected
The point on the image is set as the correction start point, and the correction origin is
Find the correction vector to be corrected and correspond to each pixel of the corrected image
The pixel position of the image before correction to be
On the other hand, the distance from the correction destination point to each pixel of the corrected image is
Multiplying by a coefficient that is inversely proportional and whose sum is "1"
Image after correcting the obtained pixel position value
Is set to the value of the pixel corresponding to .

Further, when obtaining the pixel movement position, the image to be corrected is divided into a plurality of square blocks, and if the difference between the movement vector of the vertex of each block and its average vector is small, the block is rejected. Pixels within are processed collectively according to the average vector.

[0009]

With respect to the displacement of the object appearing on the boundary between the two drawings, the point on one of the drawings to be corrected is set to the correction origin,
Assuming that a point on the other drawing is a correction destination point, a correction vector that makes the correction origin coincide with the correction destination point may be obtained, and the pixels on the drawing to be corrected may be shifted by the correction vector.

However, if the pixels are shifted at the same ratio regardless of the magnitude of the distance from the correction destination point, the positional deviation of the pixels located away from the boundary between the two drawings is too large, resulting in an unnatural correction result.

Therefore, the correction vector is multiplied by a coefficient such that the correction amount is large for a pixel close to the correction destination point, and the correction amount decreases as the distance from the pixel increases, and the pixel is shifted by the resulting vector.

Therefore, in the present invention, the position of the pixel of the drawing image before correction corresponding to each pixel of the drawing image after correction is set to the distance from the correction starting point to each pixel of the drawing image after correction with respect to the correction vector. The value of the obtained pixel position is set to the value of the corresponding pixel of the corrected drawing image by multiplying by a coefficient that is inversely proportional and whose sum is “1”.

[0013] This makes it possible to accurately connect the images on a plurality of adjacent drawings without causing a positional shift on the connection boundary, and obtain a single seamless image.

Further, the image to be corrected is divided into a plurality of square blocks, and when the difference between the moving vector of the vertex of each block and its average vector is small, the pixels in that block are grouped according to the average vector. In this case, the blocks that are farther from the seam of the image are processed in a batch according to the average vector more frequently, and a single seamless image is obtained at a much higher speed than in the case of correction in units of one pixel. be able to.

[0015]

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

FIG. 1 is a block diagram of an apparatus for executing an image synthesizing method according to the present invention.

The apparatus of the present embodiment includes an image scanner 10 as image input means for reading an image, a central processing unit 11 for synthesizing images, a disk 12 as an external storage device for storing a map, and a display for displaying a map. Display 13, a keyboard 14 and a mouse 15 which are position designation means for designating a point to be corrected.

The disk 12 may be a hard disk or a magneto-optical disk.

FIG. 2 is a diagram showing details of the central processing unit 11.

The central processing unit 11 calculates a correction vector for obtaining a correction vector for making the correction origin coincide with the correction origin by setting a point on the image to be corrected to the image as a correction origin and a point on the other image as a correction destination. The means 16 and the position of the pixel of the image before correction corresponding to each pixel of the image after correction is inversely proportional to the distance from the correction destination point to each pixel of the image after correction with respect to the correction vector, and Calculating means 17 for obtaining by multiplying by a coefficient whose sum is "1"; and pixel value setting means 18 for setting the value of the pixel position obtained by the calculating means 17 to the value of the corresponding pixel of the corrected image. Composed of

FIG. 3 is an explanatory diagram of an image display showing an actual processing result using the processing of the present invention.

The map A31 and the map B32 display images having distortion before correction read by using an image scanner. The map A33 and the map B'34 display images in which distortion has been corrected by correction.

The combination of the maps is performed by correcting one of the adjacent maps.

When a map is actually corrected, correction is performed on a boundary between the map and four adjacent maps. For the sake of simplicity, in this embodiment, one boundary between two maps is determined. A method for connecting will be described.

Here, the point on the drawing from which the correction is specified is referred to as a correction origin, the movement destination is referred to as a correction destination point, and a vector from the correction origin to the correction destination point is referred to as a correction vector.

Hereinafter, an embodiment of the present invention will be described in detail with reference to FIG.

First, two maps (map A31 and map B22) input from the image scanner 10 are displayed on the graphic display 13 as shown in the upper diagram of FIG. 3 (step 401).

In this embodiment, the map B32 is to be corrected. The corrected result is as shown in the lower diagram of FIG.

Next, an end point (端 mark) to be corrected and a correction destination point (x mark) as a correction origin are designated using the mouse 15. Therefore, a correction vector for making the correction origin coincide with the correction destination point is obtained based on the coordinate values of the correction origin and the correction destination point (step 402).

Next, the first pixel of the corrected image is set as a target pixel (step 403).

Next, the sum S of the reciprocals of the distance between the target pixel and all the correction destination points is obtained (step 404).

Further, a total sum of the products obtained by multiplying each correction vector by the reciprocal of the product of the distance from the pixel of interest and S is obtained (step 405). The corresponding pixel on the original image of the pixel of interest is obtained from the vector obtained in step 405 (step 4).
06).

It is determined whether the target pixel is the last pixel (step 407). If not, the next target pixel is designated (step 408), and the processing from step 404 to step 406 is repeated. If the target pixel is the last pixel in step 407, the process ends.

As for the entire process, the calculation for one pixel may be repeated for all pixels.

Next, the processing described in the flowchart of FIG. 4 is performed by actually executing the map B32 corresponding to the pixels of the map B'34.
A method for obtaining one pixel will be described.

Here, assuming that a vector symbol is represented by “♂” and that n correction vectors are designated, the correction origin is defined by ΔP
_i, ♂ correct destination point ♂P _{'i (where, i = 1,2, ... n)} , ♂x pixels on the map B32, Map B' pixels on 34
Expressed as x '.

As a policy, a vector ♂x-♂x 'representing the movement of a pixel is considered. At this time, the following items are followed.

(1) ♂x-♂x 'is the sum of each correction vector multiplied by a certain coefficient.

For example, ♂x-♂x '= α (♂P ₁ -♂
P ′ ₁ ) + β (♂P ₂ −♂P ′ ₂ ) (2) The coefficient is ♂P ′ of ♂x ′ so that the influence of a certain correction vector is large near the correction destination point and small at a distant point.
It is assumed to be inversely proportional to the distance from.

For example, if the distances are 3 and 5, α =
1/3, β = 1/5.

(3) However, if this coefficient is used as it is, discontinuous correction will be applied near the correction destination point. Therefore, normalization is performed so that the sum of the coefficients becomes "1". That is, it is divided by the sum of all coefficients.

When the coefficients of the example are recalculated, α = (1/3)
/ (1/3 + 1/5), β = (1/5) / (1/3 + 1)
/ 5).

A generalized calculation formula based on the above policy is shown below.

First, the sum S of the reciprocals of the distance between the target pixel and all the correction destination points is obtained (step 404).

[0045]

(Equation 1)

Next, using S, {x ′ on the map B ′ 34
Determine coordinates x on the map B 32 corresponding to the (Step 4
05, 406).

[0047]

(Equation 2)

The value of the pixel at the coordinate x on the map B32 may be set to the value of the pixel at the coordinate x 'on the map B'34.

As described above, according to the present embodiment, after specifying a plurality of sets of points to be connected on an adjacent map read by the image scanner 10, the points on the map to be corrected are set to the correction origin. A point on the other map is set as a correction point,
Find a correction vector that matches the correction origin with the correction destination point,
The position of the pixel of the image before correction corresponding to each pixel of the image after correction is inversely proportional to the distance from the correction destination point to each pixel of the image after correction with respect to the correction vector, and the sum is “1”. , And the value of the obtained pixel position is set to the value of the corresponding pixel of the corrected image. Connection can be made accurately without causing displacement, and one seamless image can be obtained.

FIG. 5 is a diagram for explaining the case where the position of the pixel of the image before correction corresponding to each pixel of the image after correction is obtained. In FIG. 5, the correspondence of the pixel of the image to be corrected is clearly shown. The circles represent one pixel. The upper side of FIG. 5 shows the state before correction, and the lower side of FIG. 5 shows the state after correction. It is also assumed that the right side of FIG. 5 is an image to be corrected, and that the pixel a is specified to correspond to the pixel o as a correction vector.

When the present invention is applied, as a result, the black pixels a to n move to ow and j, k, l, m, n.

This is given by ow, j,
This is obtained by calculating which pixel in the original image the pixel n corresponds to.

FIG. 6 is a diagram showing how the original pixel and the corrected pixel correspond to each other. For example, the pixel p corresponds to the pixel b, and the original pixel value of the pixel b is the corrected pixel value. Is the pixel p
It is shown that the pixel value is. Also, the original pixel f indicates that there is no corresponding pixel after correction, and the original pixel j indicates that it corresponds to two pixels w and j.

Next, an embodiment in which one seamless image is obtained at high speed will be described.

When one seamless image is to be obtained at a high speed, all the pixels in the image are not processed one by one, but portions which can be regarded as uniform motion vectors are processed collectively. Is desirable.

Therefore, in the second embodiment of the present invention, after specifying a plurality of sets of points to be connected on adjacent images read by the image scanner 10, the image to be corrected is placed on one side thereof. Is divided into a plurality of square blocks in which the number of pixels is a power of 2.

Then, for each vertex of each block, the pixel position before correction corresponding to the pixel position after correction is obtained by the method described in the first embodiment.

That is, the above-described equations 1 and 2 are executed, and a pixel position before correction corresponding to the pixel position after correction of each vertex is obtained. Thereby, the movement amount of each vertex can be determined. Hereinafter, in this specification, the movement amount of each vertex obtained by executing the above-described equations 1 and 2 is defined as a movement vector.

After the average vector of the movement vectors of each vertex is obtained, a difference vector between the movement vector of each vertex and the average vector is obtained.

Next, when the maximum value of the four difference vectors is smaller than a predetermined threshold value, the average vector is determined as a movement vector of the pixels of the entire block, and each pixel in the block is determined using the average vector. Is set to the value of the corresponding pixel.

If the maximum value of the difference vector is larger than a predetermined threshold value, the block is divided into four equal parts, a moving vector of each vertex of each divided block is obtained, and an average vector of the moving vectors is obtained. Then, a difference vector between the movement vector of each vertex and the average vector is obtained.

This process is repeated until the number of pixels on one side of the divided block becomes two or until the maximum value of the difference vector becomes smaller than the threshold value.

FIG. 7 is a block diagram showing a detailed configuration of the central processing unit 11 in an apparatus for performing a high-speed image synthesizing process. The image to be corrected has two pixels on one side.
Dividing means 19 for dividing into a plurality of square blocks each of which is a power of, moving vector calculating means 16 for calculating a moving vector for each vertex of each of the divided blocks, and averaging for calculating an average vector of the calculated moving vectors. A vector calculating means 20, a difference vector calculating means 21 for obtaining a difference vector between each moving vector and the average vector, a comparing means 22 for comparing whether or not the maximum value of the difference vector is smaller than a predetermined threshold value; If the maximum value of the difference vector is smaller than a predetermined threshold value based on the comparison result of the means 22, the average vector is determined as a movement vector of the pixels of the entire block, and each pixel of the image after correction of the block is determined. The difference vector is calculated based on the comparison result of the calculating means 17 for obtaining the pixel position of the corresponding image before correction and the comparing means 22. If the maximum value of the block is larger than a predetermined threshold value, the block is further divided into four equal parts, the movement vector of each vertex of each divided block is obtained, and the arithmetic unit 17 corrects the corrected image of the divided block. Repetition control means for repeatedly executing the processing for obtaining the pixel position of the image before correction corresponding to each pixel until the number of pixels on one side of the divided block becomes two pixels or the maximum value of the difference vector becomes smaller than the threshold value 23 and setting means 18 similar to that of FIG.

The high-speed image synthesizing process will be described below with reference to the flowcharts of FIGS.

First, two maps (map A31 and map B32) input from the image scanner 10 are displayed on the graphic display 13 as shown in the upper part of FIG. 3 (step 501).

In this embodiment, the map B32 is to be corrected. The corrected result is as shown in the lower diagram of FIG. Next, the user uses the mouse 15 to specify the end point of the correction target (Δ mark) which is the correction origin and the correction destination point (× mark). Therefore,
A correction vector for making the correction origin coincide with the correction destination point is obtained based on the coordinate values of the correction origin and the correction destination point (step 502).

Next, as shown in FIG. 10, the corrected image is divided into square blocks B1 to Bj each composed of pixels of which one side is “2n × 2n” (n ≧ 1 integer). Divide (step 503).

Next, the head block B1 of the blocks B1 to Bj is set as a target block (step 504). Then, for the pixels in the block of interest, corresponding pixels are obtained (step 505).

Next, it is determined whether the designated block of interest is the last (step 506). If not, the next block of interest is designated (step 507), and the block of interest is the last in step 506. If so, the process ends.

Next, a description will be given of the speed-up processing in units of blocks obtained in step 503.

The processing for this block will be described with reference to FIG.

A movement vector representing the movement amount of the pixel corresponding to the four vertices P1 to P4 of the block divided in step 503 in FIG. 8 is obtained (step 601).

If the block size is “2 × 2” pixels (step 602), the obtained vector is directly used as a movement vector (step 606).

Otherwise, first, the four vertices P1 to P1
An average vector of the movement vector of P4 is obtained (step 603). Next, a difference vector between the movement vector of the four vertices and the average vector is obtained, and it is determined whether or not the maximum value is equal to or less than “1” (threshold) (step 604).
If it is equal to or less than "1" (threshold), the average vector is adopted as a movement vector adapted to all pixels in the block (step 605).

If the maximum value of the difference vector is equal to or greater than "1", each side of the current block is further divided into four equal parts (step 607).

For example, as shown in block B3 in FIG.
The image is further divided into four, i. Thereafter, the processing from step 504 in FIG. 8 is recursively called and executed for the upper left block b1, and the same processing is performed in the upper left block b1 until the number of pixels on one side becomes 2 pixels, or the maximum value of the difference vector becomes It repeats until it becomes "1" or less (step 608).

In this case, if the maximum value of the difference is "1" or more in block b3, b31, b32, b33,
The process is further divided into b34 and the same processing is repeated.

Similarly, processing is performed on the upper right block b2, the lower right block b3, and the lower left block b4 (step 6).
09,610,611).

The detailed processing in step 505 in FIG. 8 and steps 605 and 606 in FIG.
This is the same as the processing from step 04 to step 406.

As described above, when determining the pixel movement position, the image to be corrected is divided into a plurality of square blocks, and if the difference between the movement vector of the vertex of each block and the average vector is small, the Since the pixels in the block are collectively processed according to the average vector, the cases where the distance from the seam of the image is farther from the seam of the block increases the cases where the blocks are collectively processed according to the average vector. One seamless image can be obtained.

[0081]

As described above, according to the present invention, an arbitrary number of corresponding points are designated near the boundary between a plurality of drawing images adjacent to each other, and how the other pixels move according to the designation. By calculating whether to do so, a plurality of adjacent drawing images can be accurately connected without causing a positional shift on the connection boundary, and one seamless image can be obtained.

When determining the pixel movement position, the image to be corrected is divided into a plurality of square blocks, and if the difference between the correction vector of the vertex of each block and the average vector is small, the inside of the block is determined. Pixels are collectively processed according to the average vector, so that a single seamless image can be obtained at high speed.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of an apparatus for executing an image synthesizing method according to the present invention.

FIG. 2 is a detailed configuration diagram of a central processing unit in FIG. 1;

FIG. 3 is an explanatory diagram illustrating an example of an image before combination and an image after combination.

FIG. 4 is a flowchart illustrating an image combining method according to the present invention.

FIG. 5 is an explanatory diagram of obtaining a pixel position of an image before correction corresponding to each pixel of an image after correction.

FIG. 6 is an explanatory diagram showing a positional relationship between an image after correction and an image before correction in FIG. 5;

FIG. 7 is a detailed configuration diagram of a central processing unit for realizing image synthesis at high speed.

FIG. 8 is a flowchart illustrating a processing procedure for speeding up image synthesis.

FIG. 9 is a flowchart showing details of step 505 in FIG. 8;

FIG. 10 is an explanatory diagram illustrating an example of a method for dividing an image.

[Explanation of symbols]

10 image scanner, 11 central processing unit, 12
Disk, 13 ... Graphic display, 14 ... Keyboard, 15 ... Mouse, 16 ... Correction vector calculation means, 17 ... Calculation means, 18 ... Pixel value setting means, 19 ... Division means, 20 ... Average vector calculation means, 21 ... Difference vector Calculation means, 22 ... comparison means, 23 ... repetition control means 3
1: A side serving as a reference of a map to be combined, 32: a side for adding a correction of a map to be combined (before correction), 34: a side for adding a correction for a map to be combined (after correction).

Continuation of the front page (72) Inventor Minoru Takasaki 6-81 Onoe-cho, Naka-ku, Yokohama-shi, Kanagawa Prefecture Within Hitachi Software Engineering Co., Ltd. (56) References JP-A-5-108819 (JP, A) (58) ) Surveyed field (Int.Cl. ⁶ , DB name) G06T 1/00 G09G 5/36 H04N 1/387

Claims (4)

(57) [Claims] 1. An image synthesizing method for reading images on a plurality of adjacent drawings by image input means, synthesizing the images on a computer, and displaying the images, wherein the images should be connected on adjacent images read by the image input means. After specifying a plurality of pairs of points, a point on the image to be corrected on the image is set as a correction origin, a point on the other image is set as a correction destination point, and a correction vector for matching the correction origin with the correction destination point is obtained. The position of the pixel of the image before correction corresponding to each pixel of the image is inversely proportional to the distance from the correction destination point to each pixel of the image after correction with respect to the correction vector, and the sum is “1”. An image synthesizing method, wherein the value is obtained by multiplying by a coefficient, and the value of the obtained pixel position is set to a value of a corresponding pixel of the corrected image. 2. A method for designating a plurality of points to be connected on adjacent images read by the image input means, and setting an image to be corrected to a power of 2 pixels on one side. A first step of dividing the block into a plurality of square blocks, a second step of obtaining a movement vector using the correction vector for each vertex of each block, and obtaining each of the average vectors of the movement vectors. A third step of obtaining a difference vector between the movement vector of each vertex and the average vector, a fourth step of comparing the difference vector of each vertex with a predetermined threshold, If the maximum value is smaller than the threshold value, a fifth step of determining an average vector as a movement vector of the pixels of the entire block; When the large value is larger than the threshold value, the block is divided into four equal parts, and for each divided block, the second to fifth steps are performed until the number of pixels on one side of the divided block becomes two pixels. 6. The image synthesizing method according to claim 1 , further comprising: a sixth step of recursively repeating until the maximum value of the difference vector becomes smaller than the threshold value. 3. An image synthesizing apparatus for reading images on a plurality of adjacent drawings by means of image input means, synthesizing them on a computer, and displaying the images, wherein the images should be connected on adjacent images read by said image input means. Specifying means for specifying a plurality of sets of points, a correction vector for obtaining a correction vector that sets a point on the image to be corrected to the image as a correction origin, a point on the other image as a correction destination point, and a correction origin that matches the correction destination point Calculation means, the position of the pixel of the image before correction corresponding to each pixel of the image after correction, for the correction vector, is inversely proportional to the distance from the correction destination point to each pixel of the image after correction, and the Computing means for obtaining by multiplying by a coefficient whose sum is "1", and setting means for setting the value of the pixel position obtained by the calculating means to the value of the corresponding pixel of the corrected image. Image synthesizing apparatus according to claim. 4. A designating means for designating a plurality of sets of points to be connected on adjacent images read by the image input means, wherein the number of pixels on one side of the image to be corrected is a power of 2 Dividing processing means for dividing into a plurality of square blocks; moving vector calculating means for obtaining a moving vector for each vertex of each block using the correction vector calculated by the correcting vector calculating means; After calculating an average vector of the vectors, a difference vector calculating means for calculating a difference vector between the movement vector of each vertex of each block and the average vector, and a comparing means for comparing the difference vector of each vertex with a predetermined threshold value If the maximum value of the difference vectors of the vertices is smaller than the threshold value, the average vector is adopted as the movement vector of the pixels of the entire block. Determining means; and when the maximum value of the difference vectors of the vertices is larger than the threshold value, the block is divided into four equal parts, and for each divided block, the moving vector calculating means, the difference vector calculating means, the comparing means And control means for recursively executing the processing of the determination means until the number of pixels on one side of the divided block becomes two pixels or until the maximum value of the difference vector becomes smaller than the threshold value. The image synthesizing device according to claim 3 .