JPH10294855A - Synthesis method for two-dimension image data - Google Patents

Abstract

PROBLEM TO BE SOLVED: To synthesize pluralities of 2-dimension images read by a handy scanner without positional deviation automatically and without needing manual operation such as delicate dragging. SOLUTION: Pluralities of check areas CH nearly at an equal interval in the direction Y are set around a middle of a part where reference image data and edit side image data are in duplicate in the direction X, and pixel data of the reference side check areas CH and pixel data at the edit side around them are compared. The edit side area where a total sum of differences of the pixel values is minimum through the comparison is used for an approximated area CH', it is moved in the direction X based on indication values Δx, Δy and a line number in the direction Y up to the position of the check area from an upper end is corrected. As to 2nd and succeeding check areas, an edit side image as the result of movement is used for a new edit side image and it is moved in the direction X based on indication values Δx, Δy and a line number in the direction Y up to the position of the preceding and this check areas.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a method of synthesizing two-dimensional image data by editing an overlapping part of a plurality of two-dimensional image data read by a handy scanner and synthesizing the same.

[0002]

2. Description of the Related Art As shown in FIG. 8, a handy scanner 1, which is a handy type image scanner, comprises an optical system component such as an illuminating lamp 3 and an image pickup device 4 in a housing 2 having a size operable with one hand. A subject (read image) 5 is illuminated by an illumination lamp 3, the reflected light is reflected by a mirror 7 through a slit, and is imaged on an imaging device 4 such as a CCD line sensor via a lens 8. The main scanning is performed by the image pickup device 4 and converted into an electric signal.

[0003] When the handy scanner 1 is moved in the Y direction while rolling the Y direction roller 9 on the subject 5, the rotary encoder 1 linked with the Y direction roller 9 is moved.
By 0, the moving distance in the Y direction is detected. Then, two-dimensional image data is obtained by the main operation of the reading width w for one line in the X direction and the manual sub-scan in the Y direction.

The handy scanner 1 further includes an X-direction roller 11 orthogonal to the Y-direction roller 9 and a rotary encoder 12 interlocked with the X-direction roller 11 in order to read an object 5 having a reading width w of one line or more. Is provided. That is, as shown in FIG. 9, the handy scanner 1
After performing Y ₁ ′ manual sub-scanning, the handy scanner 1 is tilted, and the X-axis
The direction roller 11 rolls and moves in the X direction.

At this time, the moving distance in the X direction is detected by the rotary encoder and the photo interrupter 12, but the object 5 is not read. After moving in the X direction a distance X ₁ ′ that does not exceed the reading width w, the Y-direction roller 9 is brought into contact with the subject 5 again, and one line of main scanning and Y scanning are performed.
Manual sub-scanning of the distance Y ₂ ′ is performed in the direction. During the sub operation in the Y direction, the X direction roller 11 is lifted from the subject, and its rotation is stopped.

In the same manner, the handy scanner 1 is
(In FIG. 9, X ₂ ′ and Y ₃ ′)
By repeating the above operation until all the subjects 5 are read,
Since the moving distances and directions (positive and negative) in the X direction and the Y direction are detected, a plurality of two-dimensional image data in which a part of the adjacent X direction overlaps can be obtained.

[0007]

According to the above-described conventional handy scanner 1, a subject 5 having a size equal to or greater than the reading width w of one line can be read by a series of operations, but the handy scanner 1 is tilted. In this case, the drive timing of the rotary encoders 10 and 12 may be shifted, or the X and Y direction rollers 9 and 11 may make sliding contact with the paper surface of the subject 5. Errors may occur.

As described above, when an error occurs in the moving distance,
As shown in FIG. 10, between two-dimensional image data A and B, B and C
, The two-dimensional image data read accurately cannot be continued.

Therefore, conventionally, a plurality of read two-dimensional image data are displayed on a display or the like, and the two-dimensional image data having a positional shift is edited by a manual operation so as to be continuous with each other. That is, two of the overlapping portions of one editing-side two-dimensional image data (for example, B in the drawing)
Pixels 13B and 14B at arbitrary positions are arbitrarily extracted, and the two-dimensional image data on the editing side is matched with the corresponding pixels 13A and 14A of the two-dimensional image data on the reference side (for example, A in the figure) by a mouse drag operation or the like. The whole is moved and both are combined.

However, in the case of a slight displacement, it is a difficult task to finely move the pixels 13B and 14B extracted by an operation such as dragging while confirming it on the display.

In addition, it is necessary to perform the above-mentioned editing work in each overlapping portion of a plurality of two-dimensional image data, and the compositing work is troublesome.

An object of the present invention is to provide a method for synthesizing two-dimensional image data in which a plurality of two-dimensional image data are synthesized without positional deviation without performing an editing operation by manual operation.

[0013]

According to a first aspect of the present invention, there is provided a method of synthesizing two-dimensional image data, comprising: disposing a handy scanner on a subject; A manual scanning in the Y direction to generate two-dimensional image data consisting of a scanning width of one line and a sub-scanning width in the Y direction;
The X-direction movement operation of moving the handy scanner in the X direction within the line reading width is repeated to generate at least two or more two-dimensional image data in which a part of the adjacent X direction overlaps, and the image data of the overlapping portion is generated. In the two-dimensional image data synthesizing method for editing the two-dimensional image data so that the two-dimensional image data coincide with each other, (a) the overlapping part between the reference two-dimensional image data and the adjacent two-dimensional image data on the editing side includes the i-th row and the j-th column. A plurality of check areas CH composed of pixel data are set in the Y direction with reference to the reference-side two-dimensional image data.
The pixel data included in (K) is referred to as reference pixel data D (i,
j), and (c) the K-th check area C
A predetermined number of comparison areas CHx, y are set at equal pitches in the X and Y directions around H (K), and (d) pixel data included in each comparison area CHx, y from the editing-side two-dimensional image data Is extracted as the comparison pixel data Dx, y (i, j), and (e) the sum of the differences between the respective comparison pixel data Dx, y (i, j) and the corresponding pixel data of the reference pixel data D (i, j). From the comparison areas CHΔx, Δy in which the coincidence rate αx, y has the minimum coincidence rate α (K).
As the approximate area CH (K) ′, the approximate area CH
The editing instruction values Δx and Δy for moving (K) ′ to the check area CH (K) are obtained, and (g) the processing from (b) to (f) is repeated for all the check areas CH,
The editing instruction values Δx and Δy are obtained for each check area CH,
(H) Approximate area CH based on edit instruction values Δx and Δy
The edit-side two-dimensional image data including 'is matched with the corresponding check area CH and edited, and the reference-side two-dimensional image data and the edit-side two-dimensional image data are combined.

The editing-side two-dimensional image data enclosed by the approximation area CH (K) 'is the one that most closely resembles the reference-side two-dimensional image data enclosed by the check area CH (K). It can be estimated that a position shift has occurred between the approximate area CH (K) ′ and the check area CH (K). Accordingly, for all the check areas CH (K) set in the overlapping portion, the relative position of the approximate area CH (K) ′ with respect to the check area CH (K) is determined by the editing instruction value Δ
x, Δy, and moving the editing-side two-dimensional image data including the approximation area CH (K) ′ by the editing instruction values Δx, Δy to match the check area CH (K). Can be automatically synthesized without displacement.

According to a second aspect of the present invention, there is provided a method of synthesizing two-dimensional image data, comprising: disposing a handy scanner on a subject, performing one main scan in the X direction, and manually sub-scanning the handy scanner in the Y direction. An image reading operation for generating two-dimensional image data including a line reading width and a sub-scanning width in the Y direction, and an X direction moving operation for moving the handy scanner in the X direction within at least one line reading width are repeated.
In a two-dimensional image data synthesizing method for generating at least two or more two-dimensional image data in which a part of adjacent X-directions overlap and editing the two-dimensional image data so as to match the image data of the overlapping part, I) In the overlapping portion between the reference-side two-dimensional image data and the adjacent editing-side two-dimensional image data, i
A plurality of check areas CH composed of pixel data in the row j column are set in the Y direction with reference to the reference-side two-dimensional image data, and the K-th check area CH (K) is determined from the (nu) reference-side two-dimensional image data. Is extracted as reference pixel data D (i, j), and a predetermined number of comparison areas CHx, y are set at equal pitches in the X and Y directions around the check area CH (K). Then, (画素) pixel data included in each comparison area CHx, y is extracted as comparison pixel data Dx, y (i, j) from the editing-side two-dimensional image data, and (v) each comparison pixel data Dx, y A matching rate αx, y is obtained from the sum of differences between (i, j) and corresponding pixel data of the reference pixel data D (i, j), and (f) a matching rate αx,
The comparison area CHΔx, where y is the minimum coincidence rate α (K),
Let Δy be the approximate area CH (K) ′ and the approximate area CH
(K) ′ are determined to be edit instruction values Δx and Δy for moving to the check area CH (K).
(K) is the first check area CH from the upper end of the overlapping portion
In the case of (1), the approximate area CH (1) ′ from the upper end of the editing-side two-dimensional image data according to the editing instruction value Δy.
The check area CH (1) and the approximate area CH (1) ′ are made to coincide on the same line, and the entire edit-side two-dimensional image data is moved in the X direction according to the edit instruction value Δx. If the edited check area CH (K-1) is the second and subsequent check areas CH (K-1), the two-dimensional image data edited in the immediately preceding check area CH (K-1) is regarded as new editing-side two-dimensional image data. Thus, for each check area CH (K), the matching rate α (K) and the editing instruction value Δ
x, Δy, and (v) the approximate area CH from the immediately preceding check area CH (K−1) according to the editing instruction value Δy.
When the number of lines up to (K) ′ is corrected and the coincidence rate α (K) of the check areas CH (K) is the minimum value α _MIN (K) of all the check areas CH (K),
According to the edit instruction value Δx, the entire edit-side two-dimensional image data is moved in the X direction, and the (S) process is repeated to edit the edit-side two-dimensional image data for all the check areas CH (K), and The two-dimensional image data and the editing-side two-dimensional image data are combined.

The editing-side two-dimensional image data enclosed by the approximation area CH (K) 'is the one that most closely resembles the reference-side two-dimensional image data enclosed by the check area CH (K). It can be estimated that a position shift has occurred between the approximate area CH (K) ′ and the check area CH (K).

The relative position of the approximate area CH (K) 'with respect to the check area CH (K) is determined by editing instruction values Δx, Δy
Since the approximate area CH (K) ′ is moved by the editing instruction value Δy and the number of lines corresponding to the movement is corrected, the number of lines between the check area CH (K) of the two-dimensional image data on the reference side and the editing side is corrected. Can be matched.

Check area CH (K) coincidence rate α
If (K) is the minimum value of all the check areas CH (K), it can be estimated that the matching rate between the approximate area CH (K) ′ and the check area CH (K) is high. By moving the entire editing-side two-dimensional image data in the X direction in accordance with the editing instruction value Δx, it is possible to eliminate the displacement in the X direction between the check areas CH (K),
It can be adjusted to the closest position.

For the second and subsequent check areas CH (K), the two-dimensional image data edited in the immediately preceding check area CH (K-1) is regarded as new two-dimensional image data on the editing side. The displacement between the two-dimensional image data and the reference-side two-dimensional image data is smaller. Therefore, the comparison pixel data Dx, y (i, j)
The difference between the pixel data of the reference pixel data D (i, j) and the reference pixel data D (i, j) becomes a small value, and the calculation of the coincidence rate αx, y becomes easy.

According to a third aspect of the present invention, there is provided a method for synthesizing two-dimensional image data, wherein (ii) when editing the two-dimensional image data on the editing side which overlaps at the overlapping part, the other overlapping part in the direction of the two-dimensional image data on the editing side. After editing all the two-dimensional image data connected with the editing-side two-dimensional image data together with the editing-side two-dimensional image data, and (ii) combining the reference-side two-dimensional image data and the editing-side two-dimensional image data, The process of (T) is repeated as new reference-side two-dimensional image data, and continuous two-dimensional image data via another overlapping portion is set as new editing-side two-dimensional image data. It is characterized by combining all two-dimensional image data in which three or more two-dimensional image data are continuous.

When editing two-dimensional image data on the editing side,
Since all the two-dimensional image data connected in the direction of the editing-side two-dimensional image data are edited at the same time, the relative position between them does not change. Therefore, the editing does not increase the positional deviation between the continuous two-dimensional image data via another overlapping portion, and does not complicate the calculation of the coincidence rate αx, y.

According to a fourth aspect of the present invention, there is provided a method for synthesizing two-dimensional image data. After editing the two-dimensional image data together with the two-dimensional image data on the editing side, and (m) combining the two-dimensional image data on the reference side and the two-dimensional image data on the editing side, the two-dimensional image data on the reference side is converted into a new two-dimensional image on the editing side. The two-dimensional image data connected to the reference-side two-dimensional image data in the direction of the reference-side two-dimensional image data via another overlapping portion is defined as new reference-side two-dimensional image data (a).
Is repeated, and (c) all two-dimensional image data in which three or more two-dimensional image data are continuous in two or more overlapping portions are synthesized.

After combining the reference-side two-dimensional image data and the edit-side two-dimensional image data, the reference-side two-dimensional image data is used as new edit-side two-dimensional image data. Is used as new reference-side two-dimensional image data, so that the editing-side two-dimensional image data can always be compared with the unedited reference-side image data even if the editing operation in a plurality of overlapping portions is repeated. . Therefore, it is possible to combine a plurality of two-dimensional image data without displacement, and to combine the overlapping portions in a state similar to the subject.

[0024]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Referring to FIGS.
A method for synthesizing two-dimensional image data according to the first embodiment of the present invention will be described.

FIG. 1 shows three types of two-dimensional image data A, B, and C to be combined. The two-dimensional image data shown in FIG. 10 is normalized before the combining process. That is, the upper and lower portions of the two-dimensional image data (indicated by broken lines in the drawing) in the Y direction are adjusted to the shortest two-dimensional image data, and the length in the Y direction is made equal to Y.

As shown in FIG. 1, two-dimensional image data A, B,
C is continuous in the X direction at two overlapping portions,
In the present embodiment, first, two-dimensional image data A is used as reference-side two-dimensional image data, and two-dimensional image data B is used as editing-side two-dimensional image data.

The reference-side two-dimensional image data A and the editing-side two-dimensional image data B overlap at an overlapping portion O _AB , and the overlapping portion O _AB
3 check areas CH1 _AB , CH2 _AB , CH3
Set _AB . Each check area CH (K)
Contains the pixel data D (i, j) of L rows and M columns, and is set so that the center thereof is located at a substantially equal interval in the Y direction in the middle of the overlapping portion O _{AB in} the X direction. For example, as shown in FIG. 1, the check area CH1 _AB is located at a position of [X ₁ + (w−X ₁ ) / 2] in the X direction with respect to the reference side two-dimensional image data A and is Y / 3 from the upper end. Check location CH2
_AB is at the same position in the X direction and 2Y / 3 from the top,
The check area CH3 _AB is set at a position near the lower end at the same position in the X direction. Similarly, check areas CH1 _BC , CH2 _BC , and CH3 _{BC which} will be described later, which are set in the overlapped portion O _BC, have Y in [X ₁ + X ₂ + (w−X ₂ ) / 2].
They are set on the axis, at Y / 3, 2Y / 3 from the upper end, and at positions near the lower end, respectively.

As shown in FIG. 2, the reference-side two-dimensional image data A of the portion represented by the check area CH1 _AB of the overlapping portion O _AB has reference pixel data DA (i, j), and the editing-side two-dimensional image data B of the same portion is comparison pixel data DB (i, j) including pixel data of L rows and M columns.

If there is no displacement between the reference-side two-dimensional image data A and the editing-side two-dimensional image data B, the reference pixel data DA (i, j) and the comparison pixel data DB (i, j) are:
Are identical. The sum α of the difference between the corresponding pixel data is

[0030]

(Equation 1)

Α is the reference pixel data DA
It becomes an index indicating the coincidence ratio between (i, j) and the image of the comparison pixel data DB (i, j), and becomes 0 if they are completely the same.

Each pixel data represents one dot of color information, and is composed of 24-bit data in which each of the three primary colors of RGB is represented by 8-bit data. In the above equation 1, DA _R (i, j) -DB _R (i, j) is R
The difference between the components is represented by DA _G (i, j) −DB _G (i, j).
The difference between the G components is obtained by calculating the difference between DA _B (i, j) and DB _B (i, j).

If there is a displacement between the two-dimensional image data A and B, the image corresponding to the reference pixel data DA (i, j) is
It has moved around the check area CH1 _AB . Therefore, a number of comparison areas CHx, y (i, j) are set with the center shifted from the check area CH1 _AB by a small distance of x, y in the X and Y directions, and each of the comparison areas CHx, y is set. Comparison pixel data DB included in (i, j)
The coincidence rate αx, y between x, y (i, j) and the reference pixel data DA (i, j) is obtained by Expression 1, and the comparison pixel data DBx, y (i, j) having the minimum value is calculated. , The pixel data of the image moved from the check area CH1 _AB .

In order to obtain the coincidence rate αx, y by the equation 1, the comparison area CHx, y (i, j) is set in the check area CH1.
This is a size including the same L * M pixel data as _AB , and in this embodiment, 121 comparison areas CHx, y (i, j) centered on the intersection of x and y shown in FIG. Is set. The directions of x and y correspond to the X and Y directions, and the pitch between the respective intersection positions is 5 pixels (5 dots). The resolution of Handy Scanner 1 is 8do
If t / mm is set, the comparison area CHx, y (i, j) is set in a range of 25 dots of 5 dots * 5 vertically and horizontally around the check area CH1 _AB.
m can be detected.

As shown in FIG. 3, the check area CH1
_The 121 comparison areas CHx, y (i, j) are sequentially moved away from the center of _AB for the comparison areas CHx, y (i, j).
comparison pixel data DBx, y included in y (i, j)
The coincidence rate αx, y between (i, j) and the reference pixel data DA (i, j) is obtained by Expression 1.

As a result, for example, in the comparison area CHx ₁ , y ₁ (i, j) indicated by the broken line where x = x ₁ and y = y ₁ , the coincidence rate αx ₁ , y ₁ becomes the minimum value. Then
At this time, it can be estimated that the image of the check area CH1 _AB is displaced from the comparison area CHx ₁ , y ₁ (i, j). Accordingly, the comparison area CHx ₁ , y ₁ (i, j) is set as the approximate area CH _{1 ′} , and the deviation amount x ₁ , y ₁ of the approximate area CH 1 ′ with respect to the check area CH 1 _AB is set to the edit instruction value x ₁ for matching the two. , Y ₁ . Further, the match rate .alpha.x _1, y _1, which resulted in this minimum coincidence ratio check area CH1 _AB alpha (1).

When the editing instruction values x ₁ and y ₁ are obtained for the check area CH 1 _AB , the editing-side two-dimensional image data B is edited based on the editing instruction values x ₁ and y _1. All two-dimensional image data C and D connected to the editing-side two-dimensional image data B are edited. Hereinafter, this editing operation will be described with reference to FIGS.

[0038] in the Y direction editing is performed on the basis of the editing instruction value y _1. If y ₁ is a positive value, the approximate area CH1
′ Is moved by y ₁ in the −Y direction to check area CH 1
_{On the} same line as _AB . For this purpose, a line n corresponding to the movement amount of y ₁ is added as shown in FIG. Since the check area CH1 _AB is the first check area CH (K) from the upper end of the overlapping portion O _AB , the line n to be supplemented is in the Y direction in the editing-side two-dimensional image data B from the upper end to the approximate area CH1 ′. At equal pitch. The pixel data of the supplemented line n is the same pixel data as one of the upper and lower pixel data so that the supplemented line n does not form a unique image. As a result of replenishing line n,
Approximate area editing side two-dimensional image data B of the lower from CH1' is y ₁ moves in the -Y direction (in the drawing below).

[0039] in the X direction editing is performed on the basis of the editing instruction value x _1. If x ₁ is a positive value, approximate area CH1
'The by x ₁ moves in the -X direction (left in the figure), to match the check area CH1 _AB. For this reason, as shown in FIG. 4, the entire editing-side two-dimensional image data B is x _{1 in the} −X direction.
Move.

As shown in FIG. 5A, the above-described editing operation using the editing instruction values x ₁ and y ₁ is performed simultaneously on all the two-dimensional image data C and D connected to the editing-side two-dimensional image data B. Do. That is, the two-dimensional image data C to a horizontal position equal to the approximate area CH1' from the upper end, supplemented with line n to D, the two-dimensional image data C, is x ₁ move the entire D in the -X direction.

As described above, the other two-dimensional image data C and D
Are simultaneously edited so that the relative position between the editing-side two-dimensional image data B and the unedited two-dimensional image data C and D is maintained in a read state.

[0042] As a result of the editing work on the check area CH1 _AB, overlapping parts from the top to check area CH1 _AB is synthesized not shift position.

Next, the same processing as described above is performed for the next check area CH2 _AB . Similarly, the reference pixel data DA (i, j) of the reference-side two-dimensional image data A included in the check area CH2 _AB and the editing-side data included in the comparison area CHx, y (i, j) set therearound. The coincidence rate αx, y of the two-dimensional image data B with the comparison pixel data DBx, y (i, j) is obtained by Expression 1.

As shown in FIG. 6, in the comparison area (approximate area CH2 ') indicated by the broken line, if the coincidence rates αx ₂ and y ₂ have become the minimum values, the approximate area CH2
The position relative to the check area CH2 _AB of 'edit instruction value -x ₂ for check area CH2 _AB, -y ₂ is obtained. Further, the match rate .alpha.x _2, y ₂ at this time as the matching rate of the check area CH2 _AB α (2), used in the editing X direction to be described later.

The editing instruction value -y ₂ is, because it is a negative value,
As shown in FIG. 5B, the approximate area CH2 ′ is moved by y ₂ in the + Y direction (upward in the figure) to be on the same line as the check area CH2 _AB . To this end, as shown in FIG y ₂
The line n 'for the moving amount is deleted. Check area CH
Since 2 _AB is the second check area CH (K) from the upper end of the overlapping portion O _AB , as shown in FIG. 4, the line n ′ to be deleted corresponds to the immediately preceding check area CH (K), ie, 1
From the second check area CH (1) to the approximate area CH2
Are extracted at equal pitches from the editing-side two-dimensional image data B up to '. By deleting the line n 'evenly (at equal intervals), the read image does not significantly change. As shown in the figure, the line correction in the Y direction is also performed on the other two-dimensional image data C and D connected to the editing-side two-dimensional image data B, similarly to the above-described editing in the check area CH (1). , Thereby approximating area CH
All of the editing-side two-dimensional image data B, C, and D below the line _{2 '} move y2 in the + Y direction.

Editing of the approximate area CH2 'in the X direction is performed by setting the coincidence rate α (2) of the check area CH (2) to
Check area CH (K) coincidence rate α obtained so far
It is determined by comparing with the minimum matching rate α _MIN (K) in (K). In this case, since the minimum matching rate α _MIN (K) is the matching rate α (1) of the check area CH (1), the matching rate α
Compared with (1), the matching rate α (2) is equal to the matching rate α (1)
If the value is larger than the above value, the editing in the X direction is not performed. That is, the editing in the X direction performed on the check area CH (K) related to the minimum matching rate α _MIN (K) is such that the entire editing-side two-dimensional image data B matches the reference-side two-dimensional image data A most in the X direction. It is presumed that the editing was made to move to the position to be edited.

The result of the editing operation for the check area CH2 _AB, is synthesized without displacement further position overlapping portion O _AB is from check area CH1 _AB to check area CH2 _AB.

The same processing is performed in the third check area C
This is performed for H3 _AB , and the matching rate α (3) of the check area CH3 _AB and the editing instruction values −x ₃ , y ₃ are obtained.

Since the editing instruction value y ₃ is a positive value, as shown in FIG. 5C, the approximate area CH _{3 ′ is} shifted in the −Y direction by y _3.
Move to match the check area CH3 _AB on the same line. For this reason, as in the case of the check area CH1 _AB , the immediately preceding check area CH (K), that is, the editing-side two-dimensional image data B from the second check area CH (2) to the approximate area CH3 ′ and the Lines n of the moving amount of y _{3 are} added to the continuous two-dimensional image data C and D at the same pitch in the Y direction. Results supplemented with line n, edit side two-dimensional image data B downward from the approximate area CH3' is y ₃ moves in the -Y direction.

Subsequently, the coincidence rate α (3) of the check area CH3 _AB is calculated using the check area CH determined so far.
Minimum matching rate α _MIN (K) among matching rates α (K) of (K)
Then, it is determined whether or not the approximate area CH3 'is edited in the X direction.

Since the minimum matching rate α _MIN (K) is α (1), the matching rate α (3) is compared with α (1), and the matching rate α
If (3) is smaller than α (1), the matching rate α
Editing in the X direction is performed with (3) as the new minimum matching rate α _MIN (K).

Since the editing instruction value in the X direction is a negative value −x ₃ , the editing-side two-dimensional image data B and the entire two-dimensional image data C and D connected thereto are moved by x ₃ in the X direction.

When the editing-side two-dimensional image data B is edited for all the check areas CH (K) set in the overlapping portion O _AB in this way, the reference two-dimensional image data A of the overlapping portion O _AB is Combined without displacement.

Thereafter, when two-dimensional image data is further continued on the editing-side two-dimensional image data B side as in the present embodiment, the editing-side two-dimensional image data B is replaced with a new reference side two-dimensional image data. The processing described above is repeated with the two-dimensional image data B and the two-dimensional image data C connected to the two-dimensional image data B on the editing side via the overlapping portion _OBC as new two-dimensional image data C on the editing side. That is, the overlapping portion O _BC, Y direction substantially equidistant check area CH1 _BC A on the same axis in the Y direction, CH
2 _BC and CH3 _BC are set, and each check area C is set.
Similarly for H (K), a number of comparison areas CHx, y
(I, j) are set, and the matching rate α (K) and the edit instruction value Δ
x and Δy are obtained. However, the number of check areas CH (K) to be set is arbitrary for each overlapping portion.
It is not necessary to set on the same line as the check area CH (K) set in another overlapping portion.

The first check area CH1 _BC from the upper end
Assuming that the coincidence rate is α (1) and the edit instruction values x ₄ and −y ₄ , as shown in FIG. 5D, based on the edit instruction value −y ₄ , the approximate area CH1 ′ from the upper end is obtained. remove from the two-dimensional image data D contiguous edit side and the two-dimensional image data C to up line n'at an equal pitch, y ₄ moves the image data of the downward Y-direction than the approximate area CH1'. Further, based on the editing instruction value x _4, to move the entire two-dimensional image data D contiguous thereto and editing side two-dimensional image data C in the -X direction. The coincidence rate α (1) of the check area CH1 _BC is the minimum coincidence rate α _MIN (K) at the overlapping portion O _BC
Check area CH (K) required thereafter
Is compared with the matching rate α (K).

In this manner, the same editing is performed for all the two-dimensional image data connected in a plurality of overlapping portions, and two-dimensional image data in which there is no displacement of the image in the overlapping portions can be obtained.

In the present embodiment, the left-most two-dimensional image data A is first used as reference two-dimensional image data among a plurality of two-dimensional image data, and the reference two-dimensional image data is sequentially shifted to the right. Although the editing-side two-dimensional image data on the right side is edited (referred to as an editing-side preceding right-direction editing method), the method is not necessarily limited to this method. As the side two-dimensional image data, the reference two-dimensional image data may be sequentially shifted leftward to edit the editing two-dimensional image data on the left side (referred to as an editing-side preceding type leftward editing method).

As described above, when these editing methods are used for a large number of two-dimensional image data, first, one end is set to the reference-side two-dimensional image data (for example, A). Editing of the two-dimensional image data (for example, D) is repeated as many times as the number of two-dimensional image data connected to the reference two-dimensional image data, and the identity of the two-dimensional image data when read may be lost.

Therefore, the intermediate two-dimensional image data (for example, B) is used as the reference-side two-dimensional image data to be edited first, and the two-dimensional image data A continuous to the left side of the two-dimensional image data A is the left side of the editing-side preceding type. Editing is performed by the direction editing method, and the two-dimensional image data C and D that are continuous on the right side are edited by the editing-side preceding type right-direction editing method (referred to as editing-side preceding type two-way editing method), and all the continuous Two-dimensional image data may be combined. According to the editing-side preceding type bidirectional editing method, the number of times of editing of the two-dimensional image data at both ends can be reduced, and the identity of the images can be maintained.

In any of these methods,
The reference-side two-dimensional image data set after repeating editing in a plurality of overlapping portions has already been edited in a large number, and does not accurately reproduce the image immediately after reading. Therefore, the editing-side two-dimensional image data edited so as to match the reference-side two-dimensional image data without any positional deviation further includes an error due to the previous editing.

FIG. 7 shows a different embodiment of the present invention which solves this problem. According to this embodiment,
Even if editing in a plurality of overlapping portions is repeated, the editing-side two-dimensional image data set thereafter can always be edited with reference to the reference-side two-dimensional image data composed of unedited two-dimensional image data.

As shown in the figure, if, for example, four two-dimensional image data are to be synthesized, first, the two-dimensional image data E on the left end is combined with the two-dimensional image data on the editing side and the two-dimensional image Using the data F as the reference-side two-dimensional image data, the editing-side two-dimensional image data is edited, and the two are combined without displacement. Editing method is to check area C
This is the same as the editing method of the description for setting H (K). For example, assuming that the editing instruction values for the first check area CH1 _EF from the upper end are −x ₅ and y ₅ , FIG.
(A), the based on the editing instruction value y _5, added line n at an equal pitch from the editing side two-dimensional image data E to the approximate area CH1' from the upper end, the approximate area CH1
'Y ₅ moves the image data of the lower downwardly from. Further, x ₅ is moved to the right across editing side two-dimensional image data E.

The same editing is performed for all the check areas CH (K) set in the overlapping portion O _EF , and the two-dimensional image data E and F are synthesized as shown in FIG. 7B.

Subsequently, the two-dimensional image data F, which was the reference-side two-dimensional image data, is used as the editing-side two-dimensional image data, and the two-dimensional image data G adjacent to the right is used as the new reference-side two-dimensional image data. The editing-side two-dimensional image data F is edited. Editing of the editing-side two-dimensional image data F is also performed on all the two-dimensional image data connected to the editing-side two-dimensional image data F at the same time. That is, as shown in FIG. 9C, if the editing instruction value is x ₆ , −y ₆ and the editing-side two-dimensional image data F is to be edited, the editing instruction value −y ₆ is used as the editing instruction value. remove the line n'from dimensional image data E and F, based on the entire F and editing side two-dimensional image data E to the edit instruction value x _6, it is moved to the left.

Thus, the two-dimensional image data F and G
After that, the editing is repeated with the two-dimensional image data H connected to the right of the synthesized two-dimensional image data G as new reference-side two-dimensional image data as shown in FIG.

In the second embodiment, the left-most two-dimensional image data E is used as editing-side two-dimensional image data to be edited first among a plurality of two-dimensional image data, and the editing-side two-dimensional image data is sequentially read. The right-sided two-dimensional image data is shifted to the right to match the reference-side two-dimensional image data on the right side (referred to as a reference-side preceding type rightward editing method). The side two-dimensional image data may be one in which the editing side two-dimensional image data is sequentially shifted to the left to match the reference side two-dimensional image data on the left side (referred to as a reference-side preceding type leftward editing method). Similarly to the above-mentioned editing-side preceding type editing method, first, the reference two-dimensional image data in the middle is used as the editing-side two-dimensional image data, and the reference-side preceding type rightward editing method and the reference-side preceding type leftward editing method are used. They may be used in combination.

In the above embodiment, the normalization process for matching the lengths in the Y direction of a plurality of read two-dimensional image data was performed before editing. There is no need to combine the two-dimensional image data.

Although the first and second embodiments have been described using the same synthesis method, in the second embodiment,
Other methods may be used as long as two-dimensional image data can be combined without positional displacement.

[0069]

As described above, according to the present invention, respective overlapping portions of a plurality of two-dimensional image data can be automatically synthesized without positional displacement. Therefore, a troublesome combining operation is eliminated, and even a minute displacement can be easily corrected.

According to the second aspect of the present invention, the editing-side two-dimensional image data edited in the previous check area CH (K) is compared with the reference-side two-dimensional image data, so that the positional deviation is smaller. It has become. Therefore, the comparison pixel data DBx, y (i, j) and the reference pixel data DA (i,
The difference between the pixel data of j) becomes a small value, and the coincidence rate α
Calculation of x and y becomes easy.

According to the third aspect of the present invention, the editing-side two-dimensional image data and the continuous two-dimensional image data are simultaneously edited via another overlapping portion. There is no enlargement by editing, and the calculation of the coincidence rate αx, y does not become complicated.

According to the fourth aspect of the present invention, the editing-side two-dimensional image data can always be compared with the unedited reference-side image data, so that the overlapping portion can be synthesized in a state similar to the subject.

[0073]

[Brief description of the drawings]

FIG. 1 shows continuous two-dimensional image data A, B, and C at an overlapping portion.
FIG.

[Figure 2] overlap O _AB check area CH1 reference pixel data DA of a portion represented by _AB (i, j) and an explanatory view showing a comparison pixel data DB (i, j) of the same portion.

FIG. 3 is an explanatory diagram showing a relationship between a check area CH1 _AB and a set comparison area CHx, y (i, j).

FIG. 4 is an explanatory diagram showing an editing operation of editing-side two-dimensional image data B for a check area CH (K).

FIG. 5 (a) in an editing-side preceding type rightward editing method
Edits the check area CH1 _AB ,
(B) is an explanatory diagram for explaining an editing operation for the check area CH2 _AB , (c) is an editing operation for the check area CH3 _AB , and (d) is an explanatory diagram for explaining an editing operation for the check area CH1 _BC . is there.

FIG. 6 is an explanatory diagram showing a relationship between a check area CH2 _AB and a set comparison area CHx, y (i, j).

FIG. 7 (a) in the reference-side preceding type rightward editing method
Edits the check area CH1 _EF ,
(B) shows the editing work synthesized in the overlapping part O _EF , (c)
Edits the check area CH1 _FG ,
(D) is an explanatory view for explaining the editing work synthesized in the overlapping portion _OFG .

FIG. 8 is a plan view of the handy scanner with a cover removed.

FIG. 9 is a plan view showing a method of reading a subject wider than a reading width w by a handy scanner.

FIG. 10 is an explanatory diagram showing a state in which a plurality of read two-dimensional image data are continuous.

[Explanation of symbols]

1 Handy Scanner 5 Subject CH Check Area CH (K) Kth Check Area D (i, j) Reference Pixel Data CHx, y Comparison Area Dx, y (i, j) Comparison Pixel Data αx, y Matching Rate CH (K ) 'Approximate area Δx, Δy Edit instruction value α _MIN (K) Minimum match rate value

Claims (4)

[Claims] 1. A handy scanner (1) is connected to a subject (5).
The main scanning of one line in the X direction is performed, and the handy scanner (1) is manually sub-scanned in the Y direction.
An image reading operation for generating two-dimensional image data having a sub-scanning width in the X-direction and an X-direction moving operation for moving the handy scanner (1) in the X-direction within at least the reading width of one line are repeated in adjacent X-directions. (B) generating at least two or more two-dimensional image data in which a part of the two-dimensional image data partially overlaps and editing the two-dimensional image data so that the image data of the overlapping part is matched; A plurality of check areas CH composed of pixel data in the i-th row and j-th column are set in the Y direction with reference to the reference-side two-dimensional image data at the overlapping portion of the two-dimensional image data and the adjacent editing-side two-dimensional image data, B) extracting pixel data included in the K-th check area CH (K) as reference pixel data D (i, j) from the reference-side two-dimensional image data; In the center of the check area CH (K),
A predetermined number of comparison areas CHx, y at equal pitches in the X and Y directions
(D) From the editing side two-dimensional image data, each comparison area CH
The pixel data included in x, y is compared with comparison pixel data Dx, y
(E) The coincidence rate αx, y is calculated from the sum of the differences between the corresponding pixel data of the comparative pixel data Dx, y (i, j) and the corresponding pixel data D (i, j). (F) The comparison areas CHΔx, Δy in which the coincidence rate αx, y has the minimum coincidence rate α (K) are set as the approximate area CH (K) ′, and the approximate area CH (K) ′ is set as the check area CH.
The editing instruction values Δx and Δy to be moved to (K) are obtained. (G) The processing from (B) to (F) is repeated for all the check areas CH to obtain the editing instruction values Δx and Δy for each check area CH. (H) Approximate area CH based on edit instruction values Δx, Δy
', The edit-side two-dimensional image data including' is matched with the corresponding check area CH and edited, and the reference-side two-dimensional image data and the edit-side two-dimensional image data are synthesized. . 2. A handy scanner (1) is connected to a subject (5).
2D image data consisting of a scanning width of one line and a sub-scanning width in the Y direction is generated by main scanning of one line in the X direction and manual sub-scanning of the handy scanner (1) in the Y direction. An image reading operation to be performed and an X direction moving operation for moving the handy scanner (1) in the X direction within a reading width of at least one line are repeated, and at least two or more two-dimensional images in which a part of adjacent X directions overlap. In the method of synthesizing two-dimensional image data, which generates two-dimensional image data so as to generate the data and match the image data of the overlapping portion, (i) editing two-dimensional image data adjacent to the reference two-dimensional image data; A plurality of check areas CH composed of pixel data in the i-th row and the j-th column are set in the Y direction with reference to the reference-side two-dimensional image data. , The pixel data included in the K-th check area CH (K) is extracted as reference pixel data D (i, j), and (x) the pitch is equal in the X and Y directions around the check area CH (K). Is used to set a predetermined number of comparison areas CHx, y. (ヲ) From the two-dimensional image data on the editing side, each comparison area CH
The pixel data included in x, y is compared with comparison pixel data Dx, y
(I) The coincidence rate αx, y is calculated from the sum of the differences between the corresponding pixel data of the comparative pixel data Dx, y (i, j) and the corresponding pixel data D (i, j). (F) The comparison areas CHΔx, Δy in which the coincidence rate αx, y has the minimum coincidence rate α (K) are set as the approximate area CH (K) ′, and the approximate area CH (K) ′ is set as the check area CH.
The edit instruction values Δx and Δy to be moved to (K) are obtained, and (Y) the check area CH (K) is 1 from the upper end of the overlapping portion.
In the case of the check area CH (1), the number of lines from the upper end of the two-dimensional image data on the editing side to the approximate area CH (1) ′ is corrected according to the editing instruction value Δy to check area CH (1). And the approximate area CH (1) ′ are made coincident on the same line, and the entire edit-side two-dimensional image data is moved in the X direction according to the edit instruction value Δx for editing. K), the two-dimensional image data edited in the immediately preceding check area CH (K-1) is regarded as new two-dimensional image data on the editing side, and the coincidence rate α is determined for each check area CH (K).
(K) and edit instruction values Δx, Δy are obtained. (F) According to the edit instruction value Δy, the number of lines from the immediately preceding check area CH (K-1) to the approximate area CH (K) ′ is corrected and checked. When the coincidence rate α (K) of the area CH (K) is the minimum value α _MIN (K) of all the check areas CH (K), the entire editing-side two-dimensional image data is converted to X according to the editing instruction value Δx. Moving in the direction, and repeating (S) processing to edit the editing-side two-dimensional image data for all the check areas CH (K), and combine the reference-side two-dimensional image data and the editing-side two-dimensional image data. A two-dimensional image data combining method. (3) When editing the editing-side two-dimensional image data overlapping at the overlapping portion, the editing-side two-dimensional image data connected to the editing-side two-dimensional image data via another overlapping portion in the direction of the editing-side two-dimensional image data. After editing together with the two-dimensional image data, (d) combining the reference two-dimensional image data and the two-dimensional image data on the editing side, the two-dimensional image data on the editing side is made new two-dimensional image data on the reference side. (2) is repeated by using the continuous two-dimensional image data as new two-dimensional image data on the editing side through the section, and (n) all the two-dimensional image data in which three or more two-dimensional image data are continuous in two or more overlapping parts. The method for synthesizing two-dimensional image data according to claim 1, wherein the two-dimensional image data is synthesized. (4) When editing the editing-side two-dimensional image data overlapping in the overlapping portion, all the two-dimensional image data connected in the direction of the editing-side two-dimensional image data are edited together with the editing-side two-dimensional image data. (M) After combining the reference-side two-dimensional image data and the edit-side two-dimensional image data, the reference-side two-dimensional image data is used as new edit-side two-dimensional image data, and the other two-dimensional image data is moved in the direction of the reference-side two-dimensional image data. The process of (a) is repeated using the two-dimensional image data connected via the overlapping portion as new reference-side two-dimensional image data, and (c) all of the three or more continuous two-dimensional image data at the two or more overlapping portions The method for synthesizing two-dimensional image data according to claim 1, wherein the two-dimensional image data is synthesized.