JPH0614311B2 - Coordinate correction method - Google Patents

Description

Description: BACKGROUND OF THE INVENTION Field of Application of the Invention The present invention relates to a coordinate correction method for correcting a designated position input by using a cursor or the like so as to become a reference position of an original drawing, and particularly a printed grid. The present invention relates to a coordinate correction method suitable for accurately obtaining a reference position of a drawing existing above a position.

[Prior Art] Conventionally, in a drawing recognition device or the like, a coordinate input device such as a cursor is used to input a reference position of the drawing into a computer,
It is conceivable that the coordinate of the position is input by moving the position near the reference position to be input by a human hand (hereinafter, the coordinate to be used as the reference position is input using a coordinate input means such as a cursor. The coordinates of "specified position"). However, if the cursor attachment accuracy is insufficient or the cursor magnification is insufficient, there is a problem that the designated position, which is the center of the cursor, does not always match the reference position of the drawing intended by a human.

[Object of the Invention]

The present invention has been made in view of the above circumstances, and an object of the present invention is to accurately convert the coordinates of a specified position input by a human hand using a cursor or the like to an accurate position on the intersection of the grid printed on the drawing (grid). It is to correct to the coordinates of (position).

[Outline of Invention]

Many drawings (particularly design drawings) have graphics drawn on the basis of a grid printed on the drawing, and the reference position of the drawing is also on the grid position. However, the designated position input with the cursor or the like includes an error although the lattice position, which is the reference position, is taken into consideration.

According to the present invention, a drawing in which a grid is written is read in advance by using the drawing reading means, coordinates to be used as a reference position are input by using the coordinate input means, and the coordinates are input by the coordinate input means by using the processing means. A coordinate correction method for correcting the coordinates of a specified position to one of the coordinates of the grid position of the read drawing as the coordinates of the reference position, wherein the processing means is used to surround the grid image of the read drawing. Obtain the distribution, estimate the coordinates of the grid position from the peripheral distribution, find the coordinates of the grid position closest to the coordinates of the specified position among the coordinates of the estimated grid position, and calculate the coordinates of the specified position. The coordinate of the reference position is corrected to the coordinate of the lattice position closest to the obtained distance.

〔Example〕

FIG. 1 shows an embodiment of a recognition apparatus for carrying out the coordinate correction method according to the present invention, which is an example of a drum type line figure recognition apparatus. In FIG. 1, a drawing 1 to be recognized is closely wound around a drum, and a drum 2 is rotated by a Y-direction motor 3 in a direction of an arrow. The change in the position of the drum 2 due to this rotation is detected by the Y-direction encoder 4, and its pulse signal is sent to the Y-direction up / down counter 5. The image input head 6 is composed of a lens 7, an illumination lamp 8, an optical sensor 9 and a ball nut 10 as shown in FIG. 2, and is moved left and right by a ball screw 13 directly connected to an X-direction motor 11 in FIG. The cursor 12 is fixed and interlocked with the image input head 6 for designating an arbitrary position on the drawing. The image input head 6 is
Linear scale 1 is used to detect changes in the X-direction position.
Head part 15 which is directly connected to the image input head 6
A position pulse is sent to the Y direction up / down counter 16 from. The X-direction motor 11 and the Y-direction motor 3 in FIG. 1 are driven by an X-direction motor drive circuit 18 and a Y-direction motor drive circuit 19 controlled by the control circuit 17. The control command and state of the control circuit 17 are reflected by the command status register 20. Command status register 20, Y-direction up / down counter 5, and X-direction up / down counter 1
6 is selected by the selector 22, and the data bus 23
Connected to. The microprocessor 24 includes a program memory 25 and a data memory 2 via the data bus 23.
6, the selector 22, the correction data memory 27, and the image memory 28 are made accessible. Microprocessor 24
Decodes and executes the instruction stored in the program memory 25. On the other hand, the motor control keyboard 21 is connected to the control circuit 17 in order to give the control for moving the cursor 12 to an arbitrary position on the drum 2 by the operator, and connects the X-direction motor drive circuit 18 and the Y-direction motor drive circuit 19 to each other. Control offline. At the same time, the position designated by the cursor at that time is read from the value of the X-direction up / down counter 16 and the value of the Y-direction up / down counter 5, and a command for transferring the value to the correction data storage memory 27 is command status. It is given to the register 20.

When the drawing 1 is attached to the drum 2 by a human hand to recognize the drawing, it is necessary to set the inclination of the drawing, the reading range of the drawing, and the like. For example, as shown in FIG. 3, inputtable areas D ₁ , D ₂ , D ₃ and D ₄ of the drum 2 are set, and drawing 1 is set to P ₁ ,
P _2, P _3, P _4, when the input range of figures _{O 1, O 2, O 3} , O 4 and (in FIG. 3 not shown), with respect to the drawings 1 is the drum reference D _2, D ₃ It has a slope of θ. So entering the _{O 1, O 2, O 3} , O 4 to be the reference position of the drawing in order to set the reading range of the correction data and drawings of the inclination angle as the designated position. However, the input of O ₄ is not always necessary. In order to input the coordinates of the designated positions to be the three reference positions, the center of the cursor 12 is moved to the reference position of the drawing by using the keyboard 21, and the coordinate values are designated as the designated positions of the cursor X.
It is read from the directional linear scale 14 and the Y-direction encoder 4. However, if the mechanical attachment accuracy of the cursor is insufficient, or the resolution of the cursor is low and the resolution by the human eye cannot reach the linear scale or the sample pitch of the encoder, enter the designated position of the center of the cursor and actually input it. A deviation from the reference position of the desired drawing occurs. Then, a lattice image having a size of horizontal DX and vertical DY is input to the image memory 28 by the image input head 6 centered on the designated position designated by the cursor, and the microprocessor 24 performs a process of correcting the shift. That S _21, S _22, S ₂₃ with respect to those regions _{S 11, S 12, S 13} , S 14, O 2 with respect to regions (O ₁ around the three specified positions specified by the cursor and keyboard operation, S _24, With respect to O ₃ , S ₃₁ , S ₃₂ ,
In steps S ₃₃ and S ₃₄ , the grid position is estimated from the peripheral distribution of the grid image, and the coordinates of the designated position input by the cursor are corrected to the coordinates of the grid position. FIG. 4 is an enlarged view of one lattice image of the area centered on the coordinate P (X, Y) of the designated position input by the cursor.
The + mark is shown as the image portion of the printed grid over the range of the vertical DY. In FIG. 4, the coordinates (OX, OY) of the reference position O of the drawing to be originally input on the grid
There is a positional deviation of OFFX and OFFY between the and the coordinates (X, Y) of the designated position P input by the cursor. Here, if the lattice spacing is G, In the case, the process of shifting the coordinates (X, Y) of the designated position P input by the cursor to the coordinates (OX, OY) of the reference position O of the drawing to be originally input is performed. The coordinates (OX, OY) of the reference position O to be originally input are shown in FIG. 5 by the microprocessor 24 using the lattice image input from the seat image input head 6 in FIG. 1 and stored in the image memory 28. It is obtained by performing a series of processing as described above. First, in STEP 1 of FIG. 5, the lattice image VXY
(I, J) (I = 1 to DX, J = 1 to DY; DX, DY
Calculates the peripheral distribution CX (I) in the x direction and the peripheral distribution CY (J) in the y direction from the image size in FIG. The image data VXY (I, J) is 0 when the background, 0
In other cases, the image input head 6 of FIG.
From the image memory 28, the VXY (I,
J) increments the count of each marginal distribution by 1 when it is other than 0.

That is, CX (I) = CX (I) +1 and CY (J) = CY (J) +1. When the marginal distribution of the lattice image as shown in FIG. 4 is calculated according to the above flow, the marginal distribution in the x direction is on the upper side of the image in FIG. 4, and the marginal distribution in the y direction is on the right side of the image in FIG. become. The problem with the method of calculating the marginal distribution is that the marginal distribution is obtained in parallel with the x-axis or the y-axis.
If the inclination of the image exceeds a certain level, the peripheral distribution cannot be accurately obtained. That is, in the grid image having the tilt θ in FIG. 7 and the grid spacing G and the sizes DX and DY, the size of θ is Then, the marginal distribution becomes flat, and there is no useful distribution difference enough to estimate the grid position. Tachibana DX
= 50, G = 10 Becomes However, if the drawing 1 of the A4 plate is mounted on the drum 2 in FIG. 1 with an inclination of 11 °, the positional deviation amount d due to the tilt becomes about 57 mm in the longitudinal direction of the A4 plate. Absent. If you wear it with normal human hands so that there is no inclination in the drawing, 2
It is experimentally known that the temperature will be below °. Therefore, it is possible to estimate the grid position from the peripheral distribution of the grid image.

Next, in STEP 2 of FIG. 5, the obtained marginal distribution CX is obtained.
The sums of (I) and CY (J) are independently obtained to obtain average values THX and THY. Then, STEP 3 of FIG.
In STEP 4 and STEP 4, elements having a value less than or equal to the average value of CX (I) and CY (J) are set to 0. Assuming that FIG. 8 (a) is the peripheral distribution in the x direction in FIG. 4, this process results in zero components below THX as shown in FIG. 8 (b). Therefore, the grid position is considered to exist in each continuous element other than 0. Therefore, the STE of FIG.
At P5, as shown in FIG. 8 (b), the position where the value of the peripheral distribution becomes a value other than 0 to a value other than 0 is SX (I), and the position EX (where the value of the peripheral distribution becomes a value other than 0 becomes 0 ( I) is obtained, the position having the maximum value of the peripheral distribution in each section [SX (I), EX (I)] is obtained, and this is used as the estimated value of the lattice position. However, when the maximum values have the same value, the intermediate position of the position having the maximum value is the estimated value of the grid position. FIG. 8 (c) shows the estimated value MX (I) of the lattice position, which is the position or intermediate position having the maximum value thus obtained. Estimated value MY of the grid position in the y direction
(I) is also obtained by the same processing (STEP in FIG. 5).
6).

In order to obtain the lattice spacing G, the spacing between the estimated values of adjacent lattice positions is first calculated, and the frequency distribution corresponding to that value is obtained. That is, in STEP 7 of FIG. 5, | MX
Frequency distribution C of (I) -MX (I + 1) | to MX (I)
MX (I) is set to MY (J) in STEP 8 of FIG.
Frequency distribution CMY from -MY (J + 1) | to MY (J)
Find (J). In a normal drawing, since the lattice spacings in the x direction and the y direction are the same, the frequency distributions CMX (I) and C
The highest value among the values of MY (J) is set as the lattice spacing G (STEP 9 in FIG. 5).

Next, from the estimated values MX (I) and MY (I) of the grid positions in the input range (O ₁ , O ₂ , O _{3 in} FIG. 4) of the drawing, the grid positions XX (having the same grid spacing G I), XY (J) calculated base point positions MMX, MMY which should be base points
Ask for. Calculation base point positions MMX, MM in the present embodiment
Y is the intersection of the grids printed thicker than the other grids in the grid as shown by the coordinates (OX, OY) of the reference position O to be originally input in FIG. This position shall be determined. That is, in the input ranges O ₁ , O ₂ , O ₃ in the drawing of FIG. 4, the marginal distribution cx (I), among the estimated values MX (I), MY (I) of the lattice positions previously obtained,
The value of CY (J) is the largest (for example, the value of the marginal distribution of MX (3) is the highest in FIG. 8 (C)), and the interval between the estimated values of two adjacent grid positions is G. If there is one that satisfies the condition of the position, the base point position MMX (STEP 10 in FIG. 5), MMY (S in FIG. 5) is calculated.
TEP11). Furthermore, if there is no position satisfying such a condition, STEP in FIG.
12, in the y direction, in STEP 14 of FIG. 5, a position at which the interval between the estimated values of two adjacent grid positions is both G is obtained, and the position is calculated as the base point position MMX or MMY, respectively. Fix it. Further, when there is no position satisfying any of the above conditions, although not shown in FIG. 5, the estimated value MX (I) of the grid position having the largest values of the peripheral distributions CX (I) and CY (J), MY (I) may be obtained and the position may be corrected to the calculated base point position MMX or MMY.

Of the calculated base point positions MMX and MMY, MM in the x direction will be described below.
FIG. 9 shows a specific processing flow in STEP 10 and STEP 12 of FIG. 5 for obtaining X. First, the position MMX where the value of the marginal distribution is maximum is stored in MMX0 (STEP1 in FIG. 9). MX adjacent to MMX
The interval between (I) and MMX is obtained, and it is checked whether the value matches G (STEP 2 in FIG. 9). If both sides match G, MMX at that time is set as the calculation base point position. If this condition is not satisfied, once MMX
Is moved to the adjacent MX (I) on the right side (STEP 3 in FIG. 9), and it is similarly checked whether or not both sides match G (STEP 4 in FIG. 9). If they do not match, the MMX is moved to (I) on the right side until the boundary is reached. When the boundary is reached, the MMX is once returned to MMX0 (STEP 5 in FIG. 9), and this time the MM is moved to the adjacent MX (I) on the left side.
X is moved and the same processing is performed (STEP 6 to FIG. 9).
STEP8). In this way, when the left boundary is reached, MMX is returned to MMX0 (STEP 9 in FIG. 9). The calculation method of MMY in the y direction is performed in the same manner as MMX (STEP 11 and STEP 14 in FIG. 5).

Calculated base point positions MMX and MMY obtained by the above processing
Centered on the position of, the STE of FIG.
For P13 and the y direction, the processing of STEP15 in FIG. 5 is performed. That is, STEP 1 in FIG. 5 in the x direction
0 and STEP 12 (see FIG. 9), the calculated reference point position MMX, STEP 11 in FIG.
Then, the grid positions XX (I) and YY (J) having a constant grid interval G until reaching the boundary are obtained with the calculated base point position MMY obtained in STEP 14 as the center.

Now, in STEP 16 of FIG. 5, the coordinates (X, Y) of the designated position P input by the cursor are respectively changed to XX (I),
The coordinates of YY (J) are compared, XX (I) and YY (J) whose difference is G / 2 or less are selected, and the coordinates of the closest grid position are obtained. In the case of the present embodiment, the center of the cursor is moved by the human hand so as to be the coordinates (OX, OY) of the reference position O on the drawing, and the closest calculated grid point positions are calculated base point positions MMX, MMY. Can be expected to match.

Now, in the configuration of FIG. 1, consider the case of recognizing a drawing written with a grid in mind. First, the cursor 12 is used to input three reference points, both for setting the inclination and positional deviation of the drawing 1 attached to the drum 2 and for setting the reading range of the drawing. That is, the operator uses the keyboard 21 for controlling the motor to move the cursor so that the center of the cursor is at the center of the reference position on the drawing, and the values of the X-direction up / down counter 16 and the Y-direction up / down counter 5 at the designated positions. Is read into the data memory 26 via the selector 22 and the data bus 23. However, since it is considered that the designated position of the center of the cursor stored in the data memory 26 includes an error of OFFX and OFFY with respect to the reference position on the drawing, the size of DX and DY is centered on the designated position of the cursor. The grid drawing is read from the image input head 6 into the image input memory 28, and the coordinate correction processing according to the present invention is performed. 3 corrected by this
If the reference position of one drawing is ORG, ORGX, ORGY, ORG-ORGX is converted to x-axis, OR by coordinate conversion.
The inclination of the coordinate system with G-ORGY as the y-axis is corrected, and the coordinate values of the characteristic points such as the end points and bending points after line recognition are O.
Correction between RG-ORGX and ORG-ORGY is performed so that the grid position is estimated to be equally divided at grid intervals.

On the other hand, in addition to the hardware configuration shown in FIG. 1, the coordinate correction method according to the present invention is also effective for a line recognition device including a fax without a cursor and a keyboard for motor control and a CRT with a light pen. . That is, when the vicinity of the reference position of the drawing is displayed in the image input by fax and the reference position is designated by the light pen, a position shift occurs due to insufficient resolution of the light pen. Therefore, by correcting the reference position to the grid position estimated from the peripheral distribution of the grid image as described in the present invention, the coordinate value of the feature point after line recognition can be made accurate.

〔The invention's effect〕

As described above, according to the present invention, the difference between the designated position input using the cursor or the like and the grid position to be originally input having the interval G printed on the drawing sheet is G / 2 or less. You can automatically correct to the grid center position,
This has the effect of improving the positional accuracy of the feature points after line recognition.

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment of an apparatus for implementing the coordinate correction method according to the present invention, FIG. 2 is a specific block diagram of a part of FIG. 1, and FIG. A diagram showing the position,
FIG. 4 shows an example of the lattice image to be processed and its peripheral distribution, FIG. 5 shows the overall flow of the process for obtaining the corrected coordinate values, and FIG. 6 shows the peripheral distribution of the lattice image. Fig. 7 is a flow chart for the calculation, and Fig. 7 is a principle diagram for calculating a significant marginal distribution.
The figure is a principle diagram for estimating the grid positions from the peripheral distribution, and Fig. 9 is a flow chart for obtaining the grid positions at equal intervals. 1 ... Drawing, 2 ... Drum, 3 ... Image input head, 24 ... Microprocessor.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Shigeru Shimada 1-280 Higashi Koigokubo, Kokubunji City, Tokyo Inside Hitachi Central Research Laboratory (72) Inventor Ejiri 1-280 Higashi Koigokubo, Kokubunji, Tokyo Hitachi Ltd. (72) Inventor Osamu Abe 1450, Kamisuihonmachi, Kodaira-shi, Tokyo Inside the Musashi Factory, Hitachi, Ltd. (72) Teruhiro Watazumi 2100, Kamiimaizumi, Ebina, Kanagawa Prefecture Hitachi Seiko Co., Ltd. (72) Inventor Tsutomu Ikeuchi 6-81 Onoue-cho, Naka-ku, Yokohama-shi, Kanagawa Hitachi Software Engineering Co., Ltd. In-house (56) References JP 57-101988 (JP, A) Actual 57-10043 (JP) , U)

Claims (2)

[Claims] 1. A drawing in which a grid is described in advance is read by using a drawing reading means, coordinates to be used as a reference position are input by using a coordinate input means, and input by the coordinate input means by using a processing means. A coordinate correction method for correcting the coordinate of the specified position to one of the coordinates of the grid position of the read drawing as the coordinate of the reference position, wherein the processing means is used to convert the grid image of the read drawing. Obtain the marginal distribution, estimate the coordinates of the grid position from the marginal distribution, find the coordinate of the grid position closest to the coordinate of the specified position among the estimated coordinates of the grid position, and calculate the coordinate of the specified position. Is corrected to the coordinate of the lattice position having the shortest distance obtained as the coordinate of the reference position. 2. The claim 1, wherein when the processing means is used to estimate the coordinates of the grid position from the peripheral distribution, an average value of the peripheral distribution is calculated, and the calculated average value is calculated. A coordinate correction method characterized in that the coordinates of the position where the value of the peripheral distribution is maximum in the exceeding area are estimated as the coordinates of the grid position.