JPS61128389A - Linear approximating and encoding device of linear graphic - Google Patents

Abstract

PURPOSE:To attain real time processing for the linear approximation and encoding of a linear graphic by rotating the relative coordinates between the older two points and between newer two points out of three points in a time series coordinate point string of the linear graphic, and if necessary, executing interpolation processing or the like and reading out the opening angles of three points while referring a table. CONSTITUTION:Three points in the time series coordinate point string of the linear raphic through a data input part 2 are selected by a three-point selecting part 3 through a pen or the like, the relative coordinates between older two points and between newer two points are determined and the relative coordinates are rotated so that the opening angle between the relative coordinates is turned in a prescribed direction by a relative coordinate conversion part 4. When the relative coordinates are threshold or more, the relative coordinates are interpolated by a virtual coordinate interporation part 5 so that the length of an arm formed by the relative coordinates is an upper limit value or less. The table of an angle calculation part 6 is refereed by the interpolated relation coordinates to determine the opening angle between the relative coordinates. The start and end points of linear approximation are encoded by a final sampling point output part 7. Thus, the linear approximation and encoding can be processed with real time even if a microprocessor having no floating point calculating circuit is used.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention relates to a linear approximation/encoding device for line figures.

Conventional technology Conventionally, there are methods for approximating and encoding input line figures, such as those that approximate and encode input line figures only with straight lines, and those that use straight lines and circular arcs (for example, Japanese Patent Laid-Open No. 58 -1897
Publication No. 87). In either method, as a preprocess to divide a line figure into a plurality of straight lines or circular arcs, the opening angle 39 or the supplementary angle 40 of the opening angle 39 is (change in angle of point sequence) is often calculated. Then, based on these angular changes, bending points or inflection points included in the line figure are extracted, and approximated and encoded using straight lines or straight lines and circular arcs.

Problems to be Solved by the Invention In such conventional approximation/coding methods, angle calculation operations for input data point sequences are performed for all points except the start and end points using trigonometric functions, etc. This requires a huge amount of processing time. In particular, when attempting to perform the above operations using a microprocessor that does not have a floating point arithmetic circuit, real-time processing is difficult.

SUMMARY OF THE INVENTION An object of the present invention is to provide a linear approximation/encoding device for line figures that can perform real-time processing even in a microprocessor that does not have a floating-point arithmetic circuit.

Means to solve the problem Select three points from the time-series coordinate point sequence of the line figure, and
The relative coordinate values between the two older points and the newer two points are converted, and the converted relative coordinate values are converted into ! Alternatively, when the absolute value of the y component exceeds the threshold value, new low-altitude coordinates of three points are obtained by interpolating the low-altitude points, and
Compare the relationship between the relative coordinate values and opening angle between the older two points and the newer two points with a pre-stored table, find the supplementary angle of the opening angle of the three points, and sequentially accumulate them. , outputs points where the absolute value of the cumulative value exceeds a threshold value, resets the cumulative value, encodes a sequence of points, and then enables linear approximation of the line figure.

For production The effect of this technical means is as follows.

In other words, the opening angle or supplementary angle of the opening angle determined by the three input data points is stored in a table in advance for all combinations of the three points, and during online processing, the table is referenced only from the relative coordinates between the three points. This method calculates the above angle without using floating point calculations. Furthermore, by simply performing simple calculations on the above-mentioned angles at each point of the input data, linear approximation and encoding of the input data can be performed in real time.

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

FIG. 1 shows an embodiment of a linear approximation/encoding device for line figures according to the present invention. In FIG. 1, 10 is a line figure input device, which samples and outputs coordinate value data of a line figure drawn on an input surface. In this embodiment, an electronic blackboard was used as the line graphic input device. Reference numeral 1 denotes a linear approximation/encoding device for line figures according to the present invention, encoded data is stored in a file 11, and approximation results are displayed on a display device 12.

In the apparatus of the embodiment configured as described above, the operation thereof will be explained for each processing section below.

■ Data input unit 2 The data input unit 2 selectively processes only the data drawn with a pen in the writing mode out of the time-series sampling point sequence output from the electronic blackboard of this embodiment. Output to. Note that these data include coordinate values of one color and pen up/down information indicating the start and end points of a line figure.

(3) Three-point selection section 3 This processing section defines the lower limit of the arm length when calculating the supplementary angle of the opening angle of each point of the input data string (angle calculation section 6). The method for selecting two points before and after the point of interest is shown below.

In FIG. 2, the data is output from the data input section moment by moment.

An example of an input data string is shown below. For example, the supplementary angle of the opening angle of the focused point 14 (P, 3) in the input data string 13 is:
It varies depending on the length of the arm extended around PI3. That is, as shown in FIG. 3(a), three points P1°, P, 3 and P, which are simply consecutive in time without specifying the lower limit of the arm length, are
If 4 is selected, the supplementary angle 18 of the opening angle will not be appropriate because it will be strongly influenced by quantization errors due to microscopic processing.

On the other hand, if the lower limit of the arm length is set too large, problems may occur. For example, if P2''25 is selected from outside the dashed-dotted line frame 17 shown in FIG.
13. If the geometric shape of the line figure changes between P26, the supplementary angle 19 of the opening angle cannot be obtained correctly.

For these, the solid line frame 16 shown in Fig. 82 indicates both end points of the arm.
Alternatively, if the selection is made from outside the dotted line frame 16,
l, (supplementary angle 20. of the reasonable opening angle as shown in cl).
21 is obtained.

As described above, the regulation regarding the arm length when calculating the angle must normally be determined in consideration of the resolution of the input device, the geometric complexity of the input line figure, etc. In this example, the resolution of the electronic blackboard that is the input device is 1.8 mm/line, and the input line figures are often characters with geometrically complex shapes. Two points outside the solid line frame 16 shown in FIG. 2 and temporally closest to the point of interest 14 are selected from before and after. That is, for point of interest 14 (P13), Pl. , p13. p1
Three points of 6 are output. Note that for the 8-neighbor components of the start and end points and the start and end points, since there are no points outside the 8-neighbor components either before or after, only information to that effect is output.

■ Relative coordinate conversion unit 4 In this processing unit, if the relative coordinates between the two oldest points in time series among the three points output from the three point selection unit do not satisfy the predetermined conditions, Perform coordinate transformation.

The coordinates of the three points Ptn-4°Ptyu and ζ1 outputted from the three-point selection section are respectively (xt.1.y, nio.

(Xtn* 71.)e "tn+1 "",), the relative coordinates between "n-1"'n" and "n'"nc (ho, Δy0> and (ΔXH, Δ7N) are (ΔXo lΔ7o) =("tn"t,,+7t,-7
t, -1) (Δ"NlΔyN'="tn+1"H9
y"n+,"n'. In this processing section, relative coordinates (Δ
-1Δy)) force ζΔx0)o kaka. ≧0 If the condition is not satisfied, the following processing is performed to convert the converted relative coordinates iJ<, J and above, and (hship,
47 meals).

1)Δ! If 0<O and Δy0≦0, Δ=−Δ−
, Δ%=−Δy0 to J=−! N' instigation, = -fyNif)ho
≧Q and more. If 0 then h = -dyo,
Δradig = Δx0 Δtransport = -VH, Q = &N 1fi) "o≦0,> OO if lx≦
Figure 4 shows the exchange in three cases where Δy, 6 = -4xN or more. Figure 4 (a
t, (bl, tel are above + respectively),
n).

111), which is converted as shown in FIG. 4(d) by the above processing. Even in the case of misalignment, the coordinate axes are simply rotated around the origin, so the supplementary angle 22 of the opening angle does not change and is 0.As described above, the three points P1. ．．
+ Ptn' Pt)+1 is ・Δ! 0> O and d
y0≧0 is satisfied.

■ Low altitude coordinate interpolation unit 6 This processing unit defines the upper limit of the arm length when calculating the supplementary angle of the opening angle of each point of the input data string, and the angle formed by the three points output by this processing unit is calculated by the angle calculating section 6.

The upper limit of arm length is defined as follows. That is, the three points "n-1" and "n" output from the relative coordinate conversion section 4
For Ptn+1, Δx0≦N, Δy0≦N, l
It is assumed that Δ~1≦N and 1ON+≦N are satisfied. however,
N is a positive constant.

However, when a line figure is drawn on a blackboard at high speed, the interval between each input data becomes large, and the above conditions may not be met. In such a case, if two points in the sky are interpolated between the three points, a data string that satisfies the above conditions will be generated from K.

As an example, FIG. 6 shows a point sequence Pt that does not satisfy the above conditions! m-1'"n' Ptn+1 "tn+2 is shown.

Note that in this example, N==a. As shown in the figure, 3
When paying attention to the point ~-1tPt, lζ+1, ζya exists outside the upper limit frame 23, and vice versa.

ζ+1. When focusing on ζ+2, ζ exists outside the upper limit frame 24. In such a case, in order to satisfy the above condition, a low-altitude 2
Interpolate the points Qt,, O,, and create a new point sequence “n−t
'``n'Qtn'Qtne1'Ptn,'''
Generate m2. By performing such interpolation, 3
Points Pt, , P, , I Qt, and Qtn-1-1'
``n+1'' ``n+2 satisfies the above upper limit condition,
Also, the supplementary angles of the opening angles of Pt- and Ptyt are respectively pt
,,,,,p,. ,Qt,.

and Qtne ' Ptn, ' Ptn+2 is approximately given by the point sequence. Note that both Qtn and "n+1" are interpolated on the straight line connecting the two points P, n, and Ptn+1, so the three points P,..., Qt, I Qt,1 and Q,..., Qt, .

Complement of the opening angle of Qtn and Qtn+1 determined by P'
The n+1 angle can be set to zero.

This processing section uses an interpolation table 26 shown in FIG. 6 in order to quickly find the interpolated position of a low sky point. If the relative coordinates exceed the upper limit frame 26, the interpolated position can be determined by referring to the elements in the interpolation table determined by the relative coordinate values. That is, each element belonging to area ak in the interpolation table stores the address of the respective interpolation position A-. Therefore, in the example of FIG. 6, since the relative positional relationship between Ptn and ζ1 corresponds to the area shown in FIG. 6, the interpolated position H of the low air point Qtn in this case can be easily determined. Further, the interpolated position of the low-air point Qtu-1-1 is obtained by rotating the position of the low-air point Qtn by 1000 revolutions, as shown in FIG. 6, and is easily determined. Note that, as shown in FIG. 6, if the interpolation table 26 is provided only in one quarter direction with respect to Ptn, it can be referenced during rotation processing.

Furthermore, for data during high-speed drawing in which the relative coordinates between two points exceed the size of the interpolation table, the relative coordinates are divided into two equal parts repeatedly until they fit within the interpolation table. Even if such processing is performed, the influence of approximation errors on the supplementary angle of the opening angle is very small.

Note that the size of the upper limit frame and the interpolation table must be determined in consideration of the resolution of the input device and the coordinate value sampling frequency. In this example, the resolution of the electronic blackboard is 1.8 mm.
/ book and the coordinate value sampling frequency of 60 Hz, N in the above conditions is set to 6, that is, the upper limit frame is set to 13×13, and the interpolation table size is set to 4ffi×45.

■ Angle calculation unit 6 The three-point selection unit 3. Through each process in the relative coordinate conversion section 4 and the low-altitude coordinate interpolation section 6, three points "n-1"'n, I"70-1), P'n"'n') are input to this processing section.
and ζ4'1 "tn+'1" Relative coordinate between "n+1"(Δx0.Δ") = "tn-Itn-1' 71
．． -7t, -1) (Δ"N"N) = "tn+1-"n
・ytll+1−ytn) satisfies the following conditions.

1) 1Δ! 01>1 or 1Δyol>11h made 1>1 or 1 fact1>1 11) Δχ. 〉0 and Δy0≧0111)
Δx0≦N and Δy0≦N1&NI≦N and 1N1≦N N: Positive constant This processing section obtains by referring to a three-point Pt knee 1° penalty table that satisfies the above conditions. In this processing unit, the supplementary angle of the opening angle is called a penalty for linearity in the sense of deviation from the straight line, and the penalty table contains each penalty for all three-point combinations whose relative coordinates satisfy the above conditions. −
The value is pre-stored. The penalty table is (
Δ-1#. )Yo.

It consists of a small part called a partial penalty table.

The three points Ptn-/I'Ptn' shown in Figure 7 below are
Each relative coordinate between n+1 (Δ”0 * '70) + (Δ
The procedure for finding the penalty -27 determined from three points is shown below. It is assumed that each relative coordinate satisfies the above conditions. However, N=6.

FIG. 8 shows the penalty table 28.

Each grid in the table represents each partial penalty table determined by the relative coordinates (Δx0.Δy0) when ζ1 is assumed to be at the position shown in FIG.

Therefore, in this example (ΔI0.Δy0) = (4, 2
), therefore, the partial penalty to be referred to is table 2.
Find 9. Partial penalty in Figure 9 - Table 29
shows. Each element of the partial penalty table 29 includes:
The deviation from the straight line connecting two points ζA and P, that is, the penalty, is stored, and each penalty value is calculated by rotating clockwise and counterclockwise around point Ptn, as shown in the figure. has a negative sign. Therefore, the partial penalty −
In table 29, the penalty is determined depending on which element the point Pt,1 corresponds to. In this example (Δ
Since ``Hj'7N) = (5, 1), the penalty for which three points is -16, as shown in Figure 9.
It can be seen that it is.

As described above (this processing section refers to the penalty table from the relative coordinates of the three input points, and calculates the supplementary angle of the opening angle,
The part penalty is calculated quickly.

Note that the lower limit frame 15 shown in FIGS. 8 and 9 is defined by the three-point selection section 3, and the upper limit frame defined by the low-altitude coordinate interpolation section 6 is based on the penalty of this processing section. - reflected in table size.

(2) Final sampling point output unit T This processing unit determines the final sampling point based on the penalty value of each point calculated by the angle calculation unit 6, and outputs it together with the starting and ending points.

FIG. 10 shows the process of determining the final sampling point from the input data P, ~P2°. However, the starting point 30 and the ending point 34 are always sampled.

First, point p2. p3. The penalty values of p4... are accumulated in sequence, and the accumulated value reaches threshold 3 on both positive and negative sides.
Points exceeding 6 and 36 are taken as final sampling points.

In this example, since the cumulative value exceeds the positive threshold 36 at point P6, point P6 is set as the final sampling point and the cumulative value is reset. This process is sequentially repeated to obtain final sampling points 32 and 33.

Note that in addition to the threshold for the cumulative total value, this processing section also sets a threshold for each penalty at each point.

As shown in FIG. 11, this means that, for example, in a case where the cumulative value up to forelight P6 has reached just before the positive threshold 36, the next point P6 has a relatively large negative penalty. This is because if this is done, the cumulative value may not exceed the negative threshold. In this case, point P6 can be included in the final sampling points.

As described above, this processing section derives the final sampling rate including the start and end points from the J input data.

In addition, in this processing section, by changing the threshold value of the above-mentioned cumulative value, processing can be performed with any degree of approximation desired by the user.

■ Encoded data storage unit 8 This processing unit stores the coordinate values of the data output from the final sampling point output unit 7 together with color information, start and end point information, etc.
Each stroke is stored in the file 11 shown in FIG.

(2) Approximation result output unit 9 This processing unit displays the approximation result of the input data on the display device i, 12 in a specified color based on the encoded data.

Effects of the Invention As described above, according to the present invention, linear and near-biased encoding of line figures can be performed in real time even in microglossets that do not have floating-point arithmetic circuits. Therefore, it is possible to have an inexpensive system configuration, which is extremely useful in practice.

Furthermore, since approximation and encoding can be performed with any degree of approximation desired by the user, it is possible to improve the quality of the approximation results or achieve a high data compression rate.

[Brief explanation of drawings]

Figure 1 shows the linear approximation of a line figure in one embodiment of the present invention.
A block configuration diagram of the encoding device. FIG. 2 is a diagram showing the means for selecting three points from the input data point sequence. The FAS diagram is a diagram showing the supplementary angle of the opening angle determined by the three selected points. FIG. Figure 6 is an explanatory diagram of operations related to conversion of relative coordinates. Figure 6 is an explanatory diagram of operations when interpolating virtual empty coordinates. Figure 6 is a diagram showing an interpolation table in which positions for interpolating virtual empty coordinates are stored in advance. Figure 7 The figure shows the relative coordinates between the three selected points, Figure 8 shows the penalty table that stores the supplementary angle of the opening angle, that is, the penalty, and Figure 9 shows the partial penalty table. Figures 10 and 11 are diagrams explaining the operation when determining the final sampling point from the input data point sequence. Figure 12 is the diagram explaining the operation when determining the penalty determined from the three points. It is a figure which defines an angle. 1... Linear approximation/encoding device, 2...
Data input section, 3...3 point selection section, 4...
...Relative coordinate conversion section, 6...Low coordinate interpolation section,
6...Angle calculation unit, γ...Final sampling point output unit, 8...Encoded data storage unit, 9...
. . . Approximation result output section, 10 . . . Line figure input device. Name of agent: Patent attorney Toshio Nakao and one other name
1 Figure 2 Figure 3 (α) Connection l ' Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 110 37 Seventh punishment,

Claims (1)

[Claims] a data input unit that obtains a chronological coordinate point sequence of a line figure; a 3-point selection unit that selects three points from the coordinate point sequence; and a data input unit that selects three points from the coordinate point sequence; a relative coordinate conversion section that performs coordinate conversion of each relative coordinate value between two points on the other hand by rotation processing; a virtual space coordinate interpolation unit that outputs three new points by interpolating virtual space points, and a chronologically older two points outputted from the virtual space coordinate interpolation unit; an angle calculation unit that calculates a supplementary angle of the opening angle between the three points by referring to a table in which values are stored in advance from each relative coordinate value between the two points and the newer one; a final sampling point output unit that sequentially accumulates the supplementary angles of the output opening angle and outputs points where the absolute value of the accumulated value exceeds a threshold value, and resets the accumulated value; an encoded data storage unit that stores encoded data of all point sequences outputted from the start and end points and the final sampling point output unit in a file; and displays a result of approximating a line figure from the encoded data on a display device. A linear approximation encoding device for a line figure, comprising an approximation result output unit.