JPS60132280A - Straight line/curve data separating device - Google Patents

Abstract

PURPOSE:To increase a processing speed by converting a dot train to a vector, deriving each vector length of a vector train, and discriminating and separating the dot train to a linear or curve data in accordance with whether the vector length exceeds a parescribed threshold level or not. CONSTITUTION:A vector converting circuit 12c is provieded with a prescribed allowable range on both sides of a straight line for connecting the start point and the end point of a dot train, checks whether each point of the dot train is contained in this range or not, and sets the start point and the end points of the dot train to the start point and the end point of a vector in case all the points are within this range. On the other hand, in case a dot train which is not contained in said allowable range exists, the same processing is executed with regard to each dot train which has divided its dot train into two by the middle point, the processing is repeated until all the points of the dot train are contained in the allowable range, and in case all the points are contained in the allowable range, its allowable range is set as one vector. Subsequently, each vector length is derived by a vector length operating circuit 12d, and thereafter, when it is decided by a deciding circuit 12e that it exceeds a prescribed thershold level (t), the vector train concerned is supplied as a linear proposed example to a straight line recognizing device 14.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention relates to a straight line/curve data separation device, particularly for separating straight line data from figures containing a mixture of straight lines and curved figures.

The present invention relates to a straight line/curve data separation device that separates curve data.

Conventional technology Straight lines from shapes that are a mixture of straight and curved shapes.

A straight line/curve data recognition device that recognizes and separates each curve is well known. Conventional straight line/curve data recognition devices first recognize all line shapes as straight lines with a straight line recognition device, and then recognize them as curves with a curve recognition device. Or, conversely, a curve recognition device recognizes all line figures as curves, and then a straight line recognition device recognizes them as straight lines. However, in such conventional devices, there has been a problem that the processing load on the straight line recognition device and the curve recognition device is large, the processing speed is slow, and the efficiency is low. Therefore, in order to speed up the processing speed, a straight line /
A curve data separation device has been proposed, and when a straight line/curve data separation device is used, one straight line candidate and one curve candidate are separated in advance by the device, and then processing is performed by a straight line recognition device and a curve recognition device. , it becomes possible to speed up the processing speed.

FIG. 1 shows a schematic configuration of a straight line/curve data δ tracking device.

Note that the human input signal in the form of a line figure is photoelectrically converted in advance.

Δ/D conversion, filter, binary evolution, core line or contour extraction,
After being subjected to various processing operations such as tracking and digital arcing, it is supplied to the recognition device shown in FIG.

That is, in a normal case, a line figure is made into 16 dots/1llllll black pixels by extraction and tracing of skeleton lines or contours, and then digital arcing is performed to further thin the skeleton lines. Here, a digital arc is defined as only 8 points in the vicinity of each point outside the end points and intersections. A dot row with only black pixels. And Fig. 2 Δ.

Fig. 2B shows a dot array obtained by converting the straight line S1 curve C into a digital arc, and Fig. 2Δ.

In Figure 2B, each point other than the end points has only two black pixels in its 8 neighborhoods, and from this figure it can be seen that with digitalization, the dot array becomes a coordinate dot array. be understood. Normally, there are a plurality of dot rows in a figure, but in the apparatus shown in FIG. 1, processing is performed for each dot row, that is, for each dot row.

In Figure 1, the dot row converted into a digital arc is
The dot string data is stored in a dot string data storage device 10, and the dot string data of the memory device 10 is supplied to a straight line/curve data separation device 12 in units of dot strings.

As described above, this straight line/curve data separation device 12 is a device that preliminarily separates figures that appear to be straight lines and figures that appear to be curves from a dot sequence, and the separator 12 distinguishes a dot sequence into straight line candidates or curve candidates. The separated straight line candidates and curve candidates are supplied to a straight line δ recognition device 14 and a curve recognition device 16, respectively. The straight line candidates and the curve candidates are recognized and determined by the straight line recognition device 14 and the curve recognition device 16, respectively, and the dot rows determined to be straight lines or curves are stored in the straight line data storage device 18 and the curve data storage device 20, respectively. Supplied in a straight line.

It is stored in the storage device 18.20 as a parameter of the curve.

Note that each dot row that is not determined to be a straight line or a curved line by the recognition device 14.16 is
It is fed to a splitting device 22.2/I. The dot row dividing devices 22 and 24 divide the undeterminable dot row into two at the midpoint, and field-hack into the straight line/curve data separating device 12.

In other words, one dot row divided into two is the separation device 1.
After this separation operation is completed, another dot row divided into two is supplied to the separation device 12 and subjected to the same separation process.

As described above, according to the apparatus shown in FIG. 1, it is possible to distinguish and separate dot rows into straight lines or curved lines.

In the apparatus shown in FIG. 1, it is the straight line/curve data separator 12 that has the greatest effect on the data processing speed, and the straight line/curve data separator 12 will be described in detail below.

FIG. 3 shows the conventional straight line/curve data portion Ij [device 12
The separation device 12 includes a discriminant calculation circuit 12a, and the discriminant calculation circuit 12a calculates the distance from the straight line connecting the start point and the end point of the dot sequence to the midpoint of the dot sequence. .

That is, FIGS. 4A and 4B each show a dot array obtained by converting a straight line curve into a digital arc, and in FIGS. 4A and 4B, the starting point, middle point, and end point of the dot array are respectively designated as Ps. , Pm, Pt, and their X and y coordinates are (xs, ys) (xrn, ym>(XJ
.

The distance signal from the discriminant calculation circuit 12a is
The information is supplied to the determination circuit 12b, which compares the distance with a preset tolerance range R to determine the dot row. That is, as shown in Fig. 4Δ, h
If ≦R, then the dot array is determined to be a straight line candidate, while, as shown in FIG. 4B, if h > R, the dot array is determined to be a curved line candidate,
In this way, the dot array is separated into straight line candidates or curved line candidates by r.

Therefore, conventional straight line/curve data 11 Jl: device 12
In , the starting point Ps, the middle point Plnsm, and the point P of the dot row are
By determining the distance from Z and comparing this distance l) with the allowable range 1, the dot array is determined, and this makes it possible to separate the dot array into straight lines and 9 curves. .

However, such conventional straight line/curve data separation devices only process dots, so when this conventional straight line/curve data separation device is used in a straight line/curve data recognition device, In this case, the straight M821 recognition device and the curve recognition device that follow are also processed in units of dots, which has the problem of increasing the amount of processing and slowing down the processing speed. Furthermore, with conventional straight line/curve data separation devices, accurate separation is not always possible, and therefore, there may be cases where the subsequent straight line recognition device 1 curve recognition device cannot identify either straight line or 3 curves, and in this case, The data must be fed back to the linear/curve data separator via the dot row splitting device, and for this reason, conventional linear/curve data delta detection devices require a special loop for this purpose. There was a problem that the overall processing speed was slow. For example, as shown in Figure 4C, if the figure is a combination of straight lines and curved lines, h>R, so
The dot string is separated as a curve candidate, but the curve recognition device cannot determine that this dot string is a curve, so the straight line/curve data separation device This has the drawback of slowing down the overall processing speed of the device.

Purpose of the Invention The present invention has been made in view of the problems that existed in the previous period.
When used in a straight line/curve data recognition device, its purpose is to speed up the processing speed of the subsequent straight line/curve recognition device and to eliminate the need for the above-mentioned loop in the straight line/curve data recognition device. An object of the present invention is to provide a linear/curve data separation device that can increase the processing speed of the entire device.

Structure of the Invention The present invention converts a dot string into a vector, determines the length of each vector in the dot string, and converts -811 or the total EaS of the dot string into linear data depending on whether the vector length exceeds the dark value or not. Alternatively, it is characterized by discriminative separation into curve data.

In the following examples, we will mainly refer to attached A (Fig. 5 to 6 Δ of 1 drawing,
A preferred embodiment of the present invention will be described based on FIGS. 6B and 6C.

First, the present invention will be explained in detail.

In the present invention, a dot string is vectorized, each vector length of the vector string is determined, and the dot string is separated into straight line data or curved data depending on whether the vector length exceeds a threshold value. That is, if the vector length exceeds the threshold, the vector sequence' is determined to be straight line data and is supplied to the straight line recognition device, and if the vector length does not exceed the threshold, the vector sequence is The column is determined to be curve data and is supplied to a curve recognition device. In the conventional straight line recognition device and curve recognition device, the human signal is a dot train, and the dot train is processed dot by dot.
However, in the present invention, the straight line/curve data separation device processes on a vector-by-vector basis, and the straight line recognition device and curve 8+! i! The human input signal to the recognition device is a vector sequence, and since the dot sequence is processed in units of vectors, it has the advantage of high processing speed.

Therefore, according to the principles of the present invention, in the straight line recognition device and the curve recognition device, the dot string is processed in units of vectors, and the processing speed can be increased.

Furthermore, according to the principle of the present invention, not only the whole of one dot string is separated into straight line data or curved data, but also a part of one dot string is separated into straight line data or curved data. It is also possible to separate the tree dot row into straight and curved parts. Therefore, the straight line/curve data 11i1ffi device accurately separates part or all of one dot row into straight line data or curve data, and therefore, the straight line recognition device 1 curve recognition device separates straight line data.

Curve data can be recognized accurately, and the loop of dividing the dot row into two and feeding it back to the straight line/curve data separation device, as in conventional devices, can be eliminated.

Next, FIG. 5 shows a straight line/curve data recognition device to which the straight line/curve data separation device according to the embodiment of the present invention is applied.

In FIG. 5, a door column from a dot column data storage device 1O is supplied to a linear/curve data portion: i, 1 [device 12', and the linear/curve data separator 12' includes a vectorization circuit 12C; The vectorization circuit 12C vectorizes the dot array within a certain range on both sides to form a vector array. 6.DELTA.Ig+ and FIG. 6B each show a state in which straight line 1 curve is vectorized in a certain range Q, and straight line 2 curves are each vector sequence v1. vector sequence V
,, vectorized to V2v5.

There are several methods for vectorizing a dot string, but in this embodiment, vectorization is performed using the following processing method.

First, a certain tolerance range is set on both sides of a straight line connecting the starting point and end point of the dot array, and it is checked whether each point of the dot array is within this range. If all the points in the dot array are within this range, the start and end points of the dot array are taken as the start and end points of the vector. This state is shown in Fig. 6Δ, and in Fig. 6A, all the points in the dot array are within a certain range Q, so that the dot array is equal to the vector array v1.
It has been vectorized into . On the other hand, if there is a point in the dot array that does not fall within the permissible range, the same process is performed for each dot array obtained by dividing the dot array into two at the midpoint. This process is repeated until all points in the dot array fall within the tolerance range, and when all points fall within the tolerance range, the vector up to that tolerance range is treated as one vector, and the same goes for the remaining dot arrays. As a result, the dot string becomes a vector string across each point. Note that FIG. 6B shows a state in which the curve is vectorized, and the curve is a vector sequence vector sequence v1. The vectorization of ν2 v5 is calculated.

As described above, the vectorization circuit 12c converts the dot string into a vector string.

The vector string signal from the first vectorization circuit 12C is supplied to the vector length calculation circuit 12d, and the vector length calculation circuit 1.2
The length of each vector in the vector sequence is determined by d. In addition, the vector length calculation circuit 12d calculates the vector length (Cfa point, distance) based on the coordinate point sequence of the end points of each vector.
Outputs a sequence of vector lengths. Here, the starting point and ending point of any vector ■ are (X,, ,) respectively! S), (X/
,

As described above, the vector length calculation circuit 12d determines the length of each vector in the vector sequence, and this vector length sequence signal is supplied to the determination circuit 12e.

The determination circuit 12e determines that each vector length V has a preset threshold value t.
It is compared and determined whether each vector length exceeds the threshold t.
If the length of each vector is longer than the threshold t, the vector string is supplied as a straight line candidate to the straight line recognition device 14,
If the vector sequence is shorter than , the vector sequence is supplied to the curve recognition device 16 as a curve candidate. For example, in Fig. 6Δ, the vector length v1 of the vector sequence V is longer than the threshold value t, so the vector sequence ν is determined to be a straight line candidate, and in Fig. 6B, the vector sequence v1 ．． Since the vector lengths VI+V2 ``''''''' vs of v2v5 are all shorter than the darkness value t, the vector sequence V,, V2v, is determined to be a curve candidate.

Therefore, the vector sequence is separated into straight line candidates or curve candidates by the determination circuit 12e, and is supplied to the straight line recognition device 14 and the curve recognition device 16, so that the straight line recognition device 14,
The fill line δ tracking device 16 can process dot sequences in vector units, making it possible to increase the processing speed.

Furthermore, in the straight line/tflj line data separation device 12', as shown in Fig. 6Δ and Fig. 6B, all one dot row can be separated into straight line candidates or curve candidates. Instead, part of the dot sequence of the - tree is separated into straight line candidates or curved candidates, and as a result, the dot sequence of the - tree is divided into 1-11: i, iij 1'1i, = +
, Li line portion. for example,
Fig. 6C shows a state in which the combined figure [3] of straight lines and curves has been converted into a vector, and the dot string of the combined figure is the vector string νl, 'V2, V3 ''''-''' V6. has been made into

Then, the straight line/curve data separation device 12' separates the dot sequence corresponding to the vector sequence V1 into straight line candidates, and further separates the dot sequence corresponding to the vector sequence V1 into straight line candidates. The dot string corresponding to V3v, , can be separated into curve candidates, and the one-tree dot string can be separated into a straight line part and a curved part.

Therefore, even if the figure is a combination of a straight line and a curve, the straight line recognition device 14 and the curve recognition device 16 can accurately recognize one straight line candidate and one curve candidate. Therefore, the loop of dividing the dot row into two and feeding it back to the straight line/curve data separation device 12 as in the conventional device can be eliminated. It is clear from the comparison that in Figure 5,
It will be appreciated that no dot row bisecting device is provided.

As described above, according to the first embodiment of the present invention, it is possible to process dot sequences in vector units, and furthermore, it is possible to eliminate the need for loops in the neck of the straight line/curve data recognition device. Therefore, it is possible to increase the processing speed of the entire device.

In addition, in the above embodiment, the vectorization circuit 1.2
In C, the dot string is vectorized by the processing method described above, but other processing methods can be used as long as the dot string can be vectorized within a certain tolerance range.

As explained in Section 2, "Effects of the Invention", the straight line/curve data splitting device of 1JII of the present invention converts a dot string into a vector, determines the length of each vector in the vector string, and determines whether the vector length exceeds a threshold value. Depending on the situation, part or all of the dot row is divided into straight line data or curved data. Since there is no signature Q and processing is performed in vector units, when this straight line/curve data separation device is used in a straight line/curve data recognition device, the processing speed of the straight line recognition device and curve recognition device can be increased. . Further, according to the straight line/curve data separation device of the present invention, a part of a dot string is separated into straight line candidates or curve candidates, i.e., one dot string is divided into 519 straight lines and curved parts by G4. Even if the figure is a combination of straight lines and curves, the subsequent line recognition device and curve recognition device can accurately recognize one straight line candidate and one curve candidate. This eliminates the need for a loop in which an undeterminable dot row is divided into two and fed back to the straight line/curve data separation device.

Therefore, according to the present invention, it is possible to increase the processing speed of the entire straight line/curve data recognition device.

[Brief explanation of the drawing]

Figure 1 is a schematic configuration diagram of a conventional straight line/curve data recognition device. Figure Δ and Figure 2B each have one straight line curved digitally.
]'- Theory 1 does not show the converted dot row! Figures Ij and 3 are block circuit diagrams of conventional straight line/curve data separation devices, and Figures 4Δ, 4B, and 4c are straight lines, respectively. An explanatory diagram showing a dot array obtained by converting a curve, a straight line, and a combined figure of curves into digital gamma data, and showing the single point, middle point, and end point. The block circuit diagram of the applied straight line/curve data delta ninja device, Figures 6A, 6B, and 6C are straight lines, respectively. FIG. 3 is an explanatory diagram showing a vector sequence obtained by vectorizing a curve, a straight line, and a combined figure of curves. 12' Straight line/curve data separation device, 12c Vectorization circuit, 12d Vector length calculation circuit, 12e Judgment circuit. Figure 1 Figure 3 Figure 2 @2Aiffl Yt Figure 4A Figure 4B Figure 4C Figure 6A Figure 6B Figure 60

Claims (1)

[Claims] (1) A straight/curve data separation device that separates a figure containing a mixture of straight lines and curved figures into straight line data and curved data, which includes a vectorization circuit that converts a dot string representing a previous figure into a vector and outputs the vector string; a vector length calculation circuit that calculates the length of each vector in a vector sequence based on a vector sequence signal from the vector conversion circuit and outputs a numerical sequence of the vector length; 1. A determination circuit that compares and determines whether or not a dot row is present, and discriminately separates part or all of a dot row into straight line data or curve data.