JPS6079479A - Linear graphic encoding method - Google Patents

Abstract

PURPOSE:To execute a processing at a high speed by dividing a picture element of a linear graphic by determining in advance a connecting direction to an adjacent picture element, making each segment formed by connecting each picture element approximate by a polygonal line consisting of one - several straight lines, and encoding it. CONSTITUTION:A picture element 3 of a linear graphic is set, for instance, to eight directions of 0-7 by determining in advance a connecting direction to an adjacent picture element. A segment 5 is formed by connecting from a picture element of a start end 3-1 to a picture element of a final end 3-2 as shown in the figure. The maximum vertical distance l of each picture element 3 and the segment 5 is derived, and if l is within a reference value, the segment 5 becomes a final approximate straight line. If the distance l is larger than the reference value, the picture element 7 whose distance to the segment 5 is the largest is derived, and the segment is divided into 6-1 and 6-2. With regard to the segments 6-1, 6-2, too, the same decision is executed, and it is continued until the distance lbecomes below the reference value. In this way, the linear graphic can be encoded quickly within a necessary approximate accuracy.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a line figure encoding method that performs high-performance straight line approximation encoding based on chain-encoded line figure data.

Conventionally, a chain encoding method and a linear approximation encoding method are known as typical encoding methods for line graphics.

In the chain encoding method, after thinning a given line figure so that the line width is 1 pixel, the pixels constituting the line figure are to 0
, 1, 2--7, which direction of the eight directions are connected is encoded using a 3-bit/pixel code.

In addition, as shown in FIG. 2, the linear approximation encoding method applies a linear tube 2 having a width d to a line figure l, and each pixel column included in the tube has one I
Assuming that it can be approximated by a U line, the pixel positions at both end points of the pixel column in each tube are the starting points.

This is a method of approximate encoding using a straight line as the end point.

The above chain encoding method faithfully encodes without any full curve approximation, so it has the disadvantage that the number of codes is larger than the IRF line approximation encoding method, and the linear approximation encoding method As shown in Figure 2, the approximation accuracy is determined by the width d of the encoding tube 20, so after encoding, it is not possible to change the approximation accuracy and re-encode using the linear approximation code data. Ta. Therefore, if you want to obtain linear approximation encoding results with different approximation accuracy depending on the application, you must store original line figure data with a large amount of data, and also ensure that each pixel is within a tube with width d. Since it is necessary to make a judgment as to whether or not this is possible using two-dimensional coordinate values of pixels, there are drawbacks such as an increase in processing costs.

An object of the present invention is to eliminate the drawbacks of the conventional line figure encoding method as described above, by using chain encoded data that can perform faithful encoding with less data than original line figure data. Therefore, it is an object of the present invention to provide a σ encoding method for line figures that performs linear approximation encoding with high efficiency.

In the line figure encoding method t of the present invention, in which a line figure of one pixel width is encoded by approximating it with a broken line, the direction of a series of adjacent pixels is predetermined for the pixel row forming the line figure. The method is characterized in that it is divided into groups of pixel rows in adjacent directions, and each segment formed by the divided individual pixel rows is approximated and encoded with a broken line consisting of one or more straight lines. be.

Hereinafter, the method of the present invention will be explained in detail based on the drawings.

FIG. 8 is an explanatory diagram of a segment chain encoding method, which is one type of chain encoding method, in which 8 is an individual pixel forming a line figure, 4 is a chain, and 5 is a segment. The table in FIG. 4 shows the correspondence between chain codes in the chain encoding method and octant codes and direction codes in the segment chain encoding method.

The segment chain encoding method is a method of reducing code classes by focusing on the fact that line segments forming a line figure often continue in the same direction when viewed locally. (Reference: Journal of the Institute of Electronics and Communication Engineers, vol. J 65-D, 11
, p, pl 890 to 1397. (“Segment chain encoding of line figures”) Specifically, as shown in FIG. and each segment 5
According to the table shown in the fourth figure, the range of pixel rows in which two types of chain codes in an adjacency relationship shown in the chain code column follow is defined as segment 5 according to the octant codes corresponding to those chain codes in the upper column. Each is a method of encoding. At the same time, the distinction between the two types of chain codes in segment 5 encoded in the octant code is encoded using the direction code on the lower axis of the table in the same figure, and the line shape is created using this direction code and the octant code. The coded data is obtained.

Such a segment chain encoding method can be easily implemented using conventional techniques. For example, two-dimensionally optically scan the line diagram II, accumulate the image signals obtained by photoelectric conversion, and detect the pixels adjacent to the pixels of the reference plane using a known method, as shown in the HBi diagram. According to the adjacent direction, the chain code is encoded as shown in FIG.
The configuration shift may be such as to generate a direction code of 3 bits/segment and an octant code of 3 bits/segment.

5 and 6, a pixel row forming a line figure is divided into pixel rows in which the directions of a series of adjacent pixels are in predetermined adjacent directions, and each segment formed by the divided pixel rows is shown in FIG. FIG. 4 is an explanatory diagram of an embodiment of the method of the present invention when segment chain encoded data of segment 5 in the segment chain encoding method described above is used.

Figure 5 is an explanatory diagram when one segment 5 is approximated by one straight line, where 8 is a pixel, 3-1 and 8-2 are pixels at the start and end points of segment 5, and 6 is a segment. 5 shows an approximate straight line corresponding to 5.

On the other hand, in FIG. 6, one segment 5 is connected to two iMMs.
FIG. 3 is an explanatory diagram of approximate encoding using a broken line consisting of B21.6-2. Note that 7 indicates a breaking point.

The method of the present invention encodes each segment 5, and as shown in FIG. 1
Distance l from each pixel 8 of the pixel row forming the segment 5 to be centered
The approximation accuracy of the approximate straight line 6 is defined by the maximum value of .

The flow of the encoding process in the method of the present invention will be explained below based on FIGS. 5 and 6.

(1) In FIG. 5, a straight line whose end points are the starting point 8-1 and the ending point 8-2 of the segment 5 is the first approximate curve 6.

(Phrase) Measure the distance between each pixel 8 and the first approximation straight line 6, check whether the maximum distance l is within the allowable range, and if the maximum distance l is within the allowable range, the first approximate line 6 as the final approximation straight line, and the corresponding segment 5 as one i[
Approximately encode with M.

(8) If the maximum distance l in Off is outside the allowable range, select the pixel with the maximum distance from the first approximation straight line 6 (if there are multiple pixels, one of them) as shown in Figure 6. Set the break point to 1 as shown.

(4) Next, as shown in FIG. 6-1
Then, an approximate straight line 6-2 connecting the corner point 7 and the end pixel 8-2 of the segment 5 is divided into two.

(5) For each range of pixel rows divided at break point 7,
The corresponding second-order approximation straight line 6-1.6-2 and each II! iI
Check the distance to the pixel column, and if the distance is within the allowable range, the range of that pixel column is determined by the corresponding second-order approximation ilt line 6
The final approximation is to encode -1.6-2 as a line. Also,
For the range of pixel rows that include pixels whose distance is outside the allowable range, set a new breaking point within the range of that pixel row in the same manner as in 1) 11 (8), and make sure that the approximation accuracy is within the allowable range. Repeat the above-mentioned (8) and subsequent processing until the amount is within the range.

(6) As described above, each segment 6-1, 6-2 is approximated and encoded using a broken line consisting of one or a rectangular line whose approximation accuracy falls within the range of δ1.

Next, a method of calculating the distance t between the approximate straight line 6 and each pixel 8 of the corresponding pixel column will be explained with reference to FIG. Here, we will explain the first-order approximation straight line, but the second-order or higher approximation straight line can also be explained at the starting point pixel 8-1 or the ending point pixel 3.
If -2 is replaced with a corner pixel as necessary, the following explanation can be applied as is.

In FIG. 7, 8.9 indicates two types of chain pectroofs 1.9 included in one segment 5. yi, each having a length corresponding to the distance between the centers of two adjacent pixels in the direction indicated by the chain code.

For the purpose of explanation, the approximate straight line 6 is considered to be a vector having a direction from the position of the starting point pixel 8-1 to the position of the ending point pixel 3-2, and this is expressed as Bi.

At this time, the following formula holds true.

"-(曾□10=-+, 10---10ze.) +(7, 12+----+The) ---(1) In the above (formula 11), m and n are in segment 5. The number of included chain vectors W□l is 71. As a result, as shown in FIG.
Let yi be the component orthogonal to the vector function u/i r vl
If we consider it separately from the component parallel to i, we get: +L v,, 112
+---/l/2. 1m 10v')-0---(1)/n. That is, the variation ratio of the distance from the straight line f caused by each chain vector uilv is /
, /, which vary in opposite directions to each other m n . If either m or n is zero, it is clear that the segment 5 completely coincides with the straight line 6, so here it is set as m-n\0.

From the above, the pixel having the maximum distance from the approximate straight line 6 can be determined as follows. That is,
When we trace the 1110th pixel column from the starting point pixel 3-1 and reach the k-th pixel, the numbers of chain vectors v, , V, that have passed up to that point are M and N, respectively.
If the number of pixels is 1 l, then the distance between the corresponding pixel and the approximate straight line 6 is 111
Since dk is given by dk-α·lto-(8) (M×N-, α is a proportionality constant), the pixel for which this dk is maximum is the pixel to be included.

With the 7 segment chain encoding method (well, segment 5
The number of pixels f m+n ) in the chain vector uil Vi is known, and the chain vector uil Vi can be identified by making it correspond to the direction code in the table of FIG. By the above processing, it is possible to easily find the pixel with the maximum dk.

Next, in order to evaluate the approximation accuracy of the approximate straight line 6, the absolute value of the maximum distance of the dk is calculated. The processing for this need only be performed for the pixels that give the maximum distance determined by the above-described processing, so the processing can be performed in a short time using the calculation formula described below.

In other words, the coordinate value of the pixel that gives the maximum distance is (Xk+
Vk) Each position of the starting point pixel 3-1 and ending point pixel 8-2 of r segment 5 is set to 1 (Xs, y8) and (X
e, yo), the maximum distance dmaX that can be claimed is given by:

Note that the coordinate values of the segment chain-encoded pixels are based on the coordinate values of the starting point pixel 8-1, and for the segments indicated by the octant codes in the table of FIG. In relation to the sign, it is clear that it can be determined by adding and subtracting the value l to the X value or Y value %b% of the coordinate value, Hayashi*re, and both cylinders. Arithmetic processing can also be performed relatively easily.

Although the processing method for the llljm segment has been described above, it is clear that any segment can be processed in the same way.

The maximum distance rAi'p obtained through calculation processing in this way
The encoding process after evaluating the approximation accuracy using dmax is as described above with reference to FIGS. 5 and 6.

The practical example of the method of the present invention using segment chain encoded data has been described above, but the conversion of ordinary chain encoded data to segment chain encoded data can be performed using the table shown in Figure 4. Therefore, it goes without saying that the method of the present invention can be easily applied to chain-encoded data by executing this transformation as the l'l+Jα principle.

As explained in detail above, the method of the present invention allows line figures to be
1J6 pixel rows in the direction of adjacent pixels.
divided into groups of pixel rows in predetermined adjacent directions,
Since this method performs approximate encoding using broken lines consisting of one or more straight lines for each segment consisting of each divided individual pixel column, it has the effect of compressing the data of line figures and encoding them with high precision. . In particular, segment chain 1 encoded data, which is significantly compressed compared to raw drawing data expressed as a two-dimensional array of pixels, is accumulated as basic data, and based on this data, the method of the present invention is applied. If encoded, there is an advantage that linear approximation encoded data can be obtained with a higher data compression effect depending on the required approximation accuracy. This (for example, in a drawing database system, segment chain encoded data is stored as basic data at the center, and this data is transferred to a remote
When outputting to the Y plotter, it is possible to convert data into linear approximation data and transmit it within a desired approximation accuracy, which is extremely advantageous in terms of reducing transmission costs.The effects obtained by the method of the present invention are extremely significant.

[Brief explanation of drawings]

FIG. 1 is a diagram showing the direction of the chain in the chain encoding method, FIG. 2 is an explanatory diagram of the conventional linear approximation encoding method, FIG. 8 is an explanatory diagram of the segment chain encoding method + 1iD, Fig. 4 is a diagram showing the direction T' relative to the chain code: one thread between the corresponding f and octant codes, and Fig. 5 and Fig. 6 are examples of cases where segment chain encoding codes are used. FIG. 7 is an explanatory diagram of an example of a method σ) for calculating the distance between an approximate straight line and a pixel. 1... Diagram 2... Tube 3, 8-1.3-2... Pixel 4... Chain 5.
...Segment Q, 6-1. +3-2...Approximate straight line 7...Break points 8, 9...Chain vector. Figure 6 '/l;ri -1

Claims (1)

[Claims] In a method of encoding a line figure of 11 pixels and 1 width by approximating it with a broken line, the pixel rows forming the line figure are divided into a group of pixel rows in which the direction of a series of adjacent pixels is in a predetermined direction. A line figure encoding method characterized by dividing a pixel column and approximating and encoding each segment formed by the divided individual pixel columns with a broken line consisting of one or more straight lines.