JPS61874A - Coding system of binary coded picture data - Google Patents

Abstract

PURPOSE:To shorten the refresh time of pictures and to attain the satisfactory compression of data without producing any loss of information quantity, by providing a line thinning part where the pictures are converted into lineworks, a start point detecting part, a coding part, etc. CONSTITUTION:The picture data read out of a map 1 is converted into the binary coded picture data by an input part 2 and undergoes the line thinning process through a line thinning part 4 via a memory 8. Then the start point of the picture data is detected at a start point detecting part 5 to check the adjacent direction of the next picture element continuous to the start point via the memory 8. The coding process is performed after the adjacent direction is decided. When the coding is through with the picture element in such a way, this picture element is erased out of the original picture data. If the end point is detected with the picture element under coding, the procedure proceeds to the part 5 again from a coding part 6 for detection of the next start point. In such a way, the picture data is coded through the part 4, etc. and the satisfactory compression of data is attained and also the refresh time of pictures is shortened.

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to a method for processing image data represented by line image data such as road maps, and in particular to a method for encoding image data to compress the amount of such image data. Regarding the method.

[Background of the invention]

In an automobile navigation system, a large number of road maps of different ranges must be prepared in advance, and one of them must be displayed as desired.

For this reason, conventionally, road maps drawn on film or paper have been used, which poses problems in that map selection and display switching is troublesome and automation is difficult.

Therefore, a system is used that uses electronic image processing technology to convert a large number of road maps into image data, stores this in advance in memory, and reproduces the image on a CRT or the like as needed. It's starting to look like this.

By the way, in the navigation system mentioned above,
Requires quite a lot of maps. Therefore, if this is converted into image data as it is, the amount of data will be quite enormous, which will require a large memory capacity, increasing costs and making implementation difficult.

Therefore, various image data compression methods, such as vectorization methods and run-length coding methods, have been proposed in the past, but vectorization methods inevitably cause a loss of information and degrade image quality, The coding method had drawbacks such as it took too much time to reproduce images and was not suitable for navigation systems.

On the other hand, such drawbacks have been removed (as for
The invention disclosed in Publication No. 139567 has been proposed, and although it is an effective proposal in terms of reducing the loss of information,
Because the emphasis is on versatility, data compression leaves a bit to be desired.

[Aim of the invention]

An object of the present invention is to eliminate the drawbacks of the prior art described above, obtain sufficient data compression without causing loss of information, and
Moreover, it is an object of the present invention to provide a binary image data encoding method that requires a short image refresh time.

[Summary of the Invention] In order to achieve this object, the present invention makes use of the characteristics of images that mainly consist of line drawings such as road maps, treats the image as a set of multiple line segments, and then identifies each line segment as its starting point. It is characterized by a point expressed by position data, data representing the adjacent direction of each successive adjacent pixel from the starting point, and data representing the end point of each line segment.

[Embodiments of the invention]

Hereinafter, a binary image data encoding method according to the present invention will be explained in detail with reference to the illustrated embodiments.

FIG. 1 shows an embodiment of the present invention, where 1 is a map (original image) from which image data is to be read, and 2 is an input unit for reading image data, such as a television camera using a CCD or an image pickup tube, or a drum scanner. 3 is a binarization unit for binarizing the image data read by the input unit 2, 4 is a thinning unit for converting the image into a line drawing, 5 is a starting point detection unit, and 6 is a thinning unit for converting the image into a line drawing. An encoding unit, 7 an output unit, 8 a memory,
Encoding processing is performed by these parts.

Next, the operation of this embodiment will be explained.

The image data read from the map 1 by the input section 2 is converted by the binarization section 3 into binarized image data consisting of only white parts and black parts, and is temporarily stored in the memory.

However, although the image data obtained at this point is line drawing data corresponding to map 1, the line widths vary, so this is subjected to line thinning processing by the line thinning section 4. Then, line drawing data with a width of one pixel is obtained. An example of image data subjected to line thinning processing in this manner is shown in FIG. 2. After the line thinning data of map 1 is stored in the eight memories, the process proceeds to encoding processing.

First, the starting point detection section 5 detects the starting point S (FIG. 2) of the image data. This starting point detection operation is performed by sequentially scanning each pixel of the image data in the X and Y directions and detecting the position of the first pixel found. Therefore,
In the image data of FIG. 2, this is the starting point S(4,2).

Once the starting point S has been detected in this way, its coordinates (4, 2) are stored in the memory 8. In this case, the number of bits required to represent the X and Y coordinates will vary depending on the size of the map 1 to be encoded, but in this example, 1 byte is used for each; Starting point S
The data representing this is shown in FIG. In this figure, 11 represents the X coordinate (4) and 12 represents the Y coordinate (2).

Next, the adjacent force direction of the next pixel following this starting point S is searched. At this time, the adjacent directions are examined in eight directions from 1 to 8 using a 3×3 mask as shown in FIG.

In the case of the data in FIG. 2, if the center of the mask in FIG. 4 is made to coincide with the starting point S, it can be seen that the next pixel is adjacent in direction 3. Once the adjacent direction is determined in this way, it is converted into data. At this time, there are eight types of directions in total, and if there is no adjacent pixel at the starting point S, 4 bits are used to represent it as 0. Pixel adjacent to S (
4, 3), since its adjacent direction is direction 3 in Figure 4, it will be encoded as shown by 13 in Figure 5. When the encoding of the adjacent pixels is completed, the pixels will be converted to the original image. Erase from the data (thinned data, for example, FIG. 2).

After encoding pixel (4,3), use the 3×3 mask shown in Figure 4 again, and set the center of this mask to pixel (4,3).
, 3), and examine the adjacent direction of consecutive pixels.

However, although all eight directions were examined to detect adjacent directions at the starting point S, in the case of pixels other than the starting point S, the detection was performed not in all eight directions, but in the case of detecting adjacent directions at the starting point S (in this case, pixel (4,
3)) is further extended in the direction (direction 3 in this case) that was adjacent to the previous pixel (in this case, pixel (4, 2)), and the direction to the right and Only the three directions to the left are investigated.

This is because the main data in maps required for navigation systems is roads represented by lines, and unlike ordinary line drawings, the continuous direction of such road data does not change rapidly. , which makes encoding easier.

In other words, since pixel (4, 3) is adjacent in the direction 3 from the previous pixel (starting point S), the next pixel (5, 4) is adjacent in the same direction 3 and pixel ( 4, 3), only the front left direction 2 and the front right direction 4 are examined.

Now, in the image data in Figure 2, the pixel that follows pixel (4, 3) has coordinates (5, 4), which corresponds to the front diagonal left direction, so as shown in Figure 6, it is (01) 2. Therefore, the data is encoded as 14 in FIG.

Note that the forward straight direction is encoded as (10)z, as shown in FIG. 6, and the forward right diagonal direction is similarly encoded as (11)z.

Once the pixel (5, 4) has been encoded in this manner, this pixel is deleted from the original image data, as described above.

Next, as a pixel continuous to pixel (5, 4), pixel (
There are two pixels: pixel (6,5) and pixel (4,5).

However, if we use the 3x3 mask in Figure 4 and examine the adjacent direction by aligning its center with pixel (5, 4), the straight direction of pixel (5, 4) at this time will be the same as that of the 3x3 mask. There are two directions, therefore, the front left diagonal direction is 1, and the front right diagonal direction is 3, and as is clear from FIG. 6, the other directions are ignored. Therefore, at this time, pixel (4,5) is ignored because it is not in any of the adjacent directions of directions 1, 2, and 3, while pixel (6,
, 5), the adjacent direction is the straight direction, that is, the direction of 2 of the mask in Fig. 4, so from Fig. 6, (10) 2
As a result, data shown at 15 in FIG. 8 is added. After this, pixel (6, 5) is erased.

Similarly, the next pixel (7,6) is also encoded as (10)2 and added as data 16 in FIG.

Next, erase pixel (7, 6), use the 3x3 mask in Figure 4 again, match the center to (7, 6), and examine consecutive pixels. At this time, K is the pixel No pixels are adjacent in the forward direction of (7, 6), that is, in the 2 direction of the 3×3 mask, nor in the 1 direction to the left of it, nor in the 3 direction to the right.

Therefore, at this time, it is determined from FIG. 6 that it is the end point, and the data is converted into data using the code (00)2 representing this end point, and added as data 17 in FIG.

The encoded data up to this point is as shown in FIG. 9, and FIG. 10 shows this divided by byte. Note that the last 4 bits are less than 1 byte, so they are padded with 0's to make 1 byte.

When the encoding up to this point is completed, the pixels of the image data in FIG. 2 have been erased as a result of the encoding.

The image data is as shown in FIG.

On the other hand, when the end point is detected at the pixel being encoded,
The operation returns from the encoding unit 6 (FIG. 1) to the start point detection unit 5 again, and pixel (4, 5) is detected as the next start point S. Hereafter, pixels (4, 6), (4, 7) , is encoded, so the encoded data based on the image data in Fig. 11 is the first
It will look like Figure 2.

Therefore, the image data in Figure 2 ends up being the image data in Figures 10 and 12.
From the figure, it is encoded and represented as shown in FIG.

Therefore, the output unit 7 (FIG. 1) writes the data shown in FIG. 13 into a nonvolatile memory after attaching a predetermined identification code to the data shown in FIG. 13 as corresponding to the image data shown in FIG. Then, this memory is installed in an image reproducing device of a navigation system, etc., and encoded data is selected and read out as necessary to reproduce images.

Next, the thinning section 4, the starting point detection section 5, and the encoding section 6 (
The details of the image data encoding process according to FIG.
This will be explained using the flowchart in Figure (αl, (A+).Although not particularly annotated, in the above explanation, the pixel is the part that is black when the background is white on the image plane ( If the white pixel data is "0", the black pixel data is "1".This is the same in the following explanation, and the line part of the image plane is Pixel is 1", other pixels are 0"
It's Nishide.

Although it is omitted in Fig. 1, this embodiment includes a computer (such as a microcomputer), and the encoding process is performed according to a program stored in the computer. The process starts at a predetermined timing after the storage of one frame worth of images read from the map 1 from the unit 2 to the memory 8 is completed.

Starting at 1401, the next step 1402 inputs the binarized image data into the array A(1, J), and 1403
Thin the line to make the line width one pixel.

Next, from 1405 to 1408 and 1410 A(I,
Find out which pixel in J) is 11".■.

Since J increases sequentially from 1, start with 1"
The pixel where is found is set as the starting point S of the line.

I and J at this time are output as 1411. At step 1412, the direction of continuity of the points (I, J) just output is checked. The eight directions shown in FIG. 4 are considered as continuous directions. Point (I.

Centering on point J), check the points around it to find out where the value is 1''. The order is direction 1, direction 2, and direction 3. If '1'' is found, that direction is output in step 1413. If no consecutive points are found in any of the 8 directions, 1
At 416, "oo" representing the end point is output. At step 1414, the direction of the continuous points is assigned to the variable DIR. Next, from 1419 to 1423, DIR7) A direction adjacent to each other by 45 degrees in a counterclockwise direction is assigned to DIRI, DIR is assigned to DIR2, and a direction adjacent to I) IRQ by 45 degrees in a clockwise direction is assigned to DIR3. The continuous points found at 1412 are further
Check which direction of IRI, DIFL2, and DIR3 is continuous. When a continuous point is found in the direction of DIR2 at 1424, it outputs lO'' at 1431, and DI is output at 1432.
RK DI) 12 and go to 1419. DII at 1425 (11 at 1429 if a continuous point is found in the direction of 3)
”, and at 1430, substitute DIR3 for DIR and 14
Go to 19 (.If a continuous point is found in the direction of DIRI at 1426, output 01" at 1427, and output D to DIR at 1428.
Substitute IR1 and go to 1419. DI R1, DII'
If a continuous direction is not found in either t2 or DIR3, "oo" is output in 1416 and the process goes to 1404. Then, in 1408, when J=INY-t1 is reached, the process ends.

Note that INX and INY are the sizes of map data.

According to this embodiment, it is not necessary to encode every pixel of the image data, so the amount of data can be significantly reduced. Since all meaningful pixels are encoded as data, there is no loss of information during image reproduction, and only meaningful pixels are converted into data. Reading encoded data from memory and converting it to image data can be performed in a short time, and reproduction images can be selected and switched at high speed.

〔Effect of the invention〕

As explained above, according to the present invention, the amount of data required for image reproduction can be significantly reduced without causing any loss in necessary information. It is possible to provide a binary image data encoding method that can hold data in a small memory lj [1, and also allows high-speed image transfer without deteriorating the image quality of the reproduced image.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of the binary image data encoding method according to the present invention, FIGS. 2 to 13 are explanatory diagrams of data during encoding processing operation, and FIG.
Figures (α) and <h> are flowcharts for explaining the operation. 1...Map, 2...Input section, 3...
... Binarization section, 4 ... Thinning section, 5 ...
...Starting point detection unit, 6... Encoding unit, 7...
...Output section, 8...Memory. →Tame】Easy

Claims (1)

[Claims] 1. In a method for processing image data consisting of binary pixel data, the above image data is divided into a plurality of line segment data consisting of a series of one or more pixels with a width of one pixel, and each of these plural line segment data is , data representing the pixel position of the starting point of the line segment data, and a plurality of data groups representing adjacent directions of each of the plurality of pixels successively following the starting point pixel;
A binarized image data encoding method characterized in that it is configured to be represented by three types of data: data representing a pixel at the end point of line segment data.