JPS5916464B2 - Color figure encoding/decoding method and device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method for efficiently run-length encoding and decoding color image signals, and apparatus for use in its implementation.

Conventionally, PCM, ΔM, DPCM, etc. have been common as methods for digitally encoding color image signals.

These methods can digitally encode color natural images (landscapes, people, etc.) with multiple gradations;
To store this in memory, even the most efficient DPCM requires a huge amount of memory, about 1.6 x 106 bits per screen. In recent years, CAI (Computer Assistant), which performs program learning using electronic computers, has become popular.
edInstru-CtiOn) is attracting attention, but in this case, the screen presented to the learner is often limited in the colors used and is composed of relatively simple color graphics. For such screens, applying the run-length encoding method, which is generally known as a band compression technique in facsimile, can significantly compress the amount of information (
It is expected that more than 1/5 of DPCM) will be achieved. However, in such a method, when a color graphic includes many elements with short run lengths such as characters, the information compression rate generally tends to decrease. The present invention was made in order to eliminate these drawbacks, and objects with short run lengths such as letters and line drawings are often represented in achromatic colors with white or black outlines, and the human visual sense of color is Taking advantage of the fact that it does not actually cause much of a problem if only the luminance signal is used, excluding the chromaticity signal, for short run length image signals, only the run length and luminance signals are encoded for short run length image signals. For long run length signals, if the run length exceeds the range that can be encoded with a fixed length code or if it is interrupted by a short run length signal, only the remaining run length is encoded without recoding the image signal. The present invention is characterized in that the code length is shortened by doing so, and the drawings will be described in detail below.

FIG. 1 is a diagram explaining the principle of the encoding method according to the present invention.

Now consider the case where a color television signal is run-length encoded for a color figure as shown in FIG.
are encoded with 2 bits each, and for short run length signals, only the luminance signal Y is encoded with 2 bits.
In FIG. 1a, if the image signal between AB is a long run length signal, each Y, I, and Q signal is encoded with 2 bits, for example, as "101,""11" and "01", and the run length ( RL) is 80 dots, the run length encoded information between AB is represented by 2 bytes with 1 byte consisting of 8 bits as shown in Figure 1b.Here, the upper 2 bits are the long run length code (hereinafter referred to as color code). If the upper 2 bits of the code identification code are "00", the following 6 bits represent a digitally encoded image signal, and the next 1 byte represents the run length of the image signal. In addition, when the image signal between BC is a short run length signal, only the luminance signal Y is encoded with 2 bits, for example, "101", and the run length (RL
) is 2 bits, the run-length encoding information between BC is represented by 1 byte as shown in FIG. 1c. Here, the upper two bits are a code identification code indicating a short run length code (hereinafter referred to as the brightness code), and the upper two bits are 01.
If so, the next two bits represent the digitally encoded luminance signal, and the next four bits represent the run length of the luminance signal. Next, the run length between CDs is 8.
If there is 0 dot and the image signal is the same between A and B,
Only the run length is encoded, and the run length encoded information of the image signal between CDs is expressed in one byte as shown in FIG. 1d, and this is called a run length code. Here, the upper 1 bit is a code identification code indicating a run length code, and if the upper 1 bit is "1", the following 7 bits represent the run length (RL) of the interrupted long run length signal, and this The image signal of the section is assumed to be the same as the image signal before the short run length signal is inserted. Also, when the run length between AB is long and cannot be encoded with the 8-bit run length code of the color code and overflow occurs,
The run length code encodes only the run length of the overflow and connects it to the color code. At this time, the overflow portion is quite long, and the run length code is 7.
If an overflow occurs even after adding a bit run-length code, one more run-length code must be added to the color code, making a total of two run-length codes. However, in this case, two run-length codes (
Even if a color code (2 bytes) having an 8-bit run-length code is used instead of a total of 2 bytes), the amount of encoded information is the same, so there is no particular need to use a run-length code. This also applies to the above-mentioned encoding of long run length signals interrupted by short run length signals between CDs using run length codes. In the embodiment of the encoding circuit described later, a configuration is shown in which when two run-length codes are required, encoding is performed using a color code. Naturally, the run length code consists of 9 bits (of which 1
(bits are code identification codes), it is clear that the amount of encoded information is smaller when one run-length code is used than when one color code is used. In the above explanation, an example of encoding is given in byte units, but of course the number of bits given to the color signal, luminance signal, and run length is not limited to this, and it goes without saying that other bit configurations can be adopted. Nor. In addition to the I and Q signals, other methods of expressing chromaticity signals such as hue and saturation can also be used as chromaticity signals. FIG. 2 shows an embodiment of an encoding device to which the method of the present invention is applied.

In the figure, 1 is a color figure, which is imaged by an imaging device 3 consisting of a color television camera, a color flying spot scanner, etc., and color signals R, G, etc.
, B into electrical signals. 2 is a synchronization signal generator that supplies a synchronization signal to the imaging device 3.

The R, G, and B signals obtained by the imaging device 3 are converted into Y, I, and Q signals by the matrix circuit 4, and then input to the analog-to-digital converter 5, which outputs the clock oscillator 6. It is converted into a digital signal at a sampling frequency determined by a clock. In the figure, the output clock of the clock oscillator 6 is also supplied to the synchronization signal generator 2, and the clock oscillator 6 and the synchronization signal generator 2 are synchronized, but it is necessary to take into account the negative effects of edge business, etc. If not, both may be operated asynchronously. The signal digitized by the analog-to-digital converter 5 is input to a run length encoder 7 and also to a change detection circuit 8. Furthermore, as seen in the code structure of each code shown in Figure 1B, c, and d, when encoding an image signal with a small number of bits (2 bits for each separated signal in the example of Figure 1), , particulate noise is likely to occur as a result of quantization due to a small number of levels, and in particular, the I and Q signals take positive and negative values, so the quantization level is insufficient, and the I and Q signals take values near zero, making it an achromatic signal. cannot be encoded.

Therefore, when encoding an image signal using a small number of bits, the configuration of the analog-to-digital converter 5 first performs analog-to-digital conversion using a large number of bits (for example, 4 bits for each separated signal), and then performs analog-to-digital conversion using wiring logic. It is desirable that the configuration has a nonlinear quantization function that converts the multi-bit code of each separated signal into a small-bit code (for example, 2 bits for each separated signal) and outputs the code. At the time of decoding, a digital-to-analog converter is required, which converts a small-bit code into a multi-bit code using wiring logic, and then performs digital-to-analog conversion. The present inventor has already filed a patent application for encoding/decoding technology using a small number of bits involving such nonlinear quanta (Japanese Patent Application No. 105021-1971, September 3, 1978). The analog-to-digital converters and digital-to-analog converters included in the embodiments of the present invention are also intended to have the above-mentioned nonlinear quantization function. The run length is given by counting the clock pulses from the clock oscillator 6 with a counter 9, and the run length count value is input to the run length encoder 7.

The run length, which is a standard for separating short run length signals and long run length signals, generally differs between chromatic colors and achromatic colors.

For chromatic colors, the run length that can display colors is determined by the visual characteristics of the eye or the band limit of the chromaticity signal in the color television broadcasting system, and short run length signals and long run lengths are determined based on the minimum run length that can display colors. Separate run-length signals. For achromatic colors, the information is only the luminance signal, so the run length standard is the run length determined by the number of run length encoding bits in the luminance code (4 bits in the example in Figure 1 c).
In this example, 15 dots) may be used as the standard. As mentioned above, since it is necessary to switch the reference run length between chromatic colors and achromatic colors, the embodiment shown in FIG. 2 has an achromatic color detection circuit 10 as a circuit block for detecting the occurrence of an achromatic color signal. There is. Details of the achromatic color detection circuit 10 are shown in FIG. The I and Q signals are first squared in squaring circuits 101 and 102, respectively, and then added in an adding circuit 103. The addition output 12+Q2 is compared with a reference voltage in a comparator circuit 104, and if it is less than the reference voltage, it is regarded as an achromatic color, and the gate circuits 20 and 2', which switch the reference run length shown in FIG. 2, are notified of the occurrence of an achromatic signal. Send control signals. register 1
A reference run length for chromatic colors is set in register 8, and a reference run length for achromatic colors is set in register 18'. In addition to the method of detecting the analog signals of the chromaticity signals 1 and Q as described above, the achromatic color signal can be detected by using the digitally encoded output of the Y, I, and Q signals of the analog-to-digital converter 5. It is also possible to adopt a digital detection method in which a comparator is used to detect the digital code of the achromatic signal. The change detection circuit 8 outputs a change detection signal if there is a change in any one of the Y, I, and Q signals.

A comparator 19 compares the run length count value of the run length counter 9 with the run length reference value of the register 18 or 18b, and when the run length count value is greater than the reference value, that is, it is a long run length signal, a logic output is provided.
1'', opens the AND circuit 11, transmits the change detection signal to the long run length signal encoding control terminal L of the run length encoder 7, and controls encoding using the color code.

Conversely, when the run length count value is less than the reference value, the comparator 9 outputs a logic output "O", opens the AND circuit 12, and sends the change detection signal to the short run length signal encoding control of the run length encoder 7. The signal is transmitted to the terminal S to control encoding using the luminance code. The run length counter 9 is reset by the change detection signal via the 0R circuit 15 at the same time as the change detection signal is sent to the run length encoder 7. The run length counter 9 overflows for run lengths exceeding the maximum run length of 255 dots that can be encoded in 8 bits under the code structure shown in FIGS. 1b-d.

The overflow signal is regarded as a change detection signal, and is inputted to the long run length signal encoding control terminal L of the run length encoder 7 via the 0R circuit 13 to control encoding using the color code. In the encoding after overflow, when the overflowed run length is less than or equal to the run length size 127 dots represented by the 7-bit code of the run length code shown in FIG. O” and the output of the AND circuit 20 is logic “1”.
1, and an input is generated at the run-length code control terminal R of the run-length encoder 7. Here, 1Ff' is set to a reference run length of 127 dots in the register, and 21 is 0R.
It is a circuit. At this time, if a change detection signal occurs again at the control terminal L or S, since there is a control signal input at the terminal R,
Encoding is performed using the run-length code shown in FIG. 1d. If the run length is greater than 127 and the output of the comparator 19' is logic "1" at the time when the change detection signal is generated at the control terminal L after overflow, then the AND
Since the output of the circuit 20 becomes logic "01," no input is generated at the terminal R, and encoding is performed using the color code (FIG. 1b) under the control of the change detection signal input to the control terminal L.16. 16' is a register, 17 is a comparator, and 22 is an 0R circuit. These circuits are used to efficiently encode a long run length signal interrupted by a short run length signal using a run length code (Fig. 1d). control.

The register 16 obtains a timing signal from the short run length signal detection output (output of the AND circuit 12) or the long run length signal detection output (output of the AND circuit 11) obtained via the 0R circuit 22, and receives the timing signal at that time. The output signal of the analog-to-digital converter 5 is temporarily stored. On the other hand, the register 16' reads only the long run length signal from the register 16 using the input signal of the long run length signal encoding control terminal L, which is the output of the 0R circuit 13, as a timing signal and stores it for one time. When an input signal occurs at terminal L or terminal S, when the output of the analog-to-digital converter 5 and the content of the long run length signal in the register 16' match, the long run length signal is once interrupted by the short run length signal. After that, it is interpreted that a continuation signal of the long run-length signal has been generated again, and encoding is performed using a run-length code. This control is performed by analog-to-digital converter 5.
The comparator 17 checks whether the output of the register 16 matches the contents of the register 16, and if a match occurs, the output of the comparator 17 becomes a logic
1, the AND circuit 20 is opened via the 0R circuit 21, a control signal is applied to the terminal R, and encoding is performed using a run-length code. Note that when the run length is long and exceeds the set value of the register 18, the AND circuit 20 is blocked by the output of the comparator 19', and no control signal is generated at the terminal R, but only a control signal is generated at the terminal L. Yes, and encoding is performed using a color code. In Figure 2, 1
4 is a buffer memory, and 2T is a large capacity digital memory such as a magnetic disk device. The information encoded by the encoder 7 is once written into the buffer memory 14 and then stored in the digital memory 27. Next, the run-length encoder T will be described in detail according to an embodiment.

FIG. 4 is a diagram showing details of the run-length encoder 7.
The run-length encoder 7 starts encoding in response to a start signal from the control circuit 25 shown in FIG. 2, and ends encoding for one frame in response to an end signal.

The start signal and end signal are generated from the synchronization signal of the synchronization signal generator 2. In FIG. 4, first, the control circuit 2
5 to the gate circuit 701 via the control line 721, the control lines 722, 123, 124
Or according to the signal applied to 725, register TO2,
The contents of IO3 or 704 are prepared so as to be written into buffer memory 14 via signal FfB7l6. Hereinafter, the operation of the run-length encoder 7 will be explained regarding each encoding using a color code, a luminance code, and a run-length code. First, a case will be described in which a long run length signal is encoded using a color code.

The digitally encoded color signals Y, I, and Q are sent to the color code bit editing circuit 7 via the signal lines Tl2, 7l3, and 7l4, and the run length values are sent to the color code bit editing circuit 7 via the signal line 715.
05, performs the bit editing shown in FIG. Now the control line 724 (terminal R
), if there is no control signal (output of AND circuit 20), the change point detection signal output of 0R circuit 13 (Fig. 2) applied to control line 722 (terminal L) is applied to gate circuit 701 via AND circuit TO8. The contents of register 702 are written to buffer memory 14 via signal line 116. Next, a case will be described in which a short run length signal is encoded using a luminance code. The digitally encoded luminance signal Y is inputted via the signal line 712, and the run length value is inputted via the signal line 115 to the luminance code bit editing circuit TO6, and the bit editing shown in FIG. 1c is carried out. Write to register 103. As in the case of color code encoding, if there is no control signal (output of the AND circuit 20) on the control line 724 (terminal R), the change point detection signal output of the AND circuit 12 applied to the control line T23 (terminal S) is applied to the gate circuit 701 via the AND circuit 709, and the contents of the register 703 are sent to the buffer memory 1 via the signal line Tl6.
Written to 4. Finally, a case will be described in which a long run length signal whose run length overflows and a long run length signal interrupted by a short run length signal are encoded by a run length code. signal line 7
The run length value is taken into the run length code bit editing circuit 707 via the bit editing circuit 15, where the bit editing shown in FIG. When the run-length code encoding control signal output of the AND circuit 20 in FIG. 2 is applied to the control line 724 (terminal R), the AND circuit T
IO is open and the control line 722 (terminal L) or the control line T
23 (terminal S) is applied to the AND circuit 7.
10 to the gate circuit 701 and register 7
The contents of 04 are transferred to the buffer memory 14 via the signal line Tl6.
written to. When the encoding is terminated, a termination signal is applied from the control circuit 25 to the gate circuit 701 via the control line 721, and as the gate circuit is closed, run-length encoded information is no longer written to the buffer memory 14. Encoding ends with .

The above-mentioned operation is performed for each scanning line of the color figure FIG. (for example, a digital magnetic disk). In this way, a large number of screens are efficiently run-length encoded and stored in the large-capacity digital memory 27. Here, the second
The handling of the blanking period will be described with reference to FIG.

When the blanking signal BL having a negative polarity is applied from the synchronization signal generator 2 to the 0R circuit 15, the counter 9 is reset, and the BL signal is also applied to the AND circuit 26, so that the clock pulse from the clock oscillator 6 is applied to the counter 9. The output of the counter 9 remains at "O", and "O" is input to the run-length encoder 7 via the signal line 715. The BL signal is also input to the color code bit editing circuit 705 of the run-length encoder 7 via the control line 725 (terminal B), and during the blanking period, the Y, I, and Q signals are all "O". , the sign of "O" for all one word is written to the register 702. Also, BL
The signal is also led to the gate circuit 701 via the control line 725, and when the BL signal is input, the gate circuit 701 is opened and the contents of the register TO2 are written to the buffer memory 14 via the signal line 716. In other words, when the BL signal is input, coded information in which all one word is "0" is written to the buffer memory 14.Next, the color code that is run-length coded and recorded in the digital memory 27 is written to the buffer memory 27. A procedure for decoding graphic information and displaying it on a display terminal will be explained.

FIG. 5 shows an example of this case. In the figure, 30
The buffer memory reads and stores one frame's worth of run-length encoded information from the digital memory 27. The run length encoded information stored in the buffer memory 30 is read by the read control circuit 31, and if the upper two bits are "00", the information is decoded by the color code decoding circuit 32 and becomes "00". The next 6 bits are sent to 33 digital-to-analog converters, and the run length bits consisting of the next byte are 0.
It is sent to counter 35 via R circuit 34. If the upper 2 bits are "01", the information is decoded by the luminance code decoding circuit 36, and the 2 bits representing the luminance Y read as "01" are sent to the digital-to-analog converter 37. The run length bit consisting of the remaining 4 bits in 1 byte is sent to the counter 35 via the 0R circuit 34. If the upper 1 bit is "1", the run length bit is 4 bits.
The information is decoded by the run length code decoding circuit No. 2, and the next run length bit consisting of 7 bits is 0R.
It is sent to a counter 35 via a circuit 34. The counter 35 is a down counter, which is initially set to a value represented by a run length bit, and is decremented by 1 each time a clock pulse is input from the clock oscillator 38 until the zero detection circuit 39 confirms "0". The output of the digital-to-analog converter 33 or 37 is held at a constant value.
Note that the I and Q outputs of the digital-to-analog converter 37 are always 0V. Digital to analog converter 33,3
The output of T is input to the gate circuit 43, and when the output is generated to the color code decoding circuit 32 and the run length code decoding circuit 42, the output of the analog-digital converter 33 is input to the luminance code decoding circuit 36. If so, the output of the analog-to-digital converter 37 is sent to the color encoder 41. When an output is generated in the zero detection circuit 39, this output is sent to the readout control circuit 3 via the 0R circuit 40.
1, the next information is read from the buffer memory 30, and the above procedure is repeated until decoding for one frame is completed.

The color encoder 41 receives Y, I, and Q signals as input,
A color composite video signal (eg, NTSC color signal) is output.

The color encoder 41 is driven by the output signal of 44 synchronizing signal generators. Here, the synchronization signal generator 44 is synchronized with the clock oscillator 38, but it may be asynchronous. The color composite video signal, which is the output signal of the color encoder 41, is recorded in a refresh memory 47, and is repeatedly reproduced to refresh the display terminal 48. Next, decoding during the blanking (BL) period will be described.

45 is a blanking information decoding circuit, and all one word is "0"
Deciphering the BL information that is 〃 produces an output.

This output is applied to the AND circuit 46, and as soon as the BL end signal is input from the synchronization signal generator 44, an output can be generated in the AND circuit 46. At this time, the value of the counter 35 is 0 and no output is generated in the zero detection circuit 39, so the read control circuit 31 does not read the next encoded information from the buffer memory 30. When the BL end signal is input to the AND circuit 46, an output is generated from the AND circuit 46. This signal is input to the read control circuit 31 via the 0R circuit 40, and the run length code information of the next scanning line is read from the buffer memory 30. At the same time, 0R circuit 4
The output of 0 is input to the blanking information decoding circuit 45,
Set the output to "0". Another method of inserting a blanking period is to not insert BL information at the encoding stage, count the run length at the decoding stage, and insert a BL period when a certain value (one horizontal scanning period) is reached. However, any method used does not depart from the scope of the present invention.

In the embodiment, a case has been described in which digitally encoded information is recorded in a digital memory (graphic file device), but of course the present invention is also applicable as a band compression encoding method when transmitting an image using a digital transmission path. can also be applied,
In this case, it is not necessary to treat data in units of bytes, and run-length encoding can be performed with any bit configuration.

As explained above, according to the present invention, for a short run length image signal, only the run length and luminance signal are encoded, and for a long run length image signal, the run length is within the range that can be encoded with a fixed length code. In this case, the run length, luminance signal, and chroma signal are encoded, and if the run length exceeds the range that can be encoded with a fixed length code, or if interrupted by a short run length image signal, the luminance signal in the image signal is encoded. Since only the remaining run lengths are encoded in each case without recoding the signals and chromaticity signals, short run length images can be created by focusing only on the run lengths of individual image signals as in the past. For signals, the run length and luminance signal are encoded, and for the long run length image signal, the run length, luminance signal, and fishiness signal are encoded. It is possible to compress the encoded information for the luminance signal and chroma signal regarding the overflow that occurs in the range that can be encoded by the fixed-length code and the remaining portion that is interrupted by the short run length image signal, making it possible to compress the encoded information for efficient color graphics. It has advantages such as being able to store and transmit information.

[Brief explanation of the drawing]

1A, b, c, d, and e are diagrams explaining the principle of the encoding method according to the present invention, FIG. 2 is a block diagram of an embodiment of the encoding device used in the method of the present invention, and FIG. FIG. 4 is a block diagram of an embodiment of an achromatic color detection circuit, FIG. 4 is a block diagram of an embodiment of a run-length encoder, and FIG. 5 is a block diagram of an embodiment of a decoding device used in the method of the present invention. be.

Claims (1)

[Claims] 1. A color code consisting of a code identification code, a luminance signal, a chromaticity signal, and a run-length code, a luminance code consisting of a code identification code, a luminance signal, and a run-length code, and a code identification code and a run-length code. During encoding, a color image signal obtained by raster scanning a color figure is separated into a luminance signal and a chromaticity signal, and changes in luminance or chromaticity are detected to detect the color image signal. The image signal is decomposed into multiple run length signals, and long run length signals that exceed a preset run length reference value are encoded with a color code, and short run length signals that are less than the run length reference value are encoded with a brightness code. A long run length signal whose run length is too long to be encoded with one color code is encoded by a combination of a color code and a run length code, and a long run length signal interrupted by a short run length signal is reproduced. Regarding the signal, after encoding the short run-length signal using the luminance code, only the run length is encoded using the run-length code, and during decoding, the color code, luminance code, and run-length code are code-identified. Separate and decipher by code,
Encoding and decoding of color graphics characterized by decoding long run length signals from color codes and run length codes, decoding short run length signals from luminance codes, and reproducing color image signals by combining them. Method. 2. Means for sequentially obtaining color image signals by raster scanning a color figure, means for separating the color image signal into a luminance signal and chromaticity signal, and means for converting the separated luminance signal and chromaticity signal into digital signals. a means for detecting a change in the digital signal and counting the run length during that time; and comparing the counted value of the run length with a preset run length reference value to determine whether the color image signal has a long run exceeding the reference value. means for decomposing the long run length signal into a length signal and a short run length signal that is less than the reference value; and means for encoding the short run-length signal with a luminance code consisting of a code identification code, a luminance signal, and a run-length code;
means for detecting that the run length of said long run length signal exceeds a maximum run length that can be encoded by one color code; and means for identifying a long run length signal interrupted by a short run length signal. A color graphic encoding device comprising: 3. A color code consisting of a code identification code, a luminance signal, a chromaticity signal, and a run-length code, a luminance code consisting of a code identification code, a luminance signal, and a run-length code, and a run-length code consisting of a code identification code and a run-length code. Encoding information of color figures composed of codes,
means for decoding the code identification code and separating it into the color code, brightness code, and run length code;
means for decoding a color code into an image signal having luminance and chromaticity; means for decoding the luminance code into an image signal having only luminance; 1. A color graphic decoding device comprising: means for decoding an image signal into an image signal having an image signal; and means for synthesizing the image signals to obtain a color image signal.