JPH01140881A - Picture information transmitter - Google Patents

Abstract

PURPOSE:To simplify a signal processing in a decoding system, and to transmit picture information through less transmission quantity by selecting optionally plural ways of coding by considering a dynamic range, and making the code length of a transmission unit constant. CONSTITUTION:An NTSC television signal is inputted to a blocking circuit 12 through an A/D conversion circuit 11. A sample train divided into blocks and rearranged by the blocking circuit 12 is impressed to a minimum value detector 14, a maximum value detector 16, an intra-block mean value calculation circuit 18 and a delay circuit 20. The outputs of those are inputted to a conversion circuit 32, the conversion circuits 36, 38 and a data train generation circuit 44 through a subtracter 22 and the delay circuits 24, 26, 28, 30, etc. A selection circuit 42 selects the output of either of the delay circuit 24 or a parallel/serial conversion circuit 40. The index of each picture element is outputted from the conversion circuit 32, and the encoded data train of fixed length is outputted from the data train generation circuit 44.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image information transmission device that encodes and transmits image data in blocks.

[Conventional technology]

When transmitting digital image data, for example, a digitized television signal, to another device via a transmission path, the image information data has a large amount of information, so there is some method that can be used to transmit it. It is necessary to compress the amount of information. One such compression method is a block code or transmission method that blocks one screen's worth of image information into a predetermined number of samples, and compresses and encodes the samples that make up each block. It is being

Specifically, the maximum and minimum values of the samples belonging to the block of interest are detected, the dynamic range between the minimum value and the maximum value' is linearly quantized, and the quantization level to which each sample belongs is determined. Calculate the index and send the maximum value, minimum value and the corresponding index. Since the number of bits of the index can be significantly smaller than the number of bits of the original sample, the amount of transmission can be compressed.

[Problem that the invention seeks to solve]

In this conventional block encoding method, linear quantization is performed based on the dynamic range of the sample values within the block, so when the dynamic range is narrow, it can be encoded with high accuracy. If the dynamic range is very wide, such as when there is brightness or darkness or a block contains boundaries, the encoded data will have very poor accuracy, causing large distortions in the restored image.

To solve this problem, when the dynamic range is wide, the average value of the sample values is detected, and the linear quantum is divided equally between the average value and the maximum value and between the average value and the minimum value. A possible method is to adaptively switch between linear quantization between the maximum value and minimum value, and linear quantization that takes this average value into account, depending on the dynamic range. However, with this method, there are cases where average value data exists in the transmission data string and cases where it does not exist.
This results in a variable length data string. For example, when trying to digitize a television signal and record it on a video tape recorder, handling variable-length data strings requires real-time processing, especially playback processing that is difficult and requires a large amount of memory. This method is not suitable for transmission systems such as recording/reproducing systems because it not only requires a configuration like this, but also causes difficulties in signal processing during special playback (eg, search playback, slow playback).

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide an image information transmission device that takes into consideration the distribution of sample values within a block and that provides a fixed length data string.

[Means for solving problems]

The image information transmission device according to the present invention is a block that divides digital image data into blocks each consisting of a predetermined number of samples.
A locking means, a first calculation means for calculating two reference values regarding the dynamic range for each block, a second calculation means for calculating distribution situation data indicating a distribution situation of sample values within the block, and the first calculation means. a first encoding mode for encoding samples within a block based on two reference values of the first calculation means; and encoding samples within the block based on the two reference values of the first calculation means and distribution status data of a second calculation means; a second encoding mode, and encoding means for encoding each sample in either of the two encoding modes according to the dynamic range of each block; and an identifier generating means for generating an identifying identifier, and in the first encoding mode, the identifier, the two reference values by the first calculation means, and the encoded data of each sample are used as transmission units, and the second encoding In this mode, the transmission unit is the identifier, one reference value of the first calculation means, compressed data of the other reference value, distribution status data, and encoded data of each sample, and both transmission units have the same length. It is characterized by

[Effect]

Encoding is performed by selecting one of two types of encoding depending on the dynamic range, but no matter which encoding is adopted, the transmission code length is kept constant, so decoding Processing becomes easier.

Further, by adding an identifier indicating which encoding is used, even if both encodings are mixed in one screen of image information, it can be easily distinguished at the time of decoding. In other words, 1
When screen information is divided into blocks, appropriate encoding can be arbitrarily selected according to the dynamic range of each block, and the amount of transmission can be compressed more efficiently.

〔Example〕

Hereinafter, the present invention will be described in detail with reference to the drawings.

FIG. 1 shows a block diagram of a television signal encoding apparatus as an embodiment of the present invention. In Figure 1, 1
0 is an input terminal for a digital image signal. The television signal used in this embodiment is based on the NTSC system, and the NTSC television signal is input to the input terminal 10. Reference numeral 11 denotes an A/D converter that samples the television signal input to the input terminal 10 at a sampling frequency four times the carrier frequency Esc, and converts it into a digital television signal of 8 bits per sample. 12 stores the digital television signals that are input sequentially in units of horizontal scanning lines in a memory, etc., and reads them out after changing the sequential order, thereby editing the scanning order in units of horizontal scanning lines within a block and outputting the same. In the illustrated embodiment, one block consists of four samples in the horizontal direction and four lines in the vertical direction.

The sample sequence rearranged by the blocking circuit 12 is applied to a minimum value detector 14, a maximum value detector 16, an intra-block average value calculator 18, and a delay circuit 20. The average value calculation circuit 18 includes an integration circuit that integrates sample values for one block, a division circuit that divides the integrated value of the integration circuit by the number of sample points, that is, 16, and a hold that holds the division result of the division circuit. It consists of a circuit and outputs the average value m0 of sample values for one block. The subtractor 22 calculates the difference between the maximum value MAX detected by the maximum value detector 16 and the minimum value MIN detected by the minimum value detector 14.

Output MIN of minimum value detector 14, output O of subtractor 22
R(・MAX-RIIN), output MA of maximum value detector 16
X and the output m0 of the average value detector 18 are applied to the conversion circuit 32 via delay circuits 24, 26, 28, and 30, respectively.

The sample value DT formed by the A/D converter 11 is also input from the delay circuit 20 to the conversion circuit 32, and MAX,
Index ID that quantizes DR based on MIN, DR, mo and indicates which quantization step DT belongs to.
Output. The details will be described later. Note that the delay circuit 24.
26, 28, and 30 are for timing adjustment with the sample value DT applied from the delay circuit 20 to the conversion circuit 32.

In this embodiment, as a method of encoding a digital television signal, the dynamic range DR in one block of sample values DT is divided into eight equal parts as shown in FIG. 6(a), and each sample value DT is A 3-bit index indicating which divided area of the dynamic range it belongs to10. And the above MAX and MIN data are each 8 as information representing the dynamic range of each block.
MAX, instead of the sample value DT.
The first operating mode outputs MIN, ID+ and the dynamic range DR as shown in Figure 6 (bl).
Divide the space between rN and mo into four equal parts. and MAX, and a 3-bit index 10. which indicates which divided area of the dynamic range each sample value DT belongs to.
Then, as information representing the dynamic range of each block, the MAX data is formed with 8 bits, and the MIN and me data are each formed with 4 bits, and instead of the sample value DT, the MAX data is formed with 8 bits, and the MAX data is formed with 4 bits each.

A second operation mode that outputs MINI 1110+ 1Dz is adaptively switched according to the dynamic range DR.

34 is an identifier generation circuit which receives the dynamic range DR from the delay circuit 26 and outputs an identifier for identifying the operation mode. Specifically, the darkness value for operation mode selection and D
When DR is less than or equal to the threshold value, an identifier MD1 specifying the first operation mode is output, and in other cases, an identifier Moz specifying the second operation mode is output.

Conversion circuit 36, 38 and parallel/serial conversion circuit 4
0 is a circuit that functions when the second operation mode is selected. Specifically, the conversion circuit 36 converts 8-bit MIN into 4-bit MIN'', and the conversion circuit 38 converts 8-bit MIN'' into 4-bit MIN''.
o is a 4-bit m. ”. Such conversion is configured, for example, by a memory table method that outputs 4-bit data corresponding to 8-bit input data.
6 and the 4-bit signal from the conversion circuit 38 in serial form.
In other words, it is converted into an 8-bit signal. The selection circuit 42 selects the MI from the delay circuit 24 when selecting the first operation mode.
N (8 bits) is selected, and when the second operation mode is selected, the 8-bit signal from the parallel-to-serial conversion circuit 40 is selected. - MAX from the delay circuit 28, the identifier MD from the identifier generation circuit, the selection output of the selection circuit 42, and the index ID from the conversion circuit 32 are applied to the data string forming circuit 44. The data string forming circuit 44 converts these input signals into a serial data string and outputs the serial data string. Figure 5(a)
indicates the output data string when the first operation mode is selected;
) indicates an output data string when the second operation mode is selected. In this way, in either operation mode, the data string forming circuit 44
Since the length of the data strings output from the two is the same and fixed length, the signal processing can be performed with a simple configuration when transmitting, recording and reproducing the output data strings. In particular, in the case of recording and reproduction, special reproduction and the like can be performed easily.

FIG. 2 shows a specific example of the configuration of the conversion circuit 32.

Input signals MIN, MAX, DT are input to the conversion circuit 50.
52. The average value m0 is applied to the conversion circuit 52.

Output and dynamic range D of conversion circuit 50.52
R is applied to the output selection circuit 54. The conversion circuit 50 performs encoding in the first operation mode described above, and the conversion circuit 52 performs encoding in the second operation mode.

Each outputs 3-bit index ID+ and IDz. The output selection circuit 54 selects and outputs the output of the conversion circuit 50 or 52 according to the dynamic range DR under the same conditions as when the identifier generation circuit 34 generates the identifier.

FIG. 3 shows a specific example of the conversion circuit 50. The subtracter 72 is M
Subtract MIN from AX and output dynamic range DR. This dynamic range is divided by divider 74 to 1
78 and then directly to the comparator 81 and also to the multiplier 75.
-80 to comparators 82-87. Multipliers 75 to 80 multiply the input signal by 2 times, 3 times - 7 times, respectively. A subtracter 88 subtracts M from the sample value DT in the block.
Subtract IN and the subtraction result DT'(・DT-MIN)
is applied to the other inputs of comparators 81-87. Each comparator 81-87 prioritizes comparison results 01-C7.
The signal is supplied to the encoder 90. Output DT' of subtracter 88
01 to C7 are as follows according to the following.

(1) 0≦DT'<(1/8) DR C1=C2=C3=C4・C3=C6・C7・0(2)
(1/8) DR≦DT″ < (2/8) DRCI
-1, C2=C3=C4, C54C6=C7=0(31
(2/8) DR≦DT'< (3/8) DRCl・C2
=1, C3・C4・C3=C6=C7=0(4)
(3/8) DR5DT'< (4/8) DRC1=C
2=C3・1. C4=C5=C6=C7・0+51
(4/8) DR≦DT'< (5/8) DRCl・C
2・C3・C4=1, C5・C6=C7・0(6)
(5/8)DR≦DT'<(6/8)DRC1=
C2・C3・C4=C5・1. C6=C7=0(7)
(6/8)DR≦DT'<(7/8)DRCl,
C2・C3・C4・C3=C6=1, C7=0(8
) (7/8) DR≦DT” <DR.

C1・C2・C3・c4=c5=c6=c7・1 condition (
11, (2), (3), (41, (5), (61
, (7), and (8) are respectively areas AI+Affi+Aff and A4+A! in FIG. 6(a). l+A
6°A? , indicates A8. The priority encoder 90 outputs (000) at (1) above and (000 at (21)) above.
001), (3) (010), (4) (011), (51 (100), (61 (101), (7) (110), (8) The 3-bit index ID of time (111) is output.

FIG. 4 shows details of the conversion circuit 52. A subtracter 100 subtracts MIN from the sample value DT, and a subtracter 102 subtracts MIN from the sample value DT.
MIN is subtracted from MAX, and the subtracter 104 subtracts mo from MAX. The output (me-old N) of the subtracter 102 is divided into 1/4 by the divider 106, and then output to the multipliers 107 and 108.
and is applied to comparator 110. Multipliers 107, 108
respectively double and triple the input signals and send them to the comparators 111.
112. Output of subtracter 104 (MAX-mo
) is divided into 174 by the divider 114, and then the multiplier 115 .
It is applied to IL6 and adder 1]7. multiplier 115,
116 doubles and triples the input signal, respectively, and adds
8.119 feed. Adders 117, 118, 119
The output of the subtracter 102 is applied to the subtracter 102, and the addition result is applied to one input of the comparators 120, 121, 122, and 123. Comparators 110-112, 120-12
3, the output DT' (・
DT-formerly N) is supplied.

The outputs of comparators 110 to 112, 120 to 123 are D1 to
D7, the following D1 to D7 are input to the priority encoder 124 for each case. That is, (1) 0≦DT'< (1/4) (me-MIN)
Dl・D2・D3・D4・D5・D6=07=0(21
(1/4)(mo-old N)≦DT'<(2/4)(mo
-MTN)D1=1. D2=03=04=05=06
=07=0(31(2/4)(mo-MIN)≦DT'
<(3/4)(mo-MIN)D1=02=1, D
3・D4=D5=06・Dl・0(4) (3/4)
(me-MIN)≦DT'<m6-MINDl・D2
=03・1. D4・D5・D6=D7=0(5) m
o −M T N ≦DT'<(1/4) (MAX
-mo)+(no-MIN)D1=02=03=04=
1, D5・D6=D7・0(6) (1/4)(
MAX-me)+ (mo-MIN) ≦DT'
<(2/4) (MAX-mo) + (me-old N) DI=02・D3=04=05=1. D6・Dl・0
(7) (2/4) (MAX-mo) + (mo
-MIN) ≦DT'< (3/4) (MAX - custom) + (mo - MIN) DI = 02 = 03 = 04 = 05. D6=1, D
7=0(8) (3/4)(MAX-mo)+(n
+(1-MIN) ≦DT'<MAX-MIND1
=02=03=04=05=06・D7=1 Condition (1)
, (2), (3), (4), (5), (6)
, +7), and (8) are respectively in areas B+, Bz, and B3. in FIG. 6(b). B4. Bs, Bb, Bt, Bs are shown. The priority encoder 124 allocates the same three bits in each case as in the conversion circuit 50;
The index 102 is output.

FIG. 7 shows a schematic configuration of a digital television signal decoding apparatus corresponding to the encoding apparatus of FIG. 1. A digital television signal input from a transmission path (not shown) is sent from an input terminal 200 to a mode identification circuit 202, a maximum value/minimum value separation circuit 204, and a maximum value/minimum value/average value separation circuit 2.
06 and the index separation circuit 20B. The mode identification circuit 202 separates the mode identifier MD, and the maximum value/minimum value separation circuit 204 separates the maximum value MAX and minimum value MIN when the identifiers from the mode identification circuit 202 are between (first operation mode). Separate and output the maximum value
The minimum value/average value separation circuit 206 is the mode identification circuit 20
If the identifier MD force from 2 < rIDt (second operating mode), the maximum value MAX and the 4-bit minimum value old N
The index separation circuit 208 separates and outputs the index part.

Since the data strings supplied to each separation circuit are fixed-length codes for each block, their separation is extremely easy. Note that a clock generation circuit (not shown) is provided for separation between these.

The decoding circuit 210 includes separation circuits 202, 204, 206, .
Each sample value DT from each data separated by 208
is decoded and supplied to the D/A converter 212. The D/A converter 212 uses the sampling frequency 4fSc as a synchronizing signal to convert the decoded sample value DT into an analog signal, thereby restoring and outputting the NTSC television signal. Details of the decoding circuit 210 are shown in FIG. 2
20 is a restoration circuit that restores each sample value from the index ID in the first operation mode, and 222 is the index 10. in the second operation mode. A restoring circuit 224 for restoring the encoded signal that recovers each sample value from the restoring circuit 224 selects either the output DT of the restoring circuit 220 or the output DT of the restoring circuit 222 according to the mode identifier MD. It is a circuit.

FIG. 9 shows a configuration example of the restoration circuit 220. Subtractor 226
calculates DR (=MAX-MIN) and supplies it to the divider 228. The divider 228 makes the input 178 and applies it to the multipliers 230 to 235 and the selection circuit 238. Multipliers 230 to 235 multiply the input signals by 2, 3, 4, 5, 6, and 7, respectively, and apply the multiplied signals to the selection circuit 238. As a result, each boundary value when the dynamic range OR is divided into eight equal parts is input to the selection circuit 238. “0” is also input to the selection circuit 238 from the terminal 237.

The selection circuit 238 selects and outputs one of the eight inputs according to the index 10° of each sample. The relationship is shown in Table 1.

Divider 240 halves the output of divider 228 and outputs DR/
16, and an adder 242 adds the output of the divider 240 to the selected output of the selection circuit 238 to obtain a representative value. Adder 242 further adds MIN from delay circuit 241. The delay circuit 241 is for timing adjustment. The output of adder 242 is supplied as restored sample value DT to selection circuit 224 (FIG. 8).

FIG. 10 shows details of the restoration circuit 222. Conversion circuit 24
3 consists of a memory table, etc., converts 4-bit MIN' into 8-bit MIN, and converts it to a conversion circuit 245.
is also composed of a memory table, etc., and is a 4-bit m.゛ is an 8-bit m. Convert to Subtractor 244
is m. and the old N, and the subtracter 246 calculates the difference between MAX and m
Calculate the difference of o. The divider 248 sets the output (mo-old N) of the subtracter 244 to 178, and multipliers 249 and 250
, 251 and the selection circuit 252. Multiplier 249
, 250, and 251 multiply the input signal by 3, 5, and 7, respectively.
The multiplied signal is multiplied and supplied to the selection circuit 252. The divider 254 makes the output (MAX-mo) of the subtracter 246 178 and applies it to the multiplier 255°256.257 and the adder 258. Multipliers 255, 256, and 257 multiply the input signal by 3, 5, and 7, respectively, and adders 259, 260° 261
to be applied. Each adder 258 to 261 adds the output (mo-MIN) of the subtracter 244 and supplies the result to the selection circuit 252. The selection circuit 252 selects the index ID of each sample.
2, one of the eight inputs is selected and output. The relationship is shown in Table 2.

The output of selection circuit 252 is applied to adder 262. The adder 262 is connected to the delay circuit 264? IIN is applied, and the result of addition of both signals is applied as a restored sample value DT2 to the selection circuit 224 in FIG. 8.

As explained above, in this embodiment, by encoding the television signal, which is a type of image information, in the two types of operation modes described above, even if the amount of information is small, it is possible to improve the image quality etc. It can be encoded and transmitted in a state with less deterioration, and when encoding is performed based on the two types of operation modes as described above, the unit code length of the data string obtained by the encoding changes depending on the type of operation mode. In addition, even if data strings based on the two types of operation modes are mixed in the data string, it is possible to reliably determine in which operating mode the transmitted data string is encoded. can be identified and restored to the original signal.

Table 1 Although the explanation has been given by taking the encoded transmission of an NTSC television signal as an example, the present invention is not limited to this.
Similar effects can be obtained even when applied to a coding transmission device for television signals of the AL system or SECAM system, or a transmission device for image signals other than television signals, such as facsimile signals.

Table 2 [Effects of the Invention] As can be easily understood from the above explanation, according to the present invention, not only can a plurality of encodings be arbitrarily selected in consideration of the dynamic range, but also the code length of the transmission unit can be kept constant. Therefore, signal processing in the decoding system becomes extremely simple.

Therefore, image information can be transmitted (recorded and reproduced) with a smaller amount of transmission.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of the configuration of an encoding device as an embodiment of the present invention, FIG. 2 is a detailed diagram of the conversion circuit 32 shown in FIG. 1, and FIGS. 3 and 4 are the conversion circuits shown in FIG. circuit 50.5
2, FIG. 5 is an example of the output code string of the data string forming circuit 44 of FIG. 1, and FIG. 6 is a detailed diagram of the conversion circuit 50, 52 of FIG. 2.
Fig. 7 shows an example of the quantization division in
The figure shows a detailed example of the decoding circuit 210 in FIG. 7, FIG.
0 are detailed examples of the restoration circuits 220 and 222 of FIG. 8, respectively. 14-- Minimum value detector 16- Maximum value detector 18-
Intra-block average value calculator 32-conversion circuit 34...
Identifier generation circuit 42-Selection circuit 44-.-Data string formation circuit

Claims (1)

[Claims] a blocking means for dividing digital image data into blocks each consisting of a predetermined number of samples; a first calculating means for calculating two reference values regarding dynamic range for each block; and a distribution state indicating a distribution state of sample values within the block. a second calculation means for calculating data, a first encoding mode for encoding samples within a block based on two reference values of the first calculation means, two reference values of the first calculation means and a second calculation; and a second encoding mode that encodes the samples in the block based on the distribution situation data of the means, and encodes each sample by either of the two encoding modes depending on the dynamic range of each block. and an identifier generating means for generating an identifier for identifying the encoding performed by the first encoding means,
In the encoding mode, the identifier is 2 by the first calculation means.
The encoded data of one reference value and each sample is used as a transmission unit, and in the second encoding mode, the identifier, the first
one reference value of the calculation means, compressed data of the other reference value,
An image information transmission device characterized in that distribution status data and encoded data of each sample are used as a transmission unit, and both transmission units have the same length.