SU1709555A2 - Two-dimension adaptive reconstructing filter - Google Patents

Abstract

The invention relates to the field of communications, can be used in the construction of digital television and video television systems, deals with improving the quality of the reproduced image of digital television systems based on pulse code modulation systems and / or differential pulse code modulation using effective discretization structures, in particular a chessboard discretization structure in the field. The aim of the invention is to increase the accurate filtering. The two-dimensional adaptive recovery filter contains a multi-frequency delay line, seven adders 2–4, 13–16, three subtraction blocks 5, 6, and 10. three comparators ?. 8 and 20. Two multiplexers to and 11. Two control signal formers 18 and 21. Shaper threshold 9 and switch 12. The purpose of the invention is achieved by taking into account the sign of the difference between the elements A and C and the result of comparing this difference with the threshold value. 4 il. . ^^ "^^ The 3rd ate ate > & th

Description

The invention relates to communications, can be used in the construction of digital television and video television systems, concerns the issues of improving the quality of the reproduced image of digital television systems built on the basis of pulse code modulation systems or differential pulse code modulation using effective sampling structures, in particular, the chessboard pattern of discretization in the field, and is complementary to the main auth. No. 1438023.

The purpose of the invention is to improve the accuracy of filtration.

FIG. Figure 1 shows a structural electrical circuit diagram of a two-dimensional adaptive recovery filter; 2 shows the structure of the location of samples in the field; fig. 3 shows fragments of the image; figure 4 - the algorithm of the adaptive recovery filter.

The two-dimensional adaptive recovery filter contains a multi-tap delay line 1, first, second and third adders 2-4, first and second blocks 5 and

6readings, first and second comparators

7 and 8, the threshold driver 9, the first and second multiplexers 10 and 11, the switch 12, the fourth, fifth, sixth and seventh adders 13-16, element 14 17, the first driver 18 of the control signal, the third subtraction unit 19, the third comparator 20, second shaper 21 control signals.

Two-dimensional adaptive recovery filter works as follows.

We will first consider the question of the structure of the location of samples, since it is the main function of the proposed filter.

FIG. 2.1 illustrates the orthogonal discretization structure, i.e. the orthogonal location of the samples (in Fig.2.1 and below, the rows of only one field are shown). For the sake of clarity, it is assumed here that one line contains 10 elements, i.e. (for the case, for example, when the sampling frequency of fg8 MHz elements).

Figure 2.2 shows the checkerboard structure of samples in the field, with the same number of elements. It should be noted that in spite of the fact that in adjacent rows the number of elements does not match, with a chessboard discretization structure in the field it is customary to assume PS is the maximum number of elements in a row, i.e. for our specific example, Ps ° Psmaks 10. As you know, chess

a sample structure in the field is created if

sequences of sampling pulses change the phase by 180 from line to line and from frame to frame. The operation of the proposed filter is implied with such a sampling structure, therefore, a digital video signal is supplied to the information input of the filter, for example, represented by an eight-bit binary code with such a sample positioning structure. Fig. 2.2 also shows the counts A, B, C, D analyzed in the filter. In order for the two-dimensional adaptive recovery filter to produce different arithmetic, operations (cy 4, rotation, division, etc.) over the values of these elements their simultaneous presence, for this it is necessary to delay the elements D, C and A with respect to element B and, as can be seen from FIG. 2.2, the element D needs to be delayed (at) 9 elements, i.e. on Ps-1 the number of elements, the element C on 10 elements, i.e. on Ls the number of elements and A on 19

elements, i.e. on 2pe-1 the number of elements, which ensures the dividing line 1 of the delay.

Signals from the input (B) and outputs (O, C, A)

digital delay lines are fed to the inputs from the first to the seventh adders 2-4 and 13-16. The obtained values at the outputs of the first adder 2 - (D - "- C) / 2, the second adder 3 - (B + A) / 2 and the third adder 4 (A + B + C + D) / 4 are fed to the inputs of the first multiplexer 10 , and the values from the outputs of the fourth adder 13 - (AU) / 2, the fifth adder 14 - () / 2, the sixth adder 15 - () / 2 and the seventh adder 16 (C + A) / 2 are received at the inputs of the second multiplexer 11, the values of elements A, B, C, D are also received respectively at the inputs of the first, second and third subtraction blocks 5, b and 19, where operations are performed

C-D, AB and AC subtractions. Information about the magnitudes of the differences comes from the first, second, and third blocks 5, b, and 19 of the subtraction, respectively, to the first, second, and third comparators 7.8 and 20, where they are compared with the threshold value Rn, coming from the threshold 9. The calculation of these differences and the comparison with the threshold value (i.e., threshold detection) is done in order to detect

on the studied area of contour transitions. The choice of the value of Rn is determined, on the one hand, by the necessity of detecting small luminous transitions, on the other hand, by reducing the probability of detection

spurious transitions due to noise. The optimal value of Rn is approximately 12% of the total dynamic range of the signal brightness.

Information about the signs of the differences CD and AB from the first and second blocks 5 and b of the subtraction is fed to the inputs of the first driver 18 of the control signal, and information about the sign of the difference A – C and the results of comparing the differences A – C, A – B and CD with the threshold Rn values come from blocks 19, 20, 7 and 8, respectively, to the second control signal generator 21, which provides control (operation) of the operation of the first multiplexer 10, where, depending on the nature of the image, i.e. Luminance distributions may consider the following possible cases (Fig. 4).

I If (A-B) Rn (C-D) Rnj

This section of the image is taken to be equal to the capacitive or one-element capacitive transition, and for more accurate perception and reproduction of the differential, the difference A – C and its sign indicator sign (A – C) are introduced into the analysis. We introduce the notation at A C sign (A – C) 1, and at A C sign (A – C) 0. Under the condition (A-B) Rn and (C-D) R, the generation of the control signal of the blocks 21 depends on the ratio of the difference A-C to the threshold Rn-and its sign (A-C). If conditions (A-B) of Rn and (C-D) R are not satisfied: the values of (A-C) and sign (A-C) are not taken into account. The second control signal generator 21 makes it possible to distinguish equally the focal areas (Fig.3.5) from the vertical and horizontal one-element drops (Fig. 3, 18 and 3, 19), respectively, to correctly reproduce them.

The algorithm of operation in this case looks as follows.

If (A – C) R p

sign (A – C) 0,

This section of the image is taken as equal to the capacitance and the output of the third adder 4 - (A + B + C- -D) / 4 is passed to the output of the first multiplexer 10 of the blocks 21, and the element 17, in turn, produces a control signal the output of the first multiplexer 10 passes to the output of the second multiplexer 11 (fig.Z, 5).

Ec (A-C) Rn sign (A-C) T,

then this section of the image is also taken as equal to the capacitance and at the output of the second multiplexer 11 control signals C of block 21 and AND 17

5, the output of the third adder 4 (A + B + C + D) / 4 (FIG. 3.5) is passed. If (A – C) Rn (A – C) 0, then this portion of the image is taken

10, a single-element horizontal differential and, at the output of the second multiplexer 11, the control signals from block 21 and element 17 pass the output of the second adder 3 - (A + B) / 2 (FIG. 19).

15 If (A – C) Rn

sign (A – C) 1,

then this image section is taken as a single-element vertical differential and, at the output of the second multiplexer 11, the control signals from block 21 AND of the element 17 pass the output of the first adder 2- (C + D) / 2 (FIG. 18) ./

ii. If (A-B) Rn (C-D) Rn,

5 then this section of the image is taken as a horizontal differential and at the first and second control inputs of the first multiplexer 10 a control signal is formed, which at the output of the first multiplexer 10 passes the output of the second adder 3 - (B + A) / 2, and And 17, in turn, it generates a control signal that the output of the first multiplexer 10 transmits to the output of the second multiplexer 11 (Fig.3-3.4, b, 7). HI. If (AB) Rn

(C-D) Rn,

then this section of the image is taken as the vertical drop on the control

0 inputs of the first multiplexer 10, a signal is formed, which passes the output of the first adder 2 - (D + C) / 2 to the output of the first multiplexer 10, and the element 17, in turn, generates a control signal that the output of the first multiplexer 10 also passes on the output of the second multiplexer 11 (figure 3-1,2,8.9). Iv. If (A-B) Rn (C-A) Rn,

This section of the image is assumed to be diagonal or angular, and, for a more accurate perception and reproduction of the direction of the differential, the signs outlined above are also taken into account.

5 In this case, the And element 17 generates a control signal locking the output of the first multiplexer 10 at the input of the second multiplexer 11 and allowing the first driver 18 of the control signal, taking into account the signs of differences (AB) and

(C-D), i.e. sign (A-B) and sign (C-D), control the passage through the second multiplexer of 11 outputs of the fourth 13, fifth 14, sixth 15 or seventh 16 adders.

We introduce the notation; at A B sign (A-B) 1, at A B sign (A-B) 0 and respectively at C D sign (C-DhI, and at C D slgn (C-DhO.

In a word, with (A – B) Rn and (C – O) Rn, the sign indicators of these differences enter into the analysis, where four cases are possible.

1.If sign (

stgn (C-D) 0.

then the output of the second multiplexer 11 receives the output of the n adder 14 (D + B) / 2. (Fig.3-13, 14).

2.If sign (А-В) 0

slgn (

then the output of the second multiplexer 11 receives the output of the sixth adder 15 (B + C) / 2 (figure 3-10,16).

3.If sign (А-В) 1

slgn (C-D) 0,.

then the output of the second multiplexer 11 receives the output of the fourth adder 13 (A + O) / 2 (fig.Z-12,17).

4.If sign (А-В): 1

. sign (C-D) 1,

then the output of the second multiplexer 11 receives the output of the seventh adder 15 (C- | -A) / 2 (FIG. 3-11.15).

The sampling frequency in the image signal at the output of the filter is doubled, as in the prototype, and takes the form shown in Fig. 2.3. Each second element (shown by dashed lines) is an interpolated element X reconstructed from four adjacent elements (highlighted by a diamond ), therefore, the need arises to temporarily separate the elements C and X located adjacent to the line. This is accomplished in switch 12, the control input of which is supplied with sampling pulses of duration equal to half

period, i.e. Doubled 2iff sample rate. The operation of the switch 12 in this case is identical to the operation of the switch in the prototype.

Thus, the distortions introduced by

by discretization with the chess structure in the field, are eliminated almost completely, and the algorithm of the filter allows to correct not only horizontal, vertical, diagonal and angular, but

and single-element drops in brightness. This makes it possible to significantly improve the subjective quality of the reproduced image and to improve the objective indicators of such a filter during the transmission of test signals, while the two-dimensional adaptive recovery filter becomes only slightly more complicated.

Claims (2)

Invention Formula Two-dimensional adaptive recovery filter on. Avt.st. No. 1438023, which means that, in order to increase the filtration accuracy, the third unit connected in series is introduced subtracting, the first and second inputs of which are connected respectively to the second and third outputs of the multi-split delay line, the third comparator, the second input of which is connected to the output of the imaging unit the threshold, and the second driver of the control signal, the second input of which is connected to the second output of the third subtraction unit, the outputs of the first and second comparators being connected via the third and fourth the inputs of the second control signal generator with the first and second control inputs of the second multiplexer, о о 00 9 QQ о о о о 5f3g-о о о о о о С С С O о о о о о о о о о о о о о oh oooooo 3. 3. o 9 99 o o 9 9 99 9 o-ee-eig o o o o o o 2.. ;,; C) O o o o o o o o o o o o o 6 o o o o o o o o o o o o o o o o o o o o o o o o o o o o o oa o 6 CO 1: O. , in h .--. , r / rr.r,) dsOOO; with Fig.Z .. 3g. Oh oh zk - arbitrary cocTO: HU € FIG.