JP2006246400A - Video data repairing apparatus and method, and video compression coding system employing the same - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To select a repair source line relative to an error occurrence line in accordance with the moving state of video. <P>SOLUTION: When an error is detected in video data by a CRC error detection part 2, predetermined interpolation data and selection coefficients K1-K3 indicating the degree of similarity between the interpolation data and the original video data are outputted, respectively from a line interpolation part 20, a field interpolation part 40 and a frame interpolation part 30. In an interpolation data selection part 60, the selection coefficients are compared with one another to decide which interpolation data are closest to the original video data and the interpolation data are selected. When no CRC error is detected, a video data selection part 3 selects an input video data side and when the error is detected, the video data selection part selects (interpolates, substitutes) an interpolation data side. A CRC code re-calculation part 4 re-calculates CRC of the replaced video data and outputs the video data to a video compression coding device 5. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a video data restoration apparatus and a restoration method, and more particularly to an apparatus and method for performing error restoration of video data input to a video compression coding apparatus.

  In a video compression encoding apparatus for transmitting video data, video data having an error cannot be encoded correctly. When compression encoding is performed using the video data in error, the peripheral video data is affected, and the image quality is deteriorated. For this reason, the video compression coding apparatus performs error detection for each piece of data for one horizontal scanning line (1 line: 1H) for the input video data. Throw away and encode. As a result, encoding is performed with the data size in the vertical direction when the screen is displayed smaller than the original video data.

  In order to perform compression coding without changing the data size in the vertical direction and without degrading the image quality, the data for one line in which an error is detected can be restored and then input to the video compression coding apparatus. That's fine. For this reason, various techniques for repairing data by replacing error detected line data with interpolation data generated by referring to other related line data have been proposed.

  For example, Patent Document 1 discloses interpolation data of a missing line as one or two lines that are separated in time (usually adjacent fields) but are close as vertical positions on the screen (usually, A technique for calculating from data of adjacent lines) is described.

JP 2003-264717 A (page 1-2, FIG. 4)

  When repairing an error occurrence line, another related line that is a reference source (restoration source) may be an adjacent line in the case of a video with motion, but an adjacent frame in the case of a video with little motion. The same line is more suitable. That is, when the optimum interpolation data (closest to the original video data) for the error occurrence line is to be generated, the time and position of the reference source line of the interpolation data for the error occurrence line according to the motion state of the video data. It is necessary to make the target relationship variable.

  However, with the technique described in Patent Document 1, interpolation data can be generated only by lines that have a preset temporal and positional relationship with respect to an error occurrence line. In other words, the technique described in Patent Document 1 cannot always generate optimal interpolation data according to the motion state of video data.

  SUMMARY OF THE INVENTION An object of the present invention is to provide a video data restoration apparatus and restoration method capable of always generating optimum interpolation data for error restoration (closest to the original video data) according to the state of motion of the video data, and the same An object of the present invention is to provide a video compression encoding system using the above.

  According to a first aspect of the present invention, there is provided a video data restoration device that receives video data comprising a predetermined unit of data sequentially and recovers video data having an error, for each predetermined unit. Error detection means for performing error detection of video data, interpolation data generation means for generating a plurality of interpolation data based on a plurality of unit data having a predetermined temporal and positional relationship with the error detection target unit data, and Among the plurality of interpolation data generated by the interpolation data generation means, there is provided data selection means for determining the interpolation data closest to the original video data and replacing the video data in which an error is detected with the corresponding interpolation data.

  According to a second aspect of the present invention, there is provided a video data repairing apparatus for receiving video data comprising a predetermined unit of data successively in sequence, and repairing video data having errors in each predetermined unit. From error detection means for detecting an error in video data, first interpolation data generation means for generating interpolation data from unit data before and after error detection target unit data, and unit data in a field different from the field of error detection target unit data The second interpolation data generation means for generating the interpolation data and the third interpolation data generation means for generating the interpolation data from the unit data of a frame different from the frame of the error detection target unit data are arbitrarily selected. Output from the interpolation data generation means and the arbitrarily selected interpolation data generation means, respectively. Among between data, it determines interpolation data closest to the original image data, and a data selection means for replacing the video data error is detected in the relevant interpolation data.

  According to a third aspect of the present invention, there is provided a video data restoration device comprising: an error detecting means for detecting an error of input video data; a first line memory for delaying input video data by one line and two lines; A second line memory; a field memory for delaying input video data by one field—one line; a third line memory for delaying an output of the field memory by one line and two lines; Line memory, line interpolation data generating means for generating line interpolation data from the input video data and video data delayed by two lines output from the second line memory, and the third line memory 1 field delayed video data and 1 field output from the fourth line memory Field interpolation data generation means for generating field interpolation data from video data delayed by +1 line, and video delayed by one line output from the first line memory if no error is detected by the error detection means When the data is normal video data and an error is detected by the error detection means, the interpolation data closest to the original video data is determined from the line interpolation data and the field interpolation data, and an error is detected with the corresponding interpolation data. Data selection means for replacing the video data.

  According to a fourth aspect of the present invention, there is provided a video data restoration device comprising: an error detection means for detecting an error of input video data; a first line memory for delaying the input video data by one line and two lines; A second line memory; a frame memory that delays the input video data by one frame; a third line memory that delays an output of the frame memory by one line; the input video data; Line interpolation data generating means for generating line interpolation data from the video data delayed by two lines output from the line memory, and frame interpolation from the video data delayed by one frame + 1 line output from the third line memory If no error is detected by the frame interpolation data generation means for generating data and the error detection means When the video data delayed by one line output from the first line memory is set as normal video data and an error is detected by the error detection means, the most original video among the line interpolation data and the frame interpolation data is detected. A data selection unit that determines interpolation data close to the data and replaces video data in which an error is detected in the interpolation data;

  According to a fifth aspect of the present invention, there is provided a video data restoration apparatus, an error detection means for detecting an error of input video data, a first line memory for delaying the input video data by one line and two lines, and A second line memory; a field memory for delaying input video data by one field—one line; a third line memory for delaying an output of the field memory by one line and two lines; A line memory, a frame memory that delays input video data by one frame, a fifth line memory that delays output of the frame memory by one line, the input video data, and the second line memory Interpolation data for generating line interpolation data from video data delayed by 2 lines output from A field for generating field interpolation data from the generating means, video data delayed by one field output from the third line memory, and video data delayed from one field + 1 line output from the fourth line memory Interpolation data generation means, frame interpolation data generation means for generating frame interpolation data from video data delayed from 1 frame + 1 line output from the fifth line memory, and the error detection means must detect errors. When the video data output from the first line memory is delayed by one line as normal video data and an error is detected by the error detection means, the line interpolation data, the field interpolation data, and the frame interpolation data Among them, the interpolation data closest to the original video data is judged and And a data selection means for replacing the video data error is detected between the data.

  The video data restoration device according to claim 6 of the present invention is the video data restoration device according to any one of claims 3 to 5, wherein the error detection means detects an error of the video signal by CRC added for each line. Do.

  The video data restoration device according to claim 7 of the present invention is the video data restoration device according to any one of claims 3 to 5, wherein the error detection means detects an error of the video signal by using a parity added for each pixel. Do.

  According to an eighth aspect of the present invention, there is provided a video compression encoding system that performs compression encoding of the video data restoration device according to any one of the first to seventh aspects and the video data output from the video data restoration device. A video compression encoding device.

  A video data restoration method according to claim 9 of the present invention is a video data restoration method in a video data restoration device that receives video data comprising a predetermined unit of data successively and restores video data having errors. Error detection of video data is performed for each predetermined unit, a plurality of interpolation data is generated based on a plurality of unit data having a predetermined temporal and positional relationship with the error detection target unit data, and the plurality of interpolation data Of these, the interpolation data closest to the original video data is determined, and the video data in which an error is detected is replaced with the corresponding interpolation data.

  A video data restoration method according to claim 10 of the present invention is a video data restoration method in a video data restoration device that receives video data comprising a predetermined unit of data successively and restores video data having an error. First interpolation data generating means for performing error detection of video data for each predetermined unit and generating interpolation data from unit data before and after the error detection target unit data; unit data of a field different from the field of the error detection target unit data At least two of the second interpolation data generation means for generating the interpolation data from the second interpolation data generation means for generating the interpolation data from the unit data of the frame different from the frame of the error detection target unit data. Of the interpolated data output by the interpolated interpolated data generating means, the most original video data is output. Determining interpolation data close to data, replace the video data error is detected in the relevant interpolation data.

  In the video data restoration method according to claim 11 of the present invention, the error detection means detects the error of the input video data, and the input video data is detected by the first line memory and the second line memory, respectively. The input video data is delayed by one field and one line in the field memory, and the output of the field memory is output by one line in each of the third line memory and the fourth line memory. Line interpolation data is generated from the input video data and the video data delayed by two lines output from the second line memory, and output from the third line memory. 1 field delayed video data and 1 field + 1 line delayed video output from the fourth line memory Field interpolation data is generated from the data, and if no error is detected by the error detection means, the video data delayed by one line output from the first line memory is set as normal video data, and the error detection is performed. When an error is detected by the means, the interpolation data closest to the original video data is determined from the line interpolation data and the field interpolation data, and the video data in which the error is detected is replaced with the corresponding interpolation data.

  In the video data restoration method according to claim 12 of the present invention, the error detection means performs error detection of the input video data error detection means, and the input video data is detected by the first line memory and the second line memory. The input video data is delayed by one frame and two lines, respectively, the input video data is delayed by one frame in the frame memory, the output of the frame memory is delayed by one line in the third line memory, and the input video is delayed. Line interpolation data is generated from the data and the video data delayed by two lines output from the second line memory, and a frame is generated from the video data delayed from one frame + 1 line output from the third line memory. If interpolation data is generated and no error is detected by the error detection means, 1 is output from the first line memory. When the video data delayed by IN is set as normal video data and an error is detected by the error detection means, the interpolation data closest to the original video data is determined from the line interpolation data and the frame interpolation data, and the corresponding interpolation is performed. Replace video data in which error is detected with data.

  In the video data restoration method according to claim 13 of the present invention, the error detection means detects an error of the input video data, and each of the input video data is 1 by the first line memory and the second line memory. The input video data is delayed by one field and one line in the field memory, and the output of the field memory is output by one line in the third line memory and the fourth line memory, respectively. The input video data is delayed by one frame and two lines, the input video data is delayed by one frame in the frame memory, the output of the frame memory is delayed by one line in the fifth line memory, the input video data, Line interpolation data is generated from the video data delayed by two lines output from the second line memory, and the third line is generated. Field interpolation data is generated from the video data delayed by one field output from the memory and the video data delayed by one field + 1 line output from the fourth line memory, and output from the fifth line memory. If the frame interpolation data is generated from the video data delayed by one frame + one line and no error is detected by the error detection means, the video data delayed by one line output from the first line memory is normal. When an error is detected by the error detection means as video data, the interpolation data closest to the original video data is determined from the line interpolation data, the field interpolation data, and the frame interpolation data. Replace detected video data.

  A video data restoration method according to a fourteenth aspect of the present invention is the video data restoration method according to any one of the eleventh to thirteenth aspects, wherein error detection of a video signal is performed by a CRC added for each line.

  A video data restoration method according to a fifteenth aspect of the present invention is the video data restoration method according to any one of the eleventh to thirteenth aspects, wherein error detection of a video signal is performed using a parity added for each pixel.

  According to the present invention, when an error in video data is detected, a plurality of pieces of interpolation data are generated based on a plurality of unit data having different temporal and positional relationships, and any one of these pieces of interpolation data is generated. Is the closest to the original video data, and the video data in which the error is detected is replaced with the corresponding interpolation data. Accordingly, it is possible to generate interpolation data for error correction that is always optimal (closest to the original video data) according to the motion state of the video data.

  First, the outline of the present invention will be described. When an error in video data is detected, there are the following three types of methods for repairing video data having an error (video data for one line). (1) Interpolation data is generated from video data of lines before and after the line with an error. (2) Interpolation data is generated from the video data of the field preceding the field of the line where the error occurred. (3) Interpolation is performed with the video data of the frame before the frame of the line with the error. Among these interpolation methods, an interpolation method closest to the original video data is determined, and video data having an error is interpolated (replaced). Thereby, it is possible to prevent the video data from being changed in the vertical direction of the screen due to an error, and to suppress the deterioration of the image quality.

  FIG. 1 is a block diagram showing an embodiment (Embodiment 1) of a video data restoration apparatus of the present invention, showing an example applied to a video compression coding system.

  In FIG. 1, the video data restoration apparatus of this example includes a line memory 1, a CRC error detection unit 2, a video data selection unit 3, a CRC code recalculation unit 4, a line interpolation unit 20, and a frame interpolation unit 30. And a field interpolation unit 40 and an interpolation data selection unit 60. The output of the CRC code recalculation unit 4 is input to the video compression encoding device 5, and the video data restoration device and the video compression encoding device together constitute a video compression encoding system. The video data restoration device selects the input video data when there is no error, and selects the interpolated video data when there is an error, and sends it to the video compression encoding device. As the interpolated video data, optimum interpolation data is selected from line interpolation data, frame interpolation data, and field interpolation data.

  Next, the configuration and operation of each unit will be described in detail.

  The video data signal processed by the video data restoration apparatus of this example is a digital signal of an interlace scanning method, the frame rate is about 30 frames / second, and one frame consists of two fields. The number of horizontal scanning lines (line: hereinafter referred to as H) per field is, for example, about 250 (about 550 is also possible). The number of pixels per line (1H) (sampling number) is, for example, about 900 pixels. Each 1H video data is assigned a data value of a predetermined number of bits corresponding to each pixel (sampling point). Also, a CRC (Cyclic Redundancy Check) code is added to the end of each 1H video data.

  The line memory 1 delays the video data input to the video data restoration device by one line (1H). Interpolation data generation, interpolation data selection, and video data selection processing are performed using the output of the line memory 1 as “delay reference data”. From the viewpoint of the delay reference data, the video data input to the line memory 1 becomes 1H advanced data, that is, -1H delayed data.

  The CRC error detection unit 2 performs a CRC check every 1H to check whether there is an error in the input video data (delay reference data). The CRC error detection unit 2 sends a “video data selection control signal” to the video data selection unit 3 when there is an error in the video data.

  If the video data selection control signal is not received, the video data selection unit 3 selects and outputs the input video data. When the video data selection control signal is received, the video data selection unit 3 selects and outputs the interpolated video data.

  The CRC code recalculation unit 4 recalculates the CRC code for the interpolated video data being output from the video data selection unit 3 and adds the calculation result to the interpolated video data. The video data is sent to the compression encoding device 5.

  The CRC code recalculator 4 recalculates the CRC code for all output video data (input video data and interpolated video data) of the video data selector 3 in order to simplify the control process. It is also possible to replace or add the calculation result.

  The line interpolation unit 20 includes a line memory 21, a line memory 22, a subtracter 23, and an interpolation data generation unit 24.

  The line memory 21 delays the delay reference data by 1H and outputs it as “+ 1H delay data”. The + 1H delay data is sent to the video data selection unit 3 as video data to be selected when there is no error.

  The line memory 22 delays the + 1H delay data by 1H and outputs it as “+ 2H delay data”.

  The interpolation data generation unit 24 adds the delay reference data and the + 2H delay data to generate “line interpolation data” and sends it to the interpolation data selection unit 60. This addition process will be described in more detail. The delay reference data and + 2H delay data are added for each pixel and divided by 2, that is, the average of the delay reference data and + 2H delay data is obtained for each pixel, and the line interpolation data And

  The subtracter 23 calculates the difference between the delay reference data and the + 2H delay data, and sends the calculation result to the interpolation data selection unit 60 as “interpolation selection coefficient K1”. This difference calculation process will be described in more detail. The difference (absolute value) for each pixel between the delay reference data and the + 2H delay data is obtained, and the total sum for all the pixels is obtained as the interpolation selection coefficient K1.

  The frame interpolation unit 30 includes a frame memory 31, two line memories 32 and 33, two subtractors 34 and 35, an adder 36, and an interpolation data generation unit 37.

  The frame memory 31 delays the delay reference data by one frame and outputs it as “one frame delay data”.

  The line memory 32 delays the 1-frame delay data by 1H and outputs it as “1 frame + 1H delay data”.

  The line memory 33 delays 1 frame + 1H delay data by 1H and outputs it as “1 frame + 2H delay data”.

  The interpolation data generation unit 37 sends the 1 frame + 1H delay data output from the line memory 32 to the interpolation data selection unit 60 as “frame interpolation data”. Frame interpolation data is also data for each pixel.

  The subtractor 34 calculates a difference between the delay reference data and the one-frame delay data (difference absolute value for each pixel), and outputs the calculation result to the adder 36.

  The subtractor 35 calculates a difference (difference absolute value for each pixel) between + 2H delay data obtained by the line interpolation unit 20 and 1 frame + 2H delay data, and outputs the calculation result to the adder 36.

  The adder 36 adds the output of the subtracter 34 and the output of the subtractor 35 and sends the addition result to the interpolation data selection unit 60 as “interpolation selection coefficient K2”. That is, the interpolation selection coefficient K2 is the sum of all pixels of the difference (absolute value) for each pixel between the delay reference data and 1 frame delay data, and the difference (absolute value) between + 2H delay data and 1 frame + 2H delay data. Is the sum of the sum of all pixels.

  The field interpolation unit 40 includes a field memory 41, three line memories 42, 43, and 44, three adders 45, 47, and 49, two subtractors 46 and 48, and an interpolation data generation unit 50. is doing.

  The field memory 41 delays “−1H delay data”, which is input video data before being delayed by 1H in the line memory 1, by 1 field, and outputs it as “1 field-1H delay data”.

  The line memory 42 delays the 1 field-1H delay data by 1H and outputs it as “1 field delay data”.

  The line memory 43 delays the 1-field delay data by 1H and outputs it as “1 field + 1H delay data”.

  The line memory 44 delays 1 field + 1H delay data by 1H and outputs it as “1 field + 2H delay data”.

  The interpolation data generation 50 adds the 1 field delay data output from the line memory 42 and the 1 field + 1H delay data output from the line memory 43 (calculates an average value for each pixel) to obtain “field interpolation data”. And is sent to the interpolation data selection unit 60.

  The adder 45 adds 1 field-1H delay data and 1 field delay data (calculates an average value for each pixel), and outputs the addition result to the subtractor 46. This addition result (average value) indicates delay reference data or data corresponding to the line position of 1-frame delay data.

  The subtracter calculates a difference (difference absolute value for each pixel) between the output of the adder 45 and the delay reference data, and outputs the calculation result to the adder 49.

  The adder 47 adds 1 field + 1H delay data and 1 field + 2H delay data (calculates an average value for each pixel), and outputs the addition result to the subtractor 48. This addition result (average value) indicates data corresponding to the line position of + 2H delay data or 1 frame + 2H delay data.

  The subtractor 48 calculates a difference (difference absolute value for each pixel) between the output of the adder 47 and the + 2H delay data from the line interpolation unit 20, and outputs the calculation result to the adder 49.

  The adder 49 adds the output of the subtractor 46 and the output of the subtractor 48 and sends the addition result to the interpolation data selection unit 60 as “interpolation selection coefficient K3”. That is, the interpolation selection coefficient K3 is the sum of all pixels of the difference absolute value between the average value of 1 field-1H delay data and 1 field delay data and the delay reference data, 1 field + 1H delay data, and 1 field + 2H delay data. Is the sum of the absolute value of the difference between the average value of + 2H delay data and the sum of all pixels.

  The interpolation data selection unit 60 includes a selection unit 61 and a minimum value comparison unit 62.

  The minimum value comparison unit 62 compares the interpolation selection coefficients K1, K2, and K3 obtained by the line interpolation unit 20, the frame interpolation unit 30, and the field interpolation unit 40 with each other, and selects the interpolation value having the smallest value. An interpolation data selection control signal indicating a coefficient is output to the selection unit 61.

  The selection unit 61 selects the interpolation data from any of the line interpolation unit 20, the frame interpolation unit 30, and the field interpolation unit 40 according to the value of the interpolation data selection control signal, and sends it to the video data selection unit 3. That is, the selection unit 61 selects the line interpolation data when the interpolation selection coefficient K1 is the smallest, selects the frame interpolation data when the interpolation selection coefficient K2 is the smallest, and the interpolation selection coefficient K3 is the smallest. In this case, field interpolation data is selected.

  Here, generation and interpolation of interpolation data for each interpolation type for video data in which an error has been detected will be described with reference to FIG. FIG. 2 explains part of the video data in the vertical direction centering on image data having an error. The horizontal direction is video data for 1H.

  In FIG. 2, dots indicated by black circles are video data without errors and indicate image data used for generating interpolation data, and dots indicated by white circles indicate video data having errors to be interpolated.

  Point 1-3 indicates interpolation by line interpolation data, which is effective when the difference between the peripheral data (points 1-1 and 1-2) of the video data in error is small, and is suitable for moving video data. .

  Point 2-3 indicates interpolation based on frame interpolation data, which is effective when the difference between the video data in error and the video data of one frame before (point 2-1) is “0” or small. Suitable for video data.

  A point 3-3 indicates interpolation using field interpolation data. When the difference between the peripheral data (points 3-1 and 3-2) of the error video data is small, the video data moves faster than the line interpolation. Suitable for.

  The video compression encoding device 5 compresses and encodes video data restored by the video data restoration device and having no error.

  Next, the operation of the present embodiment will be described in detail with reference to the time charts shown in FIGS.

  FIG. 3 is a time chart showing the operation of the line interpolation unit 20 and also shows the operations of the line memory 1 and the CRC error detection unit 2.

  “B1 ′” of the video input of the line memory 1 is video data including an error (video data for 1H), “a1”, “c1” to “g1” are normal video data without errors (videos for 1H each) Data).

  The video inputs a1 to g1 are delayed by 1H in the line memory 1 and output as delay reference data. The video input is further output after being delayed by 1H in the line memory 21 and the line memory 22 of the line interpolation unit 20 and output as video data delayed by 1H such as + 1H delay data and + 2H delay data.

  The CRC error detection unit 2 inputs the CRC code of the video data b1 ', detects an error, and sets "1" (logical value 1: high level signal) to the output. After detecting the error, the CRC error detection unit 2 sets the output to “1” until the next video data, in FIG. 3, “c1” can detect that there is no CRC error.

  In the line interpolation unit 20, the subtracter 23 calculates an interpolation selection coefficient K1. The approximate calculation formula is as follows. Interpolation selection coefficient K1 = (delay reference data) − (+ 2H delay data) = (a1−c1). More specifically, the absolute difference value between a1 and c1 is obtained for each pixel, and the sum of absolute difference values for all pixels is obtained.

  The interpolation data generation unit 24 calculates line interpolation data. The approximate calculation formula is as follows. Line interpolation data = (delay reference data) + (+ 2H delay data) = (a1 + c1). More specifically, an average value of a1 and c1 is obtained for each pixel.

  The line interpolation unit 20 outputs the calculated interpolation selection coefficient K1 and line interpolation data to the interpolation data selection unit 60.

  Note that error detection information (for example, video data selection control signal) in the CRC error detection unit 2 may be notified to the line interpolation unit 20 by an appropriate means (not shown). Thereby, the calculation of the interpolation selection coefficient K1 and the line interpolation data can be performed only when a CRC error is detected.

  FIG. 4 is a time chart showing the operation of the frame interpolation unit 30 and also shows the operation of the line memory 1 and the CRC error detection unit 2.

  The video data “a0” to “f0” are 1-frame delayed data obtained by delaying the output (delay reference data) of the line memory 1 by 1 frame in the frame memory 31, and the video data one frame before the video inputs a1 to f1. It is. The line memory 32 and the line memory 33 delay 1 frame delay data by 1H, respectively, and output 1 frame + 1H delay data and 1 frame + 2H delay data.

  The subtractor 34 calculates (1 frame delay data) − (delay reference data), and the subtractor 35 calculates (1 frame + 2H delay data) − (+ 2H delay data) and outputs them to the adder 36. .

  The adder 36 calculates an interpolation selection coefficient K2 from the output of the subtracter 34 and the output of the subtractor 35. The approximate calculation formula is as follows. Interpolation selection coefficient K2 = ((1 frame delay data) − (delay reference data)) + ((1 frame + 2H delay data) − (+ 2H delay data)) = ((c0−c1) + (a0−a1)). More specifically, the difference absolute value between c0 and c1 and the difference absolute value between a0 and a1 are obtained for each pixel, and the sum of the difference absolute values for all pixels is obtained.

  The interpolation data generation unit 37 calculates frame interpolation data. The approximate calculation formula is as follows. (Frame interpolation data) = (1 frame + 1H delay data) = b0. More specifically, b0 is calculated for each pixel.

  The frame interpolation unit 30 outputs the calculated interpolation selection coefficient K2 and frame interpolation data to the interpolation data selection unit 60.

  As in the case of the line interpolation unit 20, error detection information in the CRC error detection unit 2 may be notified to the frame interpolation unit 20. Thereby, the calculation of the interpolation selection coefficient K2 and the frame interpolation data can be performed only when a CRC error is detected.

  FIG. 5 is a time chart showing the operation of the field interpolation unit 40 and also shows the operation of the line memory 1 and the CRC error detection unit 2.

  Video data “A0” to “G0” are 1 field-1H delayed data obtained by delaying the input video to the line memory 1 by 1 field in the field memory 41, and are video data one field before the video inputs a1 to g1. .

  The line memory 42, the line memory 43, and the line memory delay 1 field-1H delay data by 1H, respectively, and output 1 field delay data, 1 field + 1H delay data, and 1 field + 2H delay data.

  The adder 45 calculates ((1 field-1H delay data) + (1 field delay data)) and outputs it to the subtractor 46.

  The subtractor 46 calculates ((output of the adder 45) − (delay reference data)) and outputs it to the adder 49.

  The adder 47 calculates ((1 field + 1H delay data) + (1 field + 2H delay data)) and outputs it to the subtractor 48.

  The subtractor 48 calculates ((output of the adder 47) − (+ 2H delay data)) and outputs it to the adder 49.

  The adder 49 calculates an interpolation selection coefficient K3 based on the output of the subtractor 46 and the output of the subtractor 48. The approximate calculation formula is as follows. Interpolation selection coefficient K3 = [[[(1 field-1H delay data) + (1 field delay data)]-(delay reference data)] + [[(1 field + 1H delay data) + (1 field + 2H delay data)]- (+ 2H delay data)] = [[(D0 + C0) -c1]] + [[(B0 + A0) -a1]]. More specifically, the average value for each pixel of D0 and C0 and the difference absolute value for each pixel of c1 are obtained, the average value for each pixel of B0 and A0 and the difference absolute value for each pixel of a1 are obtained, Find the sum of absolute differences for all pixels.

  The interpolation data generation unit 50 calculates field interpolation data. The approximate calculation formula is as follows. (Field interpolation data) = (1 field delay data) + (1 field + 1H delay data) = (C0 + B0). More specifically, an average value for each pixel of C0 and B0 is obtained.

  The field interpolation unit 40 outputs the calculated interpolation selection coefficient K3 and field interpolation data to the interpolation data selection unit 60.

  As in the case of the line interpolation unit 20, error detection information in the CRC error detection unit 2 may be notified to the field interpolation unit 40. Thereby, the interpolation selection coefficient K3 and the field interpolation data can be calculated only when a CRC error is detected.

  FIG. 6 is a time chart showing the operation of the interpolation data selection unit 60. The interpolation data selection unit 60 compares the interpolation selection coefficient K1, the interpolation selection coefficient K2, and the interpolation selection coefficient K3 output by each interpolation unit (line interpolation unit 20, frame interpolation unit 30, and field interpolation unit 40). Judge the one with the smallest number.

  The interpolation data selection unit 60 adopts and outputs the interpolation data of the interpolation unit that outputs the interpolation selection coefficient with the smallest value as the interpolation processing video data. For example, when it is determined that (interpolation selection coefficient K3)> (interpolation selection coefficient K1)> (interpolation selection coefficient K2), as shown in FIG. 6, frame interpolation data is adopted as the interpolation-processed video data.

  The video data selection unit 3 selects video data (selected in units of 1H) according to the value of the video data selection control signal output from the CRC error detection unit 2 and outputs the video data to the CRC code recalculation unit 4. When the value of the video data selection control signal is “0” (logical value 0: low level signal) indicating no error, the video data selection unit 3 selects the + 1H delay data output from the line memory 21. When the value of the video data selection control signal is “1” (logical value 1: high level signal) indicating that there is an error, the video data selection unit 3 selects the interpolated video data from the interpolation data selection unit 60.

  The CRC code recalculation unit 4 completes the restoration of the video data in error by recalculating and adding the CRC code to the interpolated video data being output by the video data selection unit 3. The CRC code recalculation unit 4 outputs all the video data including the restoration to the subsequent apparatus, in the case of the present embodiment, to the video compression encoding apparatus 5. As described above, the CRC code recalculation unit 4 applies CRC codes to all output video data (input video data and interpolated video data) of the video data selection unit 3 in order to simplify the control process. Can be recalculated. The CRC code recalculation unit 4 not only adds the CRC code obtained by recalculation to the interpolated video data of the video data having an error, but also adds the CRC code of the input video data having no error to the same CRC code. Replace with. In this way, it is not necessary for the CRC code recalculation unit 4 to determine when it is necessary to recalculate the CRC code, and the control process can be simplified.

  As described above, according to the present embodiment, when an error in video data is detected, predetermined interpolation data, interpolation data, and original video data from each of the line interpolation unit, the field interpolation unit, and the frame interpolation unit, A selection coefficient indicating the degree of similarity is output. By comparing the selection coefficients, it is determined which interpolation data is closest to the original video data, and the video data of the line with the error is interpolated (replaced) with the corresponding interpolation data.

  Note that the line interpolation unit generates interpolation data from video data of lines before and after the line having an error. This interpolation data is effective when the difference between the peripheral data of the video data in error is small, and is suitable for moving video data.

  The field interpolation unit generates interpolation data from the video data of the field preceding the field of the line having the error. This interpolation data is suitable for video data that moves faster than line interpolation when the difference between the peripheral data of the video data in error is small.

  The frame interpolating unit interpolates with the video data of the frame before the frame of the line having the error. This interpolation data is effective when the difference between the error video data and the video data of one frame before is “0” or small, and is suitable for video data in the case of a still image.

  As described above, according to the present embodiment, the restoration source line can be selected according to the motion state of the video, so that the optimum error correction interpolation data (closest to the original video data) can always be obtained. Can be generated. Furthermore, it is possible to output video data without failure to the video compression encoding apparatus, to prevent a change in the vertical direction of the video data due to an error in compression encoding processing, and to suppress deterioration in image quality.

  FIG. 7 is a block diagram showing another embodiment (embodiment 2) of the video data restoration apparatus of the present invention, and shows an example applied to a video compression encoding system. The basic configuration of the video data restoration apparatus of this example is the same as that of the video data restoration apparatus of Example 1 shown in FIG. The video data restoration apparatus of this example is different from that of the first embodiment in that parity is used instead of CRC as means for detecting error in video data.

  In FIG. 7, the video data restoration apparatus of this example includes a line memory 1, a parity error detection unit 8, a video data selection unit 3, a CRC code recalculation unit 4, a line interpolation unit 20, and a frame interpolation unit 30. And a field interpolation unit 40 and an interpolation data selection unit 60. The output of the CRC code recalculation unit 4 is input to the video compression encoding device 5, and the video data restoration device and the video compression encoding device together constitute a video compression encoding system. The video data restoration device selects the input video data when there is no error, and selects the interpolated video data when there is an error, and sends it to the video compression encoding device. As the interpolated video data, optimum interpolation data is selected from line interpolation data, frame interpolation data, and field interpolation data.

  Since the internal configuration and functions of other units except the parity error detection unit 8 are the same as those shown in FIG. 1, detailed description thereof is omitted. However, the interpolation data generation unit 24 of the line interpolation unit 20, the interpolation data generation unit 37 of the frame interpolation unit 30, and the interpolation data generation unit 50 of the field interpolation unit 40 have the following points as shown in FIG. Different. That is, each interpolation data generation unit 24, 37, 50 generates parity information in accordance with data corresponding to each pixel when generating interpolation data.

  Hereinafter, the parity error detection unit 8 will be described.

  Input video data to the line memory 1 is divided for each line (1H), and a CRC code is added to the end of each 1H video data. Each line of input video data is composed of data corresponding to a plurality of pixels and parity information thereof.

  The output of the line memory 1 is input to the line interpolation unit 20 and is output from the line memory 21 as “+ 1H delay data”.

  The parity error detection unit 8 receives the + 1H delay data that is the output of the line memory 21, checks the parity added to the data for each pixel, and determines whether there is an error in the input video data. The parity error detection unit 8 sends a “video data selection control signal” to the video data selection unit 3 when there is an error in the video data.

  FIGS. 3 to 6 are time charts showing the operations of the line interpolation unit 20, the frame interpolation unit 30, the field interpolation unit 40, and the interpolation data selection unit 60 in the present embodiment and are used in the description of the first embodiment. Charts can be used. However, CRC error detection is read as parity error detection.

  The video data selection unit 3 selects video data (selected in units of 1H / pixel) according to the value of the video data selection control signal output from the parity error detection unit 8 and outputs the video data to the CRC code recalculation unit 4. When the value of the video data selection control signal is “0” (logical value 0: low level signal) indicating no error, the video data selection unit 3 selects the + 1H delay data output from the line memory 21. When the value of the video data selection control signal is “1” (logical value 1: high level signal) indicating that there is an error, the video data selection unit 3 selects the interpolated video data from the interpolation data selection unit 60.

  The CRC code recalculation unit 4 completes the restoration of the video data in error by recalculating and adding the CRC code to the interpolated video data being output by the video data selection unit 3. The CRC code recalculation unit 4 outputs all the video data including the restoration to the subsequent apparatus, in the case of the present embodiment, to the video compression encoding apparatus 5. Also in this embodiment, in order to simplify the control process of the CRC code recalculation unit 4, all the output video data (input video data and interpolated video data) of the video data selection unit 3 are CRC codes. Can be recalculated. The CRC code recalculation unit 4 not only adds the CRC code obtained by recalculation to the interpolated video data of the video data having an error, but also adds the CRC code of the input video data having no error to the same CRC code. Replace with.

  As described above, according to the present embodiment, since the presence or absence of an error can be determined by the parity corresponding to the pixel, the interpolation processing can be controlled for each pixel of the input video data, which is the same as the first embodiment of FIG. The effect of can be obtained.

  Next, still another embodiment of the video data restoration device of the present invention will be described.

  In the video data restoration devices of the first and second embodiments described above, three means of the line interpolation unit 20, the frame interpolation unit 30, and the field interpolation unit 40 are provided as means for creating interpolation data. However, as a modification of the first and second embodiments, there is a first modification in which the interpolation data creating means is configured by only two of the line interpolation unit 20 and the field interpolation unit 40. Alternatively, there is a second modification example that is configured by only two interpolation data creating means of the line interpolation unit 20 and the frame interpolation unit 30. Furthermore, there is a third modification in which the line interpolation unit 20 is deleted leaving only the two line memories 21 and 22, and the frame interpolation unit 30 and the field interpolation unit 40 are configured with only two interpolation data creation means.

  As described above, the line interpolation unit generates interpolation data from the video data of the lines before and after the line having the error. The line interpolation data is effective when the difference between the peripheral data of the erroneous video data is small, and is suitable for moving video data.

  The field interpolation unit generates interpolation data from the video data of the field preceding the field of the line having the error. Field interpolation data is suitable for video data that moves faster than line interpolation when the difference between the peripheral data of the video data in error is small.

  The frame interpolating unit interpolates with the video data of the frame before the frame of the line having the error. The frame interpolation data is effective when the difference between the error video data and the video data of one frame before is “0” or small, and is suitable for video data in the case of a still image.

  Therefore, when the line interpolation unit and the field interpolation unit are provided (first modified example), it is suitable for selecting interpolation data from video data with motion and video data with faster motion. .

  When the line interpolation unit and the frame interpolation unit are provided (second modification), it is suitable for selecting the interpolation data from the moving image data and the image data in the case of a still image.

  When the field interpolation unit and the frame interpolation unit are provided (third modification), it is suitable for selecting the interpolation data from the video data with fast motion and the video data in the case of a still image.

  Since other parts are the same as those in the first and second embodiments, detailed description thereof is omitted.

  According to the present embodiment (Modifications 1 to 3), it is possible to reduce the hardware scale of the video data restoration device while maintaining the feature that enables the selection of the restoration source line according to the motion state of the video. it can.

  A video device related to broadcasting and communication that compresses and processes video data and transmits / receives it, and a device that reproduces video data recorded and distributed on a recording medium such as a hard disk or VTR.

It is a block block diagram which shows one Embodiment of the video data restoration apparatus of this invention. It is a figure for demonstrating the production | generation of the interpolation data classified by interpolation type, and the interpolation method. It is a time chart showing operation | movement of a line interpolation part. It is a time chart showing operation | movement of a frame interpolation part. It is a time chart showing operation | movement of a field interpolation part. 3 is a time chart showing the operation of an interpolation data selection unit 60. It is a block block diagram which shows other embodiment of the video data restoration apparatus of this invention.

Explanation of symbols

2 CRC Error Detection Unit 3 Video Data Selection Unit 4 CRC Code Recalculation Unit 5 Video Compression Encoder 8 Parity Error Detection Unit 20 Line Interpolation Unit 30 Frame Interpolation Unit 40 Field Interpolation Unit 60 Interpolation Data Selection Unit

Claims (15)

In a video data restoration device that receives video data composed of a predetermined unit of data sequentially and restores video data with errors,
Error detection means for detecting an error in the video data for each predetermined unit;
Interpolation data generating means for generating a plurality of interpolation data based on a plurality of unit data having a predetermined temporal and positional relationship with respect to the error detection target unit data;
A plurality of interpolation data generated by the interpolation data generating means, determining the interpolation data closest to the original video data, and data selection means for replacing the video data in which an error is detected with the corresponding interpolation data, Video data restoration device. In a video data restoration device that receives video data composed of a predetermined unit of data sequentially and restores video data with errors,
Error detection means for detecting an error in the video data for each predetermined unit;
First interpolation data generating means for generating interpolation data from unit data before and after error detection target unit data, and second interpolation data generating means for generating interpolation data from unit data in a field different from the field of error detection target unit data And at least two arbitrarily selected interpolation data generating means among the third interpolation data generating means for generating interpolation data from unit data of a frame different from the frame of the error detection target unit data;
A data selection unit that determines interpolation data closest to the original video data from among the interpolation data output by the arbitrarily selected interpolation data generation unit and replaces the video data in which an error is detected by the corresponding interpolation data; A video data restoration device characterized by that. Error detection means for detecting an error in the input video data;
A first line memory and a second line memory for delaying input video data by one line and two lines, respectively;
A field memory for delaying input video data by one field and one line;
A third line memory and a fourth line memory for delaying the output of the field memory by one line and two lines, respectively;
Line interpolation data generating means for generating line interpolation data from the input video data and video data delayed by two lines output from the second line memory;
Field interpolation data generating means for generating field interpolation data from video data delayed by one field output from the third line memory and video data delayed from one field + 1 line output from the fourth line memory When,
If no error is detected by the error detecting means, the video data delayed from one line output from the first line memory is set as normal video data, and when the error is detected by the error detecting means, the line interpolation data And a data selection unit for determining interpolation data closest to the original video data among the field interpolation data and replacing the video data in which an error is detected with the corresponding interpolation data. Error detection means for detecting an error in the input video data;
A first line memory and a second line memory for delaying input video data by one line and two lines, respectively;
A frame memory that delays input video data by one frame;
A third line memory for delaying the output of the frame memory by one line;
Line interpolation data generating means for generating line interpolation data from the input video data and video data delayed by two lines output from the second line memory;
Frame interpolation data generating means for generating frame interpolation data from video data delayed by one frame + 1 line output from the third line memory;
If no error is detected by the error detecting means, the video data delayed from one line output from the first line memory is set as normal video data, and when the error is detected by the error detecting means, the line interpolation data And a video data restoration device comprising: data selection means for determining interpolation data closest to the original video data among the frame interpolation data and replacing video data in which an error is detected with the interpolation data. Error detection means for detecting an error in the input video data;
A first line memory and a second line memory for delaying input video data by one line and two lines, respectively;
A field memory for delaying input video data by one field and one line;
A third line memory and a fourth line memory for delaying the output of the field memory by one line and two lines, respectively;
A frame memory that delays input video data by one frame;
A fifth line memory for delaying the output of the frame memory by one line;
Line interpolation data generating means for generating line interpolation data from the input video data and video data delayed by two lines output from the second line memory;
Field interpolation data generating means for generating field interpolation data from video data delayed by one field output from the third line memory and video data delayed from one field + 1 line output from the fourth line memory When,
Frame interpolation data generating means for generating frame interpolation data from video data delayed by 1 frame + 1 line output from the fifth line memory;
If no error is detected by the error detecting means, the video data delayed from one line output from the first line memory is set as normal video data, and when the error is detected by the error detecting means, the line interpolation data And a data selection means for determining an interpolation data closest to the original video data among the field interpolation data and the frame interpolation data and replacing the video data in which an error is detected with the corresponding interpolation data. Video data restoration device.   6. The video data restoration apparatus according to claim 3, wherein the error detection unit detects an error of the video signal by a CRC added for each line.   6. The video data restoration apparatus according to claim 3, wherein the error detection unit detects an error of the video signal based on a parity added for each pixel. The video data restoration device according to any one of claims 1 to 7,
A video compression encoding system comprising: a video compression encoding device that performs compression encoding of video data output from the video data restoration device. In a video data restoration method in a video data restoration device that receives video data composed of a predetermined unit of data sequentially and restores video data having an error,
Perform error detection of video data for each predetermined unit,
Generating a plurality of interpolation data based on a plurality of unit data having a predetermined temporal and positional relationship with respect to the error detection target unit data;
A video data restoration method comprising: determining interpolation data closest to the original video data among the plurality of interpolation data, and replacing video data in which an error is detected with the interpolation data. In a video data restoration method in a video data restoration device that receives video data composed of a predetermined unit of data sequentially and restores video data having an error,
Perform error detection of video data for each predetermined unit,
First interpolation data generating means for generating interpolation data from unit data before and after error detection target unit data, and second interpolation data generating means for generating interpolation data from unit data in a field different from the field of error detection target unit data , And third interpolation data generating means for generating interpolation data from unit data in a frame different from the frame of the error detection target unit data, and at least two arbitrarily selected interpolation data generating means respectively output interpolation data A video data restoration method comprising: determining interpolation data closest to the original video data, and replacing video data in which an error is detected with the corresponding interpolation data. The error detection means detects errors in the input video data,
In the first line memory and the second line memory, the input video data is delayed by one line and two lines, respectively.
In the field memory, the input video data is delayed by 1 field-1 line,
In the third line memory and the fourth line memory, the output of the field memory is delayed by one line and two lines, respectively.
Generating line interpolation data from the input video data and video data delayed by two lines output from the second line memory;
Generating field interpolation data from video data delayed by one field output from the third line memory and video data delayed from one field + 1 line output from the fourth line memory;
If no error is detected by the error detecting means, the video data delayed from one line output from the first line memory is set as normal video data, and when the error is detected by the error detecting means, the line interpolation data A video data restoration method comprising: determining interpolation data closest to the original video data among the field interpolation data, and replacing video data in which an error is detected with the interpolation data. The error detection means performs error detection of the input video data error detection means,
In the first line memory and the second line memory, the input video data is delayed by one line and two lines, respectively.
In the frame memory, the input video data is delayed by one frame,
In the third line memory, the output of the frame memory is delayed by one line,
Generating line interpolation data from the input video data and video data delayed by two lines output from the second line memory;
Generating frame interpolation data from video data delayed by 1 frame + 1 line output from the third line memory;
If no error is detected by the error detecting means, the video data delayed from one line output from the first line memory is set as normal video data, and when the error is detected by the error detecting means, the line interpolation data A video data restoration method comprising: determining interpolation data closest to the original video data among the frame interpolation data, and replacing the video data in which an error is detected with the interpolation data. The error detection means detects errors in the input video data,
In the first line memory and the second line memory, the input video data is delayed by one line and two lines, respectively.
In the field memory, the input video data is delayed by 1 field-1 line,
In the third line memory and the fourth line memory, the output of the field memory is delayed by one line and two lines, respectively.
In the frame memory, the input video data is delayed by one frame,
In the fifth line memory, the output of the frame memory is delayed by one line,
Generating line interpolation data from the input video data and video data delayed by two lines output from the second line memory;
Generating field interpolation data from video data delayed by one field output from the third line memory and video data delayed from one field + 1 line output from the fourth line memory;
Generating frame interpolation data from video data delayed by 1 frame + 1 line output from the fifth line memory;
If no error is detected by the error detecting means, the video data delayed from one line output from the first line memory is set as normal video data, and when the error is detected by the error detecting means, the line interpolation data A video data restoration method comprising: determining interpolation data closest to the original video data among the field interpolation data and the frame interpolation data, and replacing the video data in which an error is detected with the interpolation data.   14. The video data restoration method according to claim 11, wherein error detection of the video signal is performed by a CRC added for each line. 14. The video data restoration method according to claim 11, wherein error detection of the video signal is performed based on a parity added for each pixel.

Images