JP2009272672A - Method and apparatus for encoding moving image - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To make inconspicuous the deterioration in the picture quality of a moving image frame. <P>SOLUTION: A picture quality deterioration determination unit 101 determines a degree of deterioration of picture quality for each of in plurality of blocks included in a present frame and assigns a flag value to the plurality of blocks. A refresh map generation unit 102 writes an assignment result of the flag value which is assigned by the picture quality deterioration determination unit 101 to a refresh map indicating correspondence relation between the positions of blocks in a moving image frame and the flag value assigned to the blocks. A refresh frequency setting unit 103 assigns a refresh frequency to each flag value. A refresh control unit 104 selects the same number of candidate blocks as the refresh frequency corresponding to the flag value from the blocks to which the same flag value is assigned by referring to the refresh map and executes intra-encoding of blocks existing in the same positions as the candidate blocks selected out of blocks included in a succeeding frame. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a technique for encoding a moving picture frame by intra coding and inter coding, and more specifically, suppresses image quality deterioration by inserting a block coded by intra coding into a moving picture frame. Regarding technology.

  Conventionally, as an encoding method of moving image frames, intra coding (also referred to as intra-frame coding) that compresses the amount of data by removing spatial redundancy in moving image frames, and time between a plurality of moving image frames Inter coding (also referred to as interframe coding) is used in which the amount of data is compressed by removing static redundancy.

  In general, the amount of code generated by inter coding is smaller than in intra coding, but in inter coding, errors (data errors, data loss, etc.) occur in certain video frames in the transmission path. Then, an error propagates not only to the moving image frame but also to another moving image frame. In addition, since motion compensation is performed in inter coding, a frame memory in the moving image coding apparatus holds a moving image frame slightly different from the original frame (actual moving image frame). When inter coding is continuously executed, the error of the retained frame increases as time elapses, and eventually, a moving image frame that is significantly different from the original frame is retained. When a macroblock of this moving image frame is used for motion compensation, an error of this macroblock propagates to surrounding macroblocks. As described above, the propagation of the error deteriorates the image quality of the moving image frame.

  Therefore, a technique has been proposed in which an image region encoded by intra coding is inserted into a moving image frame by periodically or aperiodically performing refresh processing (forced intra coding). Since intra-coded macroblocks can be decoded without referring to other moving image frames, propagation of errors as described above can be blocked, and as a result, image quality deterioration can be suppressed. Is possible. For example, in Japanese Patent Laid-Open No. 2001-359102 (Patent Document 1), a macroblock whose difference value from a predicted macroblock is larger than a threshold (that is, a macroblock having a large amount of difference information) is refreshed, and is updated every frame period. Discloses a method for executing refresh processing on a predetermined number of target macroblocks (macroblocks to be refreshed). As shown in FIG. 9, the target macroblock is refreshed in order by a predetermined number (for example, from the upper left to the lower right of the moving image frame), thereby increasing the generated code amount and the image quality degradation due to intra coding. Suppressed.

However, when the number of refreshes (the number of macroblocks refreshed within one frame period) is fixed, the refresh target period (the last target macro after the target macroblock is selected) increases as the number of target macroblocks increases. The period required until the block is actually refreshed) becomes longer. In order to solve such a problem, in the method of Patent Document 1, the number of refreshes is determined based on the total number of target macroblocks and the elapsed time since the macroblocks became the refresh target.
JP 2001-359102 A

  However, in the conventional method, since there is only one threshold value used for determining whether or not a macroblock is to be refreshed, there is a variation in the difference value (that is, the image quality degradation degree) between the target macroblocks. In the conventional method, since the target macroblocks are refreshed in a predetermined order, there is a possibility that the refresh of the target macroblocks having a high degree of image quality degradation may be postponed. In this case, since the image quality degradation is locally intense in the moving image frame, the image quality degradation of the moving image frame becomes conspicuous (that is, the subjective image quality deteriorates).

  For example, when shooting is performed such that a subject with little movement (landscape or the like) is located in the upper half area of the screen and a subject with much movement (such as a pedestrian) is located in the lower half area of the screen (for example, a surveillance camera) In the upper half area of the moving image frame, the difference value of the target macroblock is slightly larger than the threshold value, and in the lower half area, the difference value of the target macroblock is much higher than the threshold value. Shows a large value. That is, the image quality degradation degree in the lower half area of the moving image frame is larger than that in the upper half area of the moving image frame. Here, if the target macroblock is refreshed in order from the upper left to the lower right of the moving image frame as in the conventional case, after the refresh of the upper half area (that is, the area where the image quality degradation degree is small) is completed. Thus, refreshing of the lower half area (that is, an area having a high image quality degradation degree) is executed. In this case, the image quality degradation in the lower half area becomes conspicuous.

  Accordingly, an object of the present invention is to provide a technique capable of making the deterioration of the image quality of a moving image frame less noticeable.

  According to one aspect of the present invention, a moving image encoding method is a method of sequentially encoding a plurality of moving image frames by intra-coding or inter-coding a moving image frame in units of a predetermined size block. Categorizing each of the plurality of blocks into a plurality of block groups according to the degree of image quality degradation of each of the plurality of blocks included in the first moving image frame, and classification in the above step (a). A step (b) of selecting a number of blocks corresponding to the image quality degradation degree of the block group from each of the plurality of block groups as a candidate block, and a second encoded after the first moving image frame Among the plurality of blocks included in the moving image frame, the above-described into the block corresponding to the candidate block selected in the above step (b). And a step (c) to force the coding.

  In the above moving image encoding method, an area having a high image quality degradation degree in a moving image frame can be preferentially refreshed, so that it is possible to prevent the image quality deterioration from becoming locally intense. Image quality deterioration can be made inconspicuous.

  Preferably, in the step (b), the number of candidate blocks selected from the block group is increased as the block group has a higher degree of image quality degradation.

  In the above-described moving image coding method, an area where the degree of image quality degradation is small is less necessary to be refreshed than a case where the image quality degradation is large. Therefore, an increase in the amount of generated code can be suppressed while suppressing image quality deterioration of the moving image frame. . Thereby, even when the code amount control is executed, it is possible to suppress an increase in image quality deterioration due to the code amount control.

  Preferably, the moving image encoding method further includes a step (d) of changing the number of candidate blocks selected from each of the plurality of block groups in the step (b) according to the generated code amount.

  In the moving image encoding method, an increase in the amount of generated code can be suppressed.

  Preferably, in the step (d), the plurality of block groups according to the generated code amount so that the block group having a higher image quality degradation level has a smaller increase / decrease amount of candidate blocks selected from the block group. To change the number of candidate blocks selected from each.

  In the moving image encoding method, since the number of blocks to be refreshed is stable in the area where the image quality degradation degree is large, even if the generated code amount increases or decreases, the area where the image quality degradation degree is high can be preferentially refreshed. . As a result, it is possible to suppress an increase in the amount of generated codes while suppressing image quality deterioration of the moving image frame.

  According to another aspect of the present invention, a moving image encoding device is a device that sequentially encodes a plurality of moving image frames, and encodes a moving image frame in units of blocks of a predetermined size. And an image quality deterioration degree for each of a plurality of blocks included in the moving image frame encoded by the encoding unit, and each of the plurality of blocks is divided into a plurality of block groups according to the image quality deterioration degree. The number of blocks corresponding to the image quality deterioration degree of the block group is selected as candidate blocks from each of the plurality of block groups classified by the image quality deterioration degree determination unit and the image quality deterioration degree determination unit, and then encoded. The intra coding is performed by the coding unit on a block corresponding to the selected candidate block among the plurality of blocks included in the power moving image frame. And a refresh controller to execute.

  In the moving image encoding apparatus, since a region having a high degree of image quality degradation can be preferentially refreshed in the moving image frame, it is possible to suppress local deterioration of image quality, and as a result, Image quality deterioration can be made inconspicuous.

  As described above, it is possible to make the deterioration of the image quality of the moving image frame less noticeable.

  Embodiments of the present invention will be described below with reference to the drawings.

(Embodiment 1)
FIG. 1 shows the configuration of a moving picture encoding apparatus according to Embodiment 1 of the present invention. The moving image encoding apparatus 1 sequentially encodes a plurality of moving image frames, and each of the moving image frames is intra-coded or inter-coded in a predetermined block unit (here, a macroblock unit). In addition, the moving image encoding device 1 executes refresh processing (forced intra encoding) in accordance with the image quality degradation degree of the macroblock for each frame period (period for encoding a moving image frame). The moving image encoding apparatus 1 includes a mode selection unit 11, a blocking unit 12, a subtraction unit 13, an orthogonal transformation unit 14, a quantization unit 15, a variable length encoding unit 16, and an inverse quantization unit 17. An inverse orthogonal transform unit 18, a selector 19, an addition unit 20, a memory 21, a motion detection / compensation unit 22, an intra prediction unit 23, an image quality degradation degree determination unit 101, and a refresh map generation unit 102. The refresh number setting unit 103 and the refresh control unit 104 are provided.

[Encoding section]
The mode selection unit 11 receives block correlation information indicating the degree of correlation in units of macroblocks from each of the motion detection / compensation unit 22 and the intra prediction unit 23, and either intra coding or inter coding is performed for each macroblock. Select. When the mode selection unit 11 receives a refresh control signal from the refresh control unit 104, the mode selection unit 11 forcibly selects intra coding. The blocking unit 12 divides the moving image frame into a plurality of macro blocks, and sequentially outputs the plurality of divided macro blocks. The subtraction unit 13 calculates a difference between the input macroblock (macroblock from the block forming unit 12) and the prediction macroblock from the selector 19, and outputs the calculation result as a prediction error macroblock. The orthogonal transform unit 14 performs orthogonal transform (for example, discrete cosine transform) on the prediction error macroblock from the subtraction unit 13. The quantization unit 15 quantizes the macroblock orthogonally transformed by the orthogonal transformation unit 14. The variable length coding unit 16 performs variable length coding on the macroblock quantized by the quantization unit 15.

  The inverse quantization unit 17 inversely quantizes the macroblock quantized by the quantization unit 15. The inverse orthogonal transform unit 18 performs inverse orthogonal transform (for example, inverse discrete cosine transform) on the macroblock inversely quantized by the inverse quantization unit 17. The selector 19 selects a prediction macroblock from the intra prediction unit 23 when the mode selection unit 11 selects intra coding, and moves when the mode selection unit 11 selects inter coding. A prediction macroblock from the detection / compensation unit 22 is selected. The adder 20 adds the predicted macroblock selected by the selector 19 to the macroblock that has been inversely orthogonal transformed by the inverse orthogonal transform unit 18. Thereby, the macroblock before orthogonal transformation and quantization are performed is restored. The adder 20 outputs the restored macroblock. The memory 21 stores the macro block from the adding unit 20. In the memory 21, a reference frame is reconstructed by a plurality of restored macro blocks (macro blocks from the adder 20). The motion detection / compensation unit 22 calculates a motion vector using the input macroblock (macroblock from the blocking unit 12) and the reference frame stored in the memory 21 (so-called motion detection processing). In addition, the motion detection / compensation unit 22 outputs the predicted macroblock obtained at the time of calculating the motion vector to the selector 19 (so-called motion compensation processing), and block correlation information indicating the degree of correlation in units of macroblocks. (For example, SAD) is output to the mode selection unit 11. The intra prediction unit 23 performs intra prediction using the input macroblock and the reference frame stored in the memory 22, outputs the prediction macroblock to the selector 19, and outputs block correlation information to the mode selection unit 11. .

  In this way, the moving image frame is intra-coded or inter-coded in units of macroblocks.

[Motion detection / motion compensation]
Here, the processing by the motion detection / compensation unit 22 will be described in detail. For example, the motion detection / compensation unit 22 selects a block area having the same size as the input macroblock within the set motion search range in the reference frame as a candidate macroblock, and each pixel value included in the selected candidate macroblock And the sum of absolute differences (SAD) between the average pixel value of the input macroblock and the input macroblock. Next, the motion detection / compensation unit 22 moves the position of the block region within the motion search range, selects the next candidate macroblock, and calculates the SAD for the candidate macroblock. In this way, the motion detection / compensation unit 22 selects a plurality of candidate macroblocks within the motion search range, and calculates the SAD for each candidate macroblock.

  Next, the motion detection / compensation unit 22 selects one candidate macroblock based on each calculated SAD, and outputs a motion vector indicating the distance from the input macroblock to the selected macroblock. For example, the motion detection / compensation unit 22 selects a candidate macroblock corresponding to the smallest SAD among the plurality of SADs as a macroblock having the highest degree of correlation with the input macroblock. Furthermore, the motion detection / compensation unit 22 outputs the selected macroblock as a predicted macroblock.

[Image quality degradation degree determination unit]
The image quality degradation level determination unit 101 receives information indicating the image quality degradation level from the motion detection / compensation unit 22 every time one macroblock is processed, and is included in the current frame (currently encoded video frame). The degree of image quality degradation is determined for each macroblock. For example, the image quality degradation degree determination unit 101 is supplied with SAD corresponding to each of the macroblocks from the motion detection / compensation unit 22 as information indicating the image quality degradation level, and thresholds Thr1, Thr2 (preliminarily set for the SAD and SAD). However, Thr1 <Thr2) is compared. Then, the image quality deterioration degree determination unit 101 assigns a flag value to each macroblock according to the comparison result as follows.

<Comparison result><Flagvalue>
SAD <Thr1 ・ ・ ・ ・ “0”
Thr1 ≦ SAD <Thr2... “1”
Thr2 ≦ SAD ・ ・ ・ ・ “2”
In this way, by assigning a flag value to each macroblock, the macroblock is classified into a plurality of macroblock groups. In addition, it is preferable to assign a flag value to each macroblock so that the flag value increases as the image quality degradation degree increases. Further, there may be three or more threshold values.

[Refresh map generator]
The refresh map generation unit 102 holds a refresh map (see FIG. 2A) indicating the correspondence between the position of a macroblock in a moving image frame and the flag value assigned to the macroblock, and the image quality is included in the refresh map. The classification result (flag value allocation result) by the deterioration degree determination unit 101 is written.

[Refresh number setting section]
The refresh number setting unit 103 associates each of the flag values used in the image quality degradation level determination unit 101 with the refresh number (the number of macroblocks to be refreshed within the next frame period). For example, the refresh number setting unit 103 associates the refresh number with each flag value as follows.

<Flag value><Number of refreshes>
“0” ・ ・ ・ ・ ・ ・ ・ ・ ・ ・ ・ 0
“1” 2
“2” ・ ・ ・ ・ ・ ・ 4
In this way, by associating the refresh number with each flag value, the refresh number is set for each block group classified by the image quality degradation degree determination unit 101. Note that it is desirable to assign a refresh number to each flag value so that the larger the flag value (that is, the larger the macroblock group having a higher image quality degradation degree), the larger the refresh number.

[Refresh control unit]
The refresh control unit 104 refers to the refresh map generated by the refresh map generation unit 102, and selects the refresh number assigned to the flag value by the refresh number setting unit 103 from among the macro blocks assigned the same flag value. The same number of macroblocks are selected as candidate macroblocks (macroblocks to be refreshed within the next frame period). For example, the refresh control unit 104 does not select any of the macro blocks to which the flag value “0” is assigned, selects two of the macro blocks to which the flag value “1” is assigned, and sets the flag value “2”. 4 are selected from the macroblocks to which "" is assigned (see FIG. 2B).

  In addition, the refresh control unit 104 encodes a macroblock existing at the same position as the selected candidate macroblock among a plurality of macroblocks included in the next frame (moving image frame encoded next to the current frame). In this case, a refresh control signal is output to cause the mode selection unit 11 to select intra coding. As a result, the refresh process is executed for the macroblock of the next frame corresponding to the candidate macroblock. Further, the refresh control unit 104 outputs a refresh control signal to the refresh map generation unit 102, and causes the flag value “0” to be written in the refresh position (the position of the macro block on which the refresh process has been executed) in the refresh map (see FIG. 2C).

  Next, with reference to FIG. 3, the refresh process by the moving picture encoding apparatus 1 shown in FIG. 1 will be described. Here, it is assumed that the image quality degradation degree in the lower half area of the moving image frame is larger than that in the upper half area of the moving image frame, and the flag values “0”, “1”, and “2” are respectively Assume that the refresh numbers “0”, “2”, and “4” are assigned.

  In the first frame, the upper half area of the moving image frame includes twelve macroblocks to which the flag value “1” is assigned, and the lower half area includes twelve macro blocks to which the flag value “2” is assigned. Contains macroblocks.

  Next, in the second frame, refresh processing is performed on two macroblocks in the macroblock group having the flag value “1”, and four macroblocks in the macroblock group having the flag value “2” are processed. On the other hand, a refresh process is performed. A flag value “0” is written at the position of the refreshed macroblock in the refresh map.

  Subsequently, in the 3rd and 4th frames, as in the 2nd frame, 2 refreshes are performed from the macroblock group having the flag value “1”, and 4 refreshes are performed from the macroblock group having the flag value “2”. . In the fourth frame, the refresh is completed for all the macro blocks having the flag value “2”.

  As described above, by selecting a macroblock to be refreshed in accordance with the degree of image quality degradation, it is possible to preferentially refresh an area with a large degree of image quality degradation in a moving image frame, so that the image quality degradation is locally It is possible to suppress the intensification, and as a result, it is possible to make the image quality deterioration less noticeable.

  Further, as the number of macro blocks to be refreshed increases, the amount of generated code (the amount of encoded data output from the variable length encoding unit) increases. In this case, since the code amount control circuit (not shown) increases the quantization parameter value in order to bring the generated code amount close to a predetermined value, as a result, the image quality deterioration of the macroblock occurs. However, in the moving picture encoding apparatus 1 according to this embodiment, the number of refreshes is reduced by reducing the number of refreshes corresponding to the macroblock group having a low image quality deterioration degree to the number of refreshes corresponding to the macroblock group having a high image quality deterioration degree. Therefore, the increase in the amount of generated codes can be suppressed.

(Embodiment 2)
FIG. 5 shows the configuration of a moving picture coding apparatus according to Embodiment 2 of the present invention. In this moving image encoding apparatus 2, the number of refreshes (that is, from each of the macroblock groups) according to the generated code amount of the previous frame (the code amount generated by the variable length encoding of the previous frame by the variable length encoding unit 17). The number of candidate blocks to be selected) is changed. The moving picture encoding apparatus 2 further includes a refresh number changing unit 201 in addition to the configuration shown in FIG.

  The refresh number changing unit 201 changes the refresh number associated with each flag value by the refresh number setting unit 103 according to the generated code amount of the previous frame. In addition, the refresh number changing unit 201 changes the refresh number so that the increase / decrease amount of the refresh number corresponding to the flag value decreases as the flag value increases (that is, as the macroblock group has a higher image quality degradation degree).

  For example, as shown in FIG. 5, the refresh number setting unit 103 sets the refresh numbers Nr0, Nr1, Nr2, and Nr3 corresponding to the flag values “0”, “1”, “2”, and “3” to “0”, Assume that “2”, “3”, and “5” are set. In this case, the total number of refreshes is “10”. Here, when the generated code amount increases (for example, when the generated code amount is larger than the upper limit threshold), the refresh number changing unit 201 changes the refresh numbers Nr1 and Nr2 to “0” and “2”, respectively. As a result, the total number of refreshes decreases to “7”. On the contrary, when the generated code amount decreases (for example, when the generated code amount is smaller than the lower limit threshold), the refresh number changing unit 201 changes the refresh numbers Nr1 and Nr2 to “4” and “4”, respectively. As a result, the total number of refreshes increases to “13”.

  Next, the operation of the refresh number changing unit 201 shown in FIG. 4 will be described with reference to FIG. Here, the flag values “0”, “1”, “2”, and “3” are assigned the refresh numbers “0”, “2”, “3”, and “5”, respectively. To do.

  In the first frame, the upper area of the moving image frame includes a macroblock to which a flag value “1” is assigned, the middle area includes a macroblock to which a flag value “2” is assigned, and the lower area includes a flag Includes macroblocks assigned the value “3”.

  Next, in the second frame, refresh processing is performed on two of the macro blocks having the flag value “1”, refresh processing is performed on three of the macro blocks having the flag value “2”, Refresh processing is performed on five of the macro blocks having the flag value “3”. A flag value “0” is written at the position of the refreshed macroblock in the refresh map.

  Here, when the generated code amount of the second moving image frame is large, the refresh number corresponding to the flag value “1” is changed from “2” to “0”, and the refresh number corresponding to the flag value “2”. Is changed from “3” to “2”. Thereby, as shown in the upper row of FIG. 6, the total number of refreshed macroblocks is smaller in the third and fourth frames than in the second frame. However, in the fourth frame, the refresh is completed for all the macro blocks having the flag value “3”.

  On the other hand, when the generated code amount of the second frame is small, the refresh number corresponding to the flag value “1” is changed from “2” to “4”, and the refresh number corresponding to the flag value “2” is changed. “3” is changed to “4”. As a result, as shown in the lower row of FIG. 6, the total number of macroblocks to be refreshed is larger in the third and fourth frames than in the second frame. As a result, in the fourth frame, the refresh is completed not only for the macro block having the flag value “3” but also for all the flag values “1” and “2”.

  As described above, an increase in the generated code amount can be suppressed by changing the total number of refreshes according to the generated code amount. As a result, it is possible to suppress an increase in image quality degradation associated with code amount control.

  Further, by controlling the macro block group having a higher image quality degradation level so that the increase / decrease amount of the refresh number corresponding to the macro block group is reduced, the number of blocks to be refreshed in the region having the higher image quality degradation level is stabilized. be able to. For this reason, even if the generated code amount increases or decreases, it is possible to preferentially refresh an area where the image quality degradation degree is large. As a result, it is possible to suppress an increase in the amount of generated codes while suppressing image quality deterioration of the moving image frame.

(Imaging system)
As shown in FIG. 8, the moving image encoding apparatus according to each embodiment can be applied to an imaging system. The imaging system shown in FIG. 8 includes a lens (optical system) 30, a sensor 31, an analog / digital conversion circuit 32, a recording circuit 34, in addition to the image processing circuit 33 including the moving image encoding device 1. A reproduction circuit 35, a timing control circuit 36, and a system control circuit 37 are provided.

  In the imaging system 3, the image light incident through the lens (optical system) 30 is imaged on the sensor 31 and subjected to photoelectric conversion. The electrical signal obtained by the photoelectric conversion is converted into a digital signal by the analog / digital conversion circuit 32 and then supplied to the image processing circuit 33. The image processing circuit 33 executes Y / C processing, edge processing, image enlargement / reduction, image compression / expansion processing such as JPEG and MPEG, and control of an image compressed stream. The signal processed by the image processing circuit 33 is recorded in the recording circuit 34. The signal recorded in the recording circuit 34 is reproduced by the reproduction circuit 35. The moving image encoding apparatus 1 included in the sensor 31 and the image processing circuit 33 is controlled by a timing control circuit 36, and the lens 30, the recording circuit 34, the reproduction circuit 35, and the timing control circuit 36 are respectively a system control circuit 37. Controlled by

  Instead of the recording circuit 34, a transmission circuit that transmits the signal processed by the image processing circuit 33 via the Internet or the like may be provided. In this case, the reproduction circuit 35 reproduces the signal transmitted by the transmission circuit.

  In the imaging system shown in FIG. 8, the camera device or the like that photoelectrically converts the image light from the lens 30 by the sensor 31 and inputs the image light to the analog / digital conversion circuit 32 has been described, but the present invention is limited to this. In addition, the analog video input of an AV device such as a television may be directly supplied to the analog / digital conversion circuit 32.

(Other embodiments)
In each of the above embodiments, the sum of absolute differences (SAD) has been described as an example of the information indicating the degree of image quality degradation. You may use. Furthermore, a plurality of parameters indicating the image quality degradation degree may be used in combination.

  In addition, the moving image encoding apparatus according to each embodiment may be realized by hardware. For example, a moving image encoding apparatus is typically realized as an LSI that is an integrated circuit. The moving picture coding apparatus may be made into one chip, or may be made into one chip so as to include a part or all of the functional blocks constituting the moving picture coding apparatus. Although referred to as LSI, it may be called IC, system LSI, super LSI, or ultra LSI depending on the degree of integration. The method of circuit integration is not limited to LSI, and may be realized by a dedicated circuit or a general-purpose processor. An FPGA (Field Programmable Gate Array) that can be programmed after manufacturing the LSI or a reconfigurable processor that can reconfigure the connection and setting of circuit cells inside the LSI may be used. Furthermore, if integrated circuit technology comes out to replace LSI's as a result of the advancement of semiconductor technology or a derivative other technology, it is naturally also possible to carry out function block integration using this technology.

  Moreover, the moving image encoding apparatus according to each embodiment may be realized by software. For example, the control steps of the moving picture coding apparatus may be programmed, recorded on a recording medium (for example, a semiconductor memory or a hard disk), and a computer including a CPU may be operated by the program.

  The moving image coding method and apparatus according to the present invention can make the deterioration of image quality (subjective image deterioration) of moving image frames inconspicuous, and is useful for surveillance cameras and network cameras.

BRIEF DESCRIPTION OF THE DRAWINGS The block diagram of the moving image encoder by Embodiment 1 of this invention. The figure for demonstrating a refresh map. The figure for demonstrating concretely the refresh process by the moving image encoder shown in FIG. The block diagram of the moving image encoder by Embodiment 2 of this invention. The figure for demonstrating the correspondence of generated code amount and the number of refreshes. The figure for demonstrating concretely the refresh process by the moving image encoder shown in FIG. The block diagram of the imaging system by Embodiment 3 of this invention. The figure for demonstrating the conventional refresh process.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1, 2 Moving image encoder 11 Mode control part 12 Blocking part 13 Subtraction part 14 Orthogonal transformation part 15 Quantization part 16 Variable length encoding part 17 Inverse quantization part 18 Inverse orthogonal transformation part 19 Selector 20 Addition part 21 Memory 22 motion detection / compensation unit 23 intra prediction unit 101 image quality degradation level determination unit 102 refresh map generation unit 103 refresh number setting unit 104 refresh control unit 201 refresh number change unit 30 lens (optical system)
31 Sensor 32 Analog / Digital Conversion Circuit 33 Image Processing Circuit 34 Recording Circuit 35 Reproduction Circuit 36 Timing Control Circuit 37 System Control Circuit

Claims (14)

A method of sequentially encoding a plurality of video frames by intra-coding or inter-coding the video frames in units of a predetermined size block,
Classifying each of the plurality of blocks into a plurality of block groups according to the degree of image quality degradation of each of the plurality of blocks included in the first moving image frame;
Selecting a number of blocks corresponding to the image quality degradation degree of the block group from each of the plurality of block groups classified in the step (a) as candidate blocks;
The intra coding is performed on a block corresponding to the candidate block selected in the step (b) among a plurality of blocks included in a second moving image frame encoded after the first moving image frame. And (c) forcibly executing the moving picture encoding method. In claim 1,
In the step (b), the moving picture encoding method is characterized in that the number of candidate blocks selected from a block group having a higher image quality deterioration degree is increased. In claim 1 or claim 2,
Each of the plurality of block groups corresponds to a plurality of threshold values for the degree of image quality degradation,
The plurality of threshold values respectively correspond to a plurality of flag values;
Each of the plurality of flag values corresponds to a plurality of refresh numbers,
In the step (a), the image quality degradation degree of each of the plurality of blocks is compared with the plurality of threshold values, and a flag value corresponding to the comparison result is assigned to the block,
In the step (b), the same number of candidate blocks as the number of refreshes corresponding to the flag value are selected from the blocks to which the same flag value is assigned. In any one of Claims 1-3,
A moving picture coding method, further comprising a step (d) of changing the number of candidate blocks selected from each of the plurality of block groups in the step (b) according to a generated code amount. In claim 4,
In the step (d), from each of the plurality of block groups according to the generated code amount, the block group having a higher image quality degradation degree has a smaller increase / decrease amount of candidate blocks selected from the block group. A moving picture encoding method, wherein the number of candidate blocks to be selected is changed. In any one of Claims 1-5,
In the step (a), a moving picture coding method characterized in that the degree of image quality degradation of each of the plurality of blocks is evaluated based on a sum of absolute differences between the block and a prediction block corresponding to the block. . A program for causing a computer to execute the moving picture encoding method according to claim 1. An apparatus for sequentially encoding a plurality of moving image frames,
An encoding unit that performs intra encoding or inter encoding on a moving image frame in units of a predetermined size block;
Degradation of image quality for each of a plurality of blocks included in a moving image frame encoded by the encoding unit, and classifying each of the plurality of blocks into a plurality of block groups according to the degree of image quality deterioration A degree determination unit;
From each of the plurality of block groups classified by the image quality deterioration degree determination unit, a number of blocks corresponding to the image quality deterioration degree of the block group are selected as candidate blocks, and a plurality of blocks included in the moving image frame to be encoded next are selected. A moving picture encoding apparatus comprising: a refresh control unit that causes the encoding unit to execute the intra encoding on a block corresponding to the selected candidate block among blocks. In claim 8,
The moving picture coding apparatus, wherein the refresh control unit increases the number of candidate blocks to be selected from a block group having a higher degree of image quality degradation. In claim 8 or claim 9,
Each of the plurality of block groups corresponds to a plurality of threshold values for the degree of image quality degradation,
The plurality of threshold values respectively correspond to a plurality of flag values;
Each of the plurality of flag values corresponds to a plurality of refresh numbers,
The image quality deterioration degree determination unit compares the image quality deterioration degree of each of the plurality of blocks with the plurality of threshold values, and assigns a flag value corresponding to the comparison result to the block,
The moving picture coding method apparatus, wherein the refresh control unit selects the same number of candidate blocks as the number of refreshes corresponding to the flag value from among the blocks to which the same flag value is assigned. In any one of Claims 8-10,
The moving picture encoding apparatus, further comprising: a refresh number changing unit that changes the number of candidate blocks selected from each of the plurality of block groups by the refresh control unit according to a generated code amount. In claim 11,
The refresh number changing unit determines whether the block group having a higher image quality degradation degree has a smaller increase / decrease amount of the number of candidate blocks selected from the block group, from each of the plurality of block groups according to the generated code amount. A moving picture coding apparatus characterized by changing the number of candidate blocks to be selected. In any one of Claims 8-12,
In the step (a), a moving picture coding apparatus characterized in that an image quality degradation degree of each of the plurality of blocks is evaluated based on a sum of absolute differences between the block and a prediction block corresponding to the block. . An image processing circuit including the moving image encoding device according to any one of claims 8 to 13,
An imaging circuit that converts an image of a subject into an electrical signal;
An imaging system comprising: an analog / digital conversion circuit that converts an electrical signal from the imaging circuit into a digital signal and supplies the digital signal to the image processing circuit.

Images