JP2011124822A - Semiconductor integrated circuit, coding method and imaging device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce the amount of data transfer for image memory when performing coding processing of multi-view point image data. <P>SOLUTION: When performing coding processing, an encoder 13d which performs coding processing for second view point original image data that are output from an imaging part 12 detects motion using first view point original image data that are output from an imaging part 11 and stored in a data storage part 13a as reference image data. Thus, the amount of data transfer for image memory is reduced when performing coding processing of multi-view point image data. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a semiconductor integrated circuit, an encoding method, and an imaging apparatus.

  In recent years, with the development of digital moving image encoding processing technology, development of a display (3D display) capable of displaying a moving image in 3D (3D) has been advanced. Further, development of a three-dimensional moving image photographing device and a three-dimensional moving image coding device for photographing and compression-coding three-dimensional moving image data that can be displayed and reproduced on such a 3D display has been underway. As an international standard for 3D video coding, MVC (Multi-view Video Coding) of MPEG (Moving Picture Experts Group) -2, H.264, etc. H.264 MVC has been proposed.

  Conventionally, during motion detection processing required in 3D moving image encoding processing, a shift of the subject occurs only in the parallax direction between the left-eye viewpoint image and the right-eye viewpoint image. There has been proposed a method for reducing the amount of calculation by limiting the motion detection process in the left-right direction.

  In addition, there has been proposed a method for reducing the capacity of a buffer memory for temporarily storing image data by storing and storing image data to be encoded in units of slices instead of in units of frames.

JP 2000-165909 A JP 2009-4942 A

  In the conventional 3D moving image encoding process, image data captured by the camera of each viewpoint is temporarily stored in the image memory, and the encoding unit performs the encoding process as needed while reading the image data from the image memory, Encoded data is generated. For this reason, there has been a problem that the amount of data transferred to the image memory during the encoding process is increased compared to the case of one viewpoint.

  In view of the above, it is an object of the present invention to provide a semiconductor integrated circuit, an encoding method, and an imaging apparatus that can reduce the amount of data transferred to an image memory when encoding image data of a plurality of viewpoints. .

In order to achieve the above object, the following semiconductor integrated circuit is provided.
The semiconductor integrated circuit includes a first encoding unit that encodes original image data of a first viewpoint and a second encoding unit that encodes original image data of a second viewpoint. Then, when performing the encoding process, the second encoding unit performs motion detection using the original image data of the first viewpoint as reference image data.

  According to the disclosed semiconductor integrated circuit, encoding method, and imaging apparatus, it is possible to reduce the amount of data transferred to the image memory when encoding image data of a plurality of viewpoints.

It is a figure which shows the structure of the imaging device and semiconductor integrated circuit of embodiment. It is a figure which shows the structure of an example of the encoding part for left eye viewpoints. It is a figure which shows the structure of an example of the encoding part for right eye viewpoints. It is a figure which shows the mode of reading of the original image data from an imaging part. It is a figure which shows the process unit of an encoding part. It is the figure which showed the macroblock group of the kth line in a certain frame, and the macroblock group of the k + 1st line. It is a figure explaining a motion detection (the 1). It is a figure explaining a motion detection (the 2). It is a figure which shows an example of the encoding process in the encoding part for left eye viewpoints. It is a figure which shows an example of the reference area which the motion detection part of the encoding part for right eye viewpoints reads.

Embodiments of a semiconductor integrated circuit, an encoding method, and an imaging apparatus according to the present invention will be described below with reference to the drawings.
FIG. 1 is a diagram illustrating a configuration of an imaging device and a semiconductor integrated circuit according to an embodiment.

Here, an example of the imaging apparatus 10 that captures a moving image at two viewpoints is shown.
The imaging device 10 includes imaging units 11 and 12, a semiconductor integrated circuit 13, a bit stream memory 14, image memories 15 and 16, and a bit stream rearrangement unit 17.

For example, the imaging units 11 and 12 are arranged side by side in the horizontal direction, and perform imaging at each viewpoint and output original image data.
In the following description, it is assumed that the imaging unit 11 outputs the original image data of the left eye viewpoint and the imaging unit 12 outputs the original image data of the right eye viewpoint. It may be a left eye viewpoint. Further, the distance between the imaging units 11 and 12 is not limited to the interval between the human left eye and the right eye, and the interval can be widened or enhanced as appropriate in order to enhance the stereoscopic effect.

  The imaging units 11 and 12 include an imaging device such as a CCD (Charge Coupled Device) device or a CMOS (Complementary Metal-Oxide Semiconductor) device, and a conversion unit that converts RGB format image data into a YUV format. Yes. In FIG. 1, these configurations are not shown.

The semiconductor integrated circuit 13 compresses and encodes the original image data output from the imaging units 11 and 12.
The bit stream memory 14 stores a bit stream that is encoded data generated by the encoding process.

  The image memories 15 and 16 hold image data during the encoding process. The image memories 15 and 16 are frame memories, for example. In the imaging apparatus 10 of the present embodiment, the image memories 15 and 16 are used when encoding the original image data output from the imaging unit 11 and encode the original image data output from the imaging unit 12. It is not necessary when processing.

  As the bit stream memory 14 and the image memories 15 and 16, for example, a semiconductor memory such as a DRAM (Dynamic Random Access Memory) is applicable. For example, an area of one semiconductor memory may be divided and used as the bit stream memory 14 and the image memories 15 and 16.

  The bit stream rearrangement unit 17 reads the left-eye viewpoint and right-eye viewpoint bit streams from the bit stream memory 14, and rearranges the two bit streams in an appropriate order.

The semiconductor integrated circuit 13 of this embodiment includes data holding units 13a and 13b and encoding units 13c and 13d.
The data holding units 13a and 13b are memories such as a line buffer memory, for example. The data holding unit 13a temporarily accumulates and holds original image data output from the imaging unit 11, and the data holding unit 13b temporarily accumulates and holds original image data output from the imaging unit 12.

  The encoding unit 13c performs an encoding process on the original image data output from the left-eye viewpoint imaging unit 11 held in the image memory 15 via the data holding unit 13a. The encoding unit 13c stores reconstructed image data, which will be described later, in the image memory 16, and uses it as reference image data in motion detection.

  The encoding unit 13d performs encoding processing on the original image data output from the right-eye viewpoint imaging unit 12 held in the data holding unit 13b. However, the encoding unit 13d uses original image data captured by the left-eye viewpoint imaging unit 11 and held in the data holding unit 13a as reference image data used for motion detection. As a result, the image memories 15 and 16 can be used only when encoding the original image data of the left eye viewpoint, so that the data transfer amount to the image memory can be reduced.

Hereinafter, an exemplary configuration of the encoding units 13c and 13d will be described.
(Configuration of the encoding unit 13c for the left eye viewpoint)
FIG. 2 is a diagram illustrating a configuration of an example of an encoding unit for the left eye viewpoint.

Here, as an example, 2 is a block diagram of an encoding unit corresponding to the H.264 / AVC standard.
The encoding unit 13c includes an error image generation unit 21, an orthogonal transform / quantization unit 22, an entropy encoding unit 23, an inverse quantization / inverse orthogonal transform unit 24, a reconstructed image generation unit 25, a line memory 26, and intra prediction. A portion 27 is provided. Furthermore, the encoding unit 13 c includes a deblocking filter unit 28, a motion detection unit 29, a motion compensation unit 30, and a prediction method switching unit 31.

The error image generation unit 21 generates error image data from the difference between the target macroblock of the original image data of the left eye viewpoint held in the image memory 15 and the predicted image data.
The orthogonal transform / quantization unit 22 performs orthogonal transform and quantization processing on the error image data to obtain a quantized transform coefficient.

The entropy encoding unit 23 entropy-encodes the quantized transform coefficient to generate a bitstream that is image information whose information amount is compressed.
The inverse quantization / inverse orthogonal transform unit 24 performs inverse quantization and inverse orthogonal transform on the quantized transform coefficient to restore error image data.

The reconstructed image generation unit 25 generates reconstructed image data from the restored error image data and the predicted image data used in the error image generation unit 21.
The line memory 26 stores the reconstructed image data generated by the reconstructed image generation unit 25.

The intra-screen prediction unit 27 reads image data of the same screen as the screen to which the target macroblock belongs from the line memory 26, and generates an intra-screen prediction image.
The deblocking filter unit 28 performs a process of reducing block distortion generated in the reconstructed image data by the orthogonal transform / quantization process and the inverse quantization / inverse orthogonal transform process, and the processed reconstructed image data is stored in the image memory. 16 is written.

The motion detection unit 29 detects a motion vector based on the input macro block to be processed and the reference image data.
The motion compensation unit 30 extracts an image block having coordinates corresponding to the motion vector detected by the motion detection unit 29 from the reference image data read from the image memory 16, and generates inter-screen prediction image data.

For example, the prediction method switching unit 31 switches between an intra-screen prediction method and an inter-screen prediction method as a prediction image data generation method according to a control signal from a control unit (processor or the like) (not shown) of the imaging device 10. Do.
(Configuration of the encoding unit 13d for the right eye viewpoint)
FIG. 3 is a diagram illustrating a configuration of an example of an encoding unit for the right eye viewpoint.

  The encoding unit 13d includes an error image generation unit 41, an orthogonal transform / quantization unit 42, an entropy encoding unit 43, an inverse quantization / inverse orthogonal transform unit 44, a reconstructed image generation unit 45, a line memory 46, and intra prediction. A portion 47 is provided. Furthermore, the encoding unit 13d includes a motion detection unit 48, a motion compensation unit 49, and a prediction method switching unit 50.

  The function of each unit is almost the same as the encoding unit 13c for the left eye viewpoint described above, but in the encoding unit 13d for the right eye viewpoint, the error image generation unit 41 and the motion detection unit 48 are processed from the data holding unit 13b. The difference is that the macro block is input. Further, the difference is that the motion detection unit 48 and the motion compensation unit 49 input the original image data of the left eye viewpoint held in the data holding unit 13a as reference image data.

  Unlike the encoding unit 13c for the left eye viewpoint, the encoding unit 13d uses the original image data for the left eye viewpoint stored in the data storage unit 13a as the reference image data. Read / write is unnecessary.

Hereinafter, the operation of the imaging apparatus 10 of the present embodiment will be described in detail.
(Operation of the imaging device 10)
Original image data captured and output by the imaging units 11 and 12 are temporarily stored and held in the data holding units 13a and 13b, respectively.

FIG. 4 is a diagram illustrating how the original image data is read from the imaging unit.
As shown in FIG. 4, the image capturing units 11 and 12 read the original image data in raster scan order (one line from the top line to the bottom line, with each line from the leftmost pixel to the rightmost pixel). And stored in the data holding units 13a and 13b.

  Thereafter, the original image data of the left eye viewpoint accumulated in the data holding unit 13 a is held in the image memory 15. The left-eye viewpoint encoding unit 13 c performs encoding processing while reading original image data from the image memory 15. One right-eye viewpoint encoding unit 13d performs encoding processing while reading the original image data of the right-eye viewpoint held in the data holding unit 13b.

FIG. 5 is a diagram illustrating a processing unit of the encoding unit.
The encoding units 13c and 13d perform encoding processing in a frame 60 as shown in FIG. 5, for example, using a macroblock 61 composed of 16 × 16 pixels as a processing unit.

FIG. 6 is a diagram illustrating a macroblock group in the kth row and a macroblock group in the k + 1th row in a certain frame.
The data holding unit 13b acquires and stores the original image data of the macro block group 62 (k + 1) in the (k + 1) th row from the imaging unit 12 at the right eye viewpoint line by line. Meanwhile, the data holding unit 13b sends the original image data of the macroblock 61 of the macroblock group 62 (k + 1) in the acquired k-th row to the encoding unit 13d. The data holding unit 13b repeats the same processing from the first line to the last line of the frame 60.

  Thereby, efficient encoding processing becomes possible. As the data holding unit 13b, for example, a small memory that can store image data of a macroblock group for two rows can be used. The image data of the macroblock group for two rows corresponds to image data for 32 lines when the macroblock 61 is composed of 16 × 16 pixels.

  In the other data holding unit 13a, while the right-eye viewpoint encoding unit 13d processes the k-th macroblock group 62k, the reference image data used for motion detection is used as the reference image data in the k-th row of the left-eye viewpoint frame. The image data of the macro block group is sent to the encoding unit 13d. At the same time, the data holding unit 13a acquires and stores the image data of the macro block group in the (k + 1) th row from the imaging unit 11 at the left eye viewpoint. The data holding unit 13a repeats the same processing from the first line to the last line of the frame 60.

Therefore, the capacity of the data holding unit 13a is only required to store, for example, image data of a macroblock group for two rows, and a small memory can be used.
The reason why the encoding unit 13d uses the same row as the right-eye viewpoint macroblock group 62k as the reference image data will be described later (see FIG. 10).

Next, details of the operations of the encoding units 13c and 13d that perform the encoding process in units of processing as described above will be described.
First, the encoding process in the encoding unit 13c for the left eye viewpoint will be described with reference to FIG.
(Encoding process by the encoding unit 13c for the left eye viewpoint)
When a macroblock to be processed is input from the image memory 15, the error image generation unit 21 generates error image data that is a difference image between the macroblock and the predicted image data. The orthogonal transform / quantization unit 22 receives error image data, performs orthogonal transform and quantization processing, and obtains a quantized transform coefficient. The quantized transform coefficient is supplied to the entropy coding unit 23 and the inverse quantization / inverse orthogonal transform unit 24.

  The entropy encoding unit 23 entropy-encodes the quantized transform coefficient to generate a bitstream that is image information in which the information amount is compressed. The entropy encoding unit 23 entropy-encodes information related to the motion vector together with the quantized transform coefficient, and includes information related to the motion vector in the bitstream. A motion vector detection method will be described later.

  On the other hand, the inverse quantization / inverse orthogonal transform unit 24 performs inverse quantization and inverse orthogonal transform on the quantized transform coefficients to restore error image data. Then, the reconstructed image generation unit 25 generates reconstructed image data from the restored error image data and the predicted image data used by the error image generation unit 21 and stores the reconstructed image data in the line memory 26. In addition, the deblocking filter unit 28 performs a process of reducing block distortion generated in the reconstructed image data by the orthogonal transform / quantization process and the inverse quantization / inverse orthogonal transform process, and the reconstructed image data after the process Are stored in the image memory 16.

Generation of predicted image data used in the error image generation unit 21 and the reconstructed image generation unit 25 is performed as follows.
The intra-screen prediction unit 27 reads out image data of the same screen as the screen to which the target macroblock belongs from the line memory 26, and generates intra-screen prediction image data. This intra-screen prediction image data is supplied to the entropy encoding unit 23, and when the intra-screen prediction method is selected by the prediction method switching unit 31, it is sent to the error image generation unit 21 and the reconstructed image generation unit 25. Supplied.

On the other hand, the inter-screen prediction image data is generated as follows.
7 and 8 are diagrams for explaining motion detection.
7A shows a macroblock 71 of a frame 70 at a certain time and an enlarged view thereof, and FIG. 7B shows a reference area 81 in the reference frame 80 and an enlarged view thereof.

  Further, in FIG. 8, pixel positions in the reference area 81 and the macroblock are shown in an xy coordinate system. Further, the motion vector 82 is shown with the position in the reference area 81 corresponding to the upper left corner of the macroblock 71 as the origin (0, 0).

  The motion detection unit 29 searches the reference frame for an image region that is most similar to the macro block to be processed. For example, when the motion detection unit 29 acquires the macro block 71 to be processed as shown in FIG. 7A from the image memory 15, the reference frame 80 as shown in FIG. 7B held in the image memory 16. From this, the image data of the reference area 81 is acquired as reference image data. As the reference area 81, for example, in the reference frame 80, a rectangular area centered on a macroblock located at the same position as the macroblock 71 to be processed is selected.

  Then, the motion detection unit 29 detects a position most similar to the processing target macroblock 71 from the reference area 81. The motion detection unit 29 uses, for example, an absolute difference sum (SAD) as a similarity criterion. SAD (X, Y) between the macro block 71 to be processed and the predicted image 83 at the position of the motion vector 82 shown in FIG. 8 is calculated by the following equation (1).

SAD (X, Y) = Σ | C (i, j) −R (X + i, Y + j) | (1)
Here, i = 0, 1, 2,..., 15, j = 0, 1, 2,..., 15 and Σ represents cumulative addition for 256 pixels. C (i, j) represents the pixel value of the coordinate (i, j) of the macro block 71 to be processed, and R (X + i, Y + j) represents the pixel value of the coordinate (X + i, Y + j) of the reference area 81. Since the SAD is small when the difference between C and R is small, that is, when the pixel value of the macroblock 71 to be processed is close to the pixel value of the predicted image 83, the predicted image 83 having a smaller SAD is more similar. Become. The motion detector 29 detects a motion vector 82 having the smallest SAD from arbitrary motion vectors in the reference area 81.

  In this way, if the motion vector 82 is selected so that the predicted image 83 is the image most similar to the macro block 71 to be processed, the information amount of the error image is minimized, and finally the bit after entropy coding is used. The amount of stream information is the smallest.

  The motion vector detected by the motion detection unit 29 is supplied to the motion compensation unit 30 and the entropy encoding unit 23. The motion compensation unit 30 that has acquired the motion vector extracts an image block having coordinates corresponding to the motion vector from the reference image data read out from the image memory 16, and generates inter-screen prediction image data.

The inter-screen prediction image data is supplied to the error image generation unit 21 and the reconstructed image generation unit 25 when the inter-screen prediction method is selected by the prediction method switching unit 31.
By the way, the image data obtained by the imaging by the imaging unit 11 for the left eye viewpoint is once stored in the image memory 15 and then processed while being read out in the order of performing the encoding process from the encoding unit 13c. Can be referred to.

FIG. 9 is a diagram illustrating an example of the encoding process in the encoding unit for the left eye viewpoint.
The horizontal axis is time. FIG. 9 shows an example of a frame acquired at each shooting time and an example of a frame encoded at each encoding time. Further, FIG. 9 shows the accumulation state of the original image data in the image memory 15 and the accumulation state of the reconstructed image data in the image memory 16 at each time.

  In FIG. 9, from the time t0 to t1, the original image data (frame I0) of the left viewpoint imaged by the imaging unit 11 is held in the image memory 15 via the data holding unit 13a. Similarly, the image memory 15 holds the frame B1 from time t1 to t2, the frame B2 from time t2 to t3, and the frame P3 from time t3 to t4. Subsequent left viewpoint original image data is also stored in the image memory 15 in the same manner.

  The encoding unit 13c for the left eye viewpoint performs encoding processing of frame I0 from time t3 to t4, frame P3 from time t4 to t5, frame B1 from time t5 to t6, and frame B2 from time t6 to t7. The encoding unit 13c performs the same encoding process on the subsequent left-viewpoint original image data.

  Frames I0, I15,... Represent intra frames. In an intra frame, in the encoding process shown in FIG. 9, all macroblocks in a frame are encoded only by intra prediction.

  Frames P3, P6, P9,... Represent frames that are encoded using the immediately preceding frame In or frame Pn as a reference area. For example, the frame P3 is encoded using the frame I0, the frame P6 is encoded using the frame P3, and the frame P9 is encoded using the frame P6 as the reference frame.

  Frames B1, B2, B4, B5,... Are encoded using two frames, ie, the immediately preceding frame In or frame Pn and the immediately following frame In or frame Pn as reference frames. For example, the frame B1 is encoded using the image data of the frames I0 and P3 held in the image memory 16 as reconstructed image data as reference image data. Although the frame B1 has a shooting time before the frame P3, such a reference relationship is possible. Similarly, for example, frame B2 is encoded using frame I0 and frame P3, frame B4 is encoded using frame P3 and frame P6, and frame B5 is encoded using frame P3 and frame P6 as reference frames.

  In this way, the encoding unit 13c for the left viewpoint can select a reference region from any two reference frames and detect a motion vector in the frame Bn, so that a prediction having a smaller SAD can be performed. An image can be selected.

Next, the encoding process in the encoding unit 13d for the right eye viewpoint will be described with reference to FIG.
(Encoding process by the encoding unit 13d for the right eye viewpoint)
When the processing target macroblock is input from the data holding unit 13b, the error image generation unit 41 generates error image data that is a difference image between the macroblock and the predicted image data. The orthogonal transform / quantization unit 42 receives the error image data, executes orthogonal transform and quantization processing, and obtains a quantized transform coefficient. The quantized transform coefficients are supplied to the entropy coding unit 43 and the inverse quantization / inverse orthogonal transform unit 44.

  The entropy encoding unit 43 entropy encodes the quantized transform coefficient, and generates a bit stream that is image information in which the information amount is compressed. The entropy encoding unit 43 entropy-encodes information related to the motion vector together with the quantized transform coefficient, and includes information related to the motion vector in the bitstream.

  On the other hand, the inverse quantization / inverse orthogonal transform unit 44 performs inverse quantization and inverse orthogonal transform on the quantized transform coefficients to restore error image data. Then, the reconstructed image generation unit 45 generates reconstructed image data from the restored error image data and the predicted image data used by the error image generation unit 41 and stores the reconstructed image data in the line memory 46.

  Of the prediction image data used in the error image generation unit 41 and the reconstructed image generation unit 45, the generation of the intra-screen prediction image data is the same as the processing of the encoding unit 13d for the left eye viewpoint described above. On the other hand, the generation of inter-screen prediction image data is different from the processing of the encoding unit 13d for the left eye viewpoint described above.

  The motion detection unit 48 inputs the macroblock to be processed and the original image data of the left-eye viewpoint input at the same time as reference image data from the data holding unit 13a. When the motion detection of the macroblock at the right eye viewpoint is performed using the original image data of the left eye viewpoint at the same time, the subject is captured with parallax in the horizontal direction, which is the installation direction of the two imaging units 11 and 12, and thus detection is performed The power vector is a horizontal vector. That is, the vertical component of the motion vector is zero.

FIG. 10 is a diagram illustrating an example of a reference area read by the motion detection unit of the encoding unit for the right eye viewpoint.
The case where the motion detection unit 48 performs motion detection on the macroblock 90 to be processed in the image data of the right eye viewpoint is illustrated.

  As described above, when the imaging units 11 and 12 are arranged in the horizontal direction, the motion detection unit 48 searches for a motion vector only in the horizontal direction. Therefore, the original image data of the left eye viewpoint read out from the data holding unit 13a as the reference image data by the motion detection unit 48 is, for example, an area 92 positioned in the horizontal direction with respect to the macro block 91 at the same position as the macro block 90 to be processed. It becomes. The area 92 has a size expanded by 16 × 16 pixels left and right, for example, with the macro block 91 as the center. The size of the region 92 can be changed as appropriate according to the distance between the imaging units 11 and 12.

  The bit streams generated by the encoding units 13c and 13d as described above are written to the bit stream memory, rearranged in an appropriate order by the bit stream rearrangement unit 17, and then written back to the bit stream memory 14. .

  Thereafter, the bit stream of each viewpoint is stored in, for example, a storage device such as a hard disk drive (not shown) or a storage medium such as a DVD (Digital Versatile Disc), or decoded by a 3D display device and displayed in 3D. Note that the imaging device 10 itself may include a storage device, a decoding unit, a display unit, or the like. The decoding unit may be built in the semiconductor integrated circuit 13.

  By the way, when decoding the left-eye viewpoint image data encoded as described above, a decoding unit (not shown) decodes the encoded error image data, for example, and decodes the error image data and the previous frame of the error image data. Image data is generated from the decoded data of the left eye viewpoint. In addition, when decoding the right-eye viewpoint image data at a certain time, the decoding unit decodes the encoded error image data, and the image data is obtained from the error image data and the left-eye viewpoint decoded data at the same time. Should be decrypted.

  As described above, according to the semiconductor integrated circuit 13, the encoding method, and the imaging apparatus 10 of the present embodiment, the original image data of the left eye viewpoint is used as the reference image data used for encoding the original image data of the right eye viewpoint. To do. As a result, the image memories 15 and 16 can be used only when encoding the original image data of the left-eye viewpoint. Therefore, in the encoding process of the original image data of a plurality of viewpoints, data transfer to the image memory is performed. The amount can be reduced.

  This also reduces the amount of data that is read and written per unit time, so that an inexpensive semiconductor memory can be used as the image memory, and the cost of the imaging apparatus 10 can be reduced.

  As described above, one aspect of the semiconductor integrated circuit, the encoding method, and the imaging apparatus of the present invention has been described based on the embodiments. However, these are merely examples, and the present invention is not limited to the above description.

  For example, in FIGS. Although the encoding units 13c and 13d corresponding to the H.264 / AVC standard have been described, an encoding unit corresponding to MPEG-2 may be used. In this case as well, when motion detection is performed in the encoding unit that encodes the original image data of one viewpoint, the same effect can be obtained by using the original image data of the other viewpoint as the reference image data.

  In the above description, the case of two viewpoints has been described. However, the present invention can be similarly applied when encoding original image data of three or more viewpoints. In that case, a data holding unit and an encoding unit corresponding to each viewpoint are provided, and the original image data of the reference viewpoint is used as reference image data in motion detection when encoding a macroblock of a certain viewpoint. It is possible to obtain the same effect.

DESCRIPTION OF SYMBOLS 10 Image pick-up device 11, 12 Image pick-up part 13 Semiconductor integrated circuit 13a, 13b Data holding part 13c, 13d Encoding part 14 Bit stream memory 15, 16 Image memory 17 Bit stream rearrangement part

Claims (5)

A first encoding unit that encodes original image data of the first viewpoint;
A second encoding unit that encodes the original image data of the second viewpoint;
Have
The second encoding unit, when performing the encoding process, performs motion detection using the original image data of the first viewpoint as reference image data. A first data holding unit that holds the original image data of the first viewpoint out of the first viewpoint and the second viewpoint located in a horizontal direction;
The first data holding unit converts the original image data of the first viewpoint positioned in a horizontal direction on the screen with respect to the original image data of the processing unit in the second encoding unit, and the reference image The semiconductor integrated circuit according to claim 1, wherein the data is sent as data to the second encoding unit. A second data holding unit that holds the original image data of the second viewpoint among the first viewpoint and the second viewpoint located in a horizontal direction;
While the second data holding unit acquires and stores one line of the original image data, which is a processing unit of the encoding process in the second encoding unit, the processing of the other acquired lines 3. The semiconductor integrated circuit according to claim 1, wherein the original image data of a unit is sent to the second encoding unit. The first encoding unit encodes the original image data of the first viewpoint,
When performing the encoding process, the second encoding unit that encodes the original image data of the second viewpoint uses the original image data of the first viewpoint as reference image data, and performs the second viewpoint. An original encoding method for encoding original image data. A first imaging unit that performs imaging from a first viewpoint and outputs original image data;
A second imaging unit that performs imaging from a second viewpoint and outputs original image data;
A first encoding unit that encodes the original image data of the first viewpoint;
A second encoding unit that encodes the original image data of the second viewpoint;
Have
The second encoding unit, when performing the encoding process, performs motion detection using the original image data of the first viewpoint as reference image data.

Images