JP5012493B2 - VIDEO OUTPUT DEVICE, VIDEO OUTPUT METHOD, VIDEO OUTPUT PROGRAM, VIDEO PROCESSING SYSTEM, VIDEO PROCESSING DEVICE, VIDEO PROCESSING METHOD, AND VIDEO PROCESSING PROGRAM - Google Patents

Description

  The present invention relates to an apparatus for converting image data into a YUV video signal and outputting the video data, and in particular, a video output apparatus suitable for converting floating-point format image data into a YUV video signal and outputting it, The present invention relates to a video output method, a video output program, a video processing system, a video processing apparatus, a video processing method, and a video processing program.

Conventionally, in order to compensate for the narrowness of the dynamic range (hereinafter referred to as “D range”) of an image pickup device, an image pickup device capable of shooting with two or more types of exposure times in one frame period has been proposed. In addition, in an image sensor that constitutes one frame by two fields, a method of imaging by changing the exposure time in odd and even fields has been proposed.
In other words, the image data obtained by photographing with the above two or more types of exposure time (for example, standard exposure time T1, short exposure time T2 (T1> T2), etc.) is synthesized outside or inside the sensor, and thus D Obtain image data with an expanded range. Image data having a relatively wide D range is generally referred to as HDR (High Dynamic Range) image data.

As a technique related to the expansion of the D range of captured image data as described above, for example, there is an imaging apparatus described in Patent Document 1.
The imaging device of Patent Document 1 reads out charges accumulated at different exposure times a plurality of times, and pixel data that is saturated in a longer exposure time among pixel data composed of the read charges with different exposure times. Is replaced with pixel data that is not saturated in a shorter exposure time, thereby obtaining image data with an expanded D range.

Also, image data with an expanded dynamic range obtained in an imaging apparatus such as Patent Document 1 is often output as fixed-point format data after reducing the amount of information in accordance with the performance of the output device. . For example, this range of information reduces the D range to about 50 dB.
On the other hand, the problem of overexposure and underexposure due to the narrowness of the D range is a lack of pixel information, which can be a problem when the system performs a desired operation. In particular, overexposure is fatal in machine vision because it cannot be determined whether it is an image or a defect. In addition, machine vision captures images globally (macro), such as area detection (labeling), edge detection, correlation detection, and binarization processing. Information loss greatly affects system performance such as recognition.

As a means for solving this problem, an image processing system that transmits and receives HDR video has attracted attention.
As an HDR image format, a format called Radiance that expresses pixel values in floating point is known. The Radiance format is expressed in a floating point format of 32 bits per pixel by adding an exponent E (8 bits) in addition to RGB (8 bits each) of each pixel. When the pixel value is expressed in the Radiance format (Ri, Gi, Bi, E), the actual luminance value of each color is expressed in the floating-point format shown in the following equation (1).

R = Ri × 2 ^(E-128)
G = Gi × 2 ^(E-128)
B = Bi × 2 ^(E-128) (1)

  Next, the output format of the video signal between the video signal source such as an electronic camera on the video output side and the video processing / display device (hereinafter simply referred to as the video processing device) on the video receiving side will be described. To do.

  In the case of a color image, the output from the video signal source is output to the video processing device in a format such as an RGB signal or a YUV signal (luminance signal and color difference signal). YUV signals are standardized by ITU-R BT.601 by the International Telecommunication Union (ITU). This standard is used for NTSC, JPEG, and the like in a method that reduces the amount of information in a range that is not recognized by the human eye in consideration of the visual characteristics of the human eye. NTSC employs a YUV method capable of reducing the amount of information for the purpose of effectively utilizing a limited frequency band, and JPEG for the purpose of reducing the file size.

  The YUV method includes a Y signal indicating a luminance signal, and a U signal and a V signal (collectively referred to as UV signals) indicating color difference signals. The Y signal is output to the video processing device for the total number of pixels forming the image. On the other hand, the UV signal uses the low spatial resolution of the human eye with respect to color, and the amount of information to be output is reduced (thinning-out processing). For example, the amount of information to be output for the U signal and V signal in the YUV411 format is reduced to ¼ compared to when the U signal and V signal are output for the total number of pixels forming the image.

An interface between the video signal source and the video processing apparatus will be described. For example, the YUV411 format is often used to output video data between a video signal source such as an electronic camera and a video processing device for the purpose of improving output efficiency by reducing the amount of video information.
At this time, when the video signal source outputs a video signal to the video processing device via a predetermined interface at a frame rate of 30 [fps (frame / sec)], for example, FIG. As shown, there are three 8-bit ports for Y signal output, 8-bit ports for color difference signal U and V signal output, and synchronization signal ports for establishing synchronization between the video signal source and the video processing device. The port is used to achieve high-speed data transmission. Further, as shown in FIG. 13B, a configuration for realizing output of a YUV video signal by using two ports of an 8-bit port for transmitting Y signal and UV signal and the synchronous signal port. There is also.

In FIG. 13A, the sync signal port outputs three signals, ie, a vertical sync signal indicating the start of a frame, a horizontal sync signal indicating the start of a line, and a pixel sync signal for synchronizing pixels. Is output as a decoded synchronization signal. Since the Y signal port outputs all image data, the data is output in synchronization with the pixel synchronization signal. On the other hand, since the UV signal is thinned out at the UV output port, the color difference signal U and the V signal are multiplexed and output (switched in the order of U and V). In addition, because there is no UV signal to be output in the even-numbered line due to thinning, dummy data is output.
JP-A-5-22670

  However, in the above prior art, for example, when the image data expanded in the D range in the imaging apparatus of Patent Document 1 is output to the video processing apparatus as floating point format data and in the YUV format, In addition to the Y signal port and the UV signal port, it is necessary to provide a port for outputting an exponent data signal.

  Therefore, the present invention has been made paying attention to such an unsolved problem of the prior art, and converts image data in a floating-point format into YUV video data using an existing interface. It is an object of the present invention to provide a video output device, a video output method and a video output program, a video processing system, a video processing device, a video processing method, and a video processing program that are suitable for output.

[Mode 1] In order to achieve the above object, a video output apparatus according to mode 1 includes:
A video output device that converts image data into a YUV video signal including a luminance signal Y and a color difference signal UV, and outputs the video signal obtained by the conversion via a predetermined interface,
The pixel value is expressed in a floating-point format with a mantissa part of N bits (N is a natural number of 2 or more), a radix of X (X is a natural number of 2 or more), and an exponent part of M bits (M is a natural number of 2 or more). Mantissa conversion means for converting the mantissa data of each pixel data constituting the first image data into the YUV video signal for the first image data composed of the pixel data;
Multiplexing means for multiplexing the data of the exponent part of each pixel data constituting the first image data into the color difference signal UV in the YUV video signal;
Video signal output means for outputting a YUV video signal comprising the luminance signal Y and the color difference signal UV in which the exponent data is multiplexed.

  With this configuration, the mantissa conversion means converts the pixel value into a mantissa part of N bits (N is a natural number of 2 or more), a base number of X (X is a natural number of 2 or more), and an exponent part of M bits ( For image data composed of pixel data expressed in floating point format (M is a natural number of 2 or more), the data of the mantissa part of each pixel data of the image data can be converted into the YUV video signal. It is possible to multiplex the exponent data in each pixel data of the image data into the color difference signal UV in the YUV video signal by the multiplexing means, and the luminance signal Y by the video signal output means. It is also possible to output a YUV video signal composed of the color difference signal UV in which the exponent data is multiplexed.

Accordingly, the exponent data of each pixel data in the floating-point format can be multiplexed and output on the color difference signal UV, so that the YUV is not provided separately without providing an interface (port) for outputting the exponent data. It is possible to output an image signal including exponent data by using an existing interface of the system as it is.
Further, since the floating point format image data can be handled, the floating point format HDR image data can be easily output using an existing interface. That is, it is possible to easily output HDR image data having a sufficient contrast ratio and dynamic range that does not cause overexposure or blackout. For example, in the case of HDR image data in a floating-point format with a mantissa part 8 bits and an exponent part 4 bits, it is about 32000 times (90 dB) compared to the HDR image data represented by a conventional 8-bit integer value. Dynamic range can be expressed.

[Mode 2] Furthermore, the video output device of mode 2 is the video output device of mode 1,
Exponent part conversion means for converting the M-bit exponent part of each pixel data constituting the first image data into an m-bit exponent part (m is a natural number of M>m);
In the second image data, which is image data after the exponent part is converted by the exponent part conversion means, for each pixel data of a predetermined number of pixels continuous in a predetermined scanning direction, an exponent in the pixel data of the predetermined number of pixels Packing index part data generating means for generating packing index part data composed of parts of data,
The mantissa conversion means converts mantissa data of each pixel data constituting the second image data into the YUV video signal,
The multiplexing means multiplexes the packing index part data with a color difference signal UV in the YUV video signal.

With such a configuration, the data output amount can be reduced by reducing the data amount of the exponent part of each pixel data in the floating point format.
Further, by converting the exponent data to the number of bits that can be handled easily by the YUV method, the exponent data of the floating point format image data can be easily multiplexed with the color difference signal UV. It is done.

[Mode 3] Furthermore, the video output device of mode 3 is the video output device of mode 2,
The first image data is color image data in RGB (Red, Green, Blue) format, and a value representing each of R, G, and B color elements constituting the pixel value has a mantissa part of the N bits. The radix is 2 and the exponent is represented in the M-bit floating point format, and the value indicated by the M-bit exponent is composed of pixel data having the same value in each color element,
The exponent part converting means converts the M-bit exponent part corresponding to each color element of each pixel data constituting the first image data into an exponent part of 4 bits or less having the same value in each color element. And
The mantissa conversion means converts mantissa data corresponding to each color element of each pixel data constituting the second image data into a YUV411 video signal that is one type of the YUV method,
In the second image data, the packing index part data generation unit is configured to include, for each pixel data of four pixels continuous in the horizontal scanning direction, packing index part data composed of index part data in the pixel data of the four pixels. Is generated.

With such a configuration, for example, a 4-byte luminance signal Y, a 1-byte color difference signal U, a 1-byte color difference signal V, and a 2-byte color difference signal for each pixel region of 2 pixels × 2 pixels of the image. In the video output using the YUV411 system in which a dummy signal is output, the packing index part data can be configured with 2 bytes or less.
Thereby, the packing index portion data can be multiplexed with the color difference signal UV instead of the data of 2 bytes thinned out when the color difference signal U and the color difference signal V are multiplexed (corresponding to the dummy signal). Therefore, it is possible to easily multiplex the data of the exponent part with the color difference signal UV.

[Mode 4] Furthermore, the video output device of mode 4 is the video output device of mode 3,
The first image data is expressed in a floating-point format in which values representing the R, G, and B color elements constituting the pixel value are 8 bits for the mantissa, 2 for the radix, and 8 bits for the exponent. And the value indicated by the 8-bit exponent part is composed of pixel data having the same value in each color element,
The exponent part conversion means converts the 8-bit exponent part corresponding to each color element of each pixel data constituting the first image data into a 4-bit exponent part having the same value in each color element. It is characterized by that.

With such a configuration, a 4-byte luminance signal Y, a 1-byte color difference signal U, a 1-byte color difference signal V, and a 2-byte dummy signal for each pixel region of 2 pixels × 2 pixels of the image. In the video output using the YUV411 system, the packing index part data can be composed of 2 bytes.
As a result, the 2-byte packing index part data is multiplexed in place of the 2-byte data (corresponding to the dummy signal) thinned out when the color difference signal U and the color difference signal V are multiplexed with the color difference signal UV. Therefore, it is possible to obtain an effect that the data of the exponent part can be multiplexed with the color difference signal UV without waste and easily.

[Mode 5] Furthermore, the video output device of mode 5 is the video output device of any one of modes 1 to 4,
The multiplexing means multiplexes the data of the exponent part with the color difference signal UV corresponding to either an odd line or an even line of an image.
With such a configuration, when the color difference signal UV is output, conventionally, in the signal output of an image line (either an even line or an odd line) for outputting a dummy signal, an exponent part is used instead of the dummy signal. Therefore, it is possible to easily multiplex and output the data of the exponent part without greatly changing the configuration of the existing image data processing means on the video output side using the YUV method. The effect that it can be obtained.

[Mode 6] On the other hand, in order to achieve the above object, the video processing system of mode 6
A video processing system comprising a video output device and a video processing device, the video processing device generating display data capable of displaying video in a predetermined display format based on a video signal output from the video output device. And
The video output device
The pixel value is expressed in a floating-point format with a mantissa part of N bits (N is a natural number of 2 or more), a radix of X (X is a natural number of 2 or more), and an exponent part of M bits (M is a natural number of 2 or more). A mantissa for converting mantissa data in each pixel data of the first image data into YUV video signal composed of the luminance signal Y and the color difference signal UV with respect to the first image data composed of the pixel data. Part conversion means;
Multiplexing means for multiplexing the data of the exponent part of each pixel data constituting the first image data into the color difference signal UV in the YUV video signal;
A YUV video signal comprising the luminance signal Y in the YUV video signal obtained by the conversion by the mantissa conversion means and the color difference signal UV in which the exponent data is multiplexed by the multiplexing means. Video signal output means for outputting,
The video signal output means outputs the luminance signal Y and the color difference signal UV to the video processing device independently of each other,
The video processing device includes:
A signal separation means for separating the color difference signal UV multiplexed with the exponent data input from the video output device into the mantissa color difference signal UV and the exponent data signal;
A format converting means for converting the luminance signal Y input from the video output device and the color difference signal UV obtained by the separation into a signal in a data format of a mantissa before being converted into the YUV format;
Display data generating means for generating display data based on the data signal of the exponent part separated by the signal separating means and the signal of the data format of the mantissa part obtained by conversion by the format converting means. Features.

  With such a configuration, the video output device uses the mantissa conversion unit to convert the pixel value into a mantissa part of N bits (N is a natural number of 2 or more), a radix of X (X is a natural number of 2 or more), and an exponent. For image data composed of pixel data expressed in floating-point format with M bits (M is a natural number of 2 or more), the mantissa data in each pixel data of the image data is converted into the YUV video signal. The data of the exponent part in each pixel data of the image data can be multiplexed with the color difference signal UV in the YUV video signal by the multiplexing means, and the video signal output means Thus, it is possible to output a YUV video signal composed of the luminance signal Y and the color difference signal UV obtained by multiplexing the exponent data.

  Further, the video processing device can separate the color difference signal UV input from the video output device into a mantissa color difference signal UV and an exponent data signal by means of signal separation means, The format conversion means converts the luminance signal Y and the color difference signal UV separated from the exponent data input from the video output device into a signal in the mantissa data format before conversion into the YUV format. The display data generation means generates display data based on the exponent data signal separated by the signal separation means and the mantissa data format signal obtained by conversion by the format conversion means. It is possible.

  Accordingly, it is possible to obtain an effect that the video signal of the image data in the floating point format can be output to the video processing device without changing the existing YUV interface, and from the video signal output from the video output device, a predetermined value is obtained. Since display data that can be displayed in the display format can be generated, for example, a display system that displays a video of HDR image data in a floating-point format, a monitoring system that uses HDR image data video, a security system, and the like can be simplified. The effect that it can be constructed is obtained.

  Here, the display data that can display a video in a predetermined display format corresponds to, for example, the HDR image data in the floating point format if the output side device supports the image data in the floating point format. If the output device does not support floating-point format image data, RGB image data in which each pixel value is expressed by an integer value, which is obtained by converting floating-point format image data, corresponds. Hereinafter, the same applies to the mode related to the video processing apparatus and the mode related to the video processing method.

[Mode 7] Furthermore, the video processing system of mode 7 is the same as the video processing system of mode 6,
The mantissa conversion means includes a dividing unit that divides the second image data into a plurality of data units in a pixel area composed of j pixels (j is a natural number of 2 or more) adjacent to each other, and each of the pixels The mantissa data in the pixel data corresponding to j pixels in a region is converted into one luminance signal Y for each pixel and k pixels (k is a natural number where k <j) for each pixel region. A conversion unit for converting into a YUV video signal composed of a corresponding color difference signal UV,
The video processing apparatus corresponds to (j−k) pixels thinned out for each pixel area when the mantissa data is converted into the YUV video signal by the mantissa conversion means. A color difference signal interpolation means for interpolating the color difference signal UV is provided.

With such a configuration, the second image data can be converted into video signals corresponding to various formats such as YUV411, YUV211 and YUV422.
Furthermore, since the color difference signal UV of the pixels thinned out when the second image data is converted into a YUV video signal can be interpolated, display data with higher image quality can be generated. can get.

[Mode 8] In order to achieve the above object, the video processing apparatus according to mode 8
Data of the mantissa part in each pixel data of image data composed of pixel data representing pixel values in a floating point format is converted into a luminance signal Y and a color difference signal UV, and each pixel data is converted into each color difference signal UV. A video processing device for generating display data capable of displaying a video in a predetermined display format based on a video signal obtained by multiplexing the exponent data of
A signal separating means for separating the color difference signal UV in which the exponent data is multiplexed into the mantissa color difference signal UV and the exponent data signal;
Format conversion means for converting the luminance signal Y and the color difference signal UV obtained by the separation into a signal of a data format of a mantissa part before being converted into the YUV system;
Display data generating means for generating display data based on the data signal of the exponent part separated by the signal separating means and the signal of the data format of the mantissa part obtained by conversion by the format converting means. Features.

  Accordingly, display data that can be displayed in a predetermined display format can be generated from the YUV video signal in which the exponent data is multiplexed, as in the video processing apparatus in the video processing system of aspect 6. Therefore, for example, it is possible to easily construct a display system or the like that displays a video of the HDR image data in the floating point format.

[Mode 9] In order to achieve the above object, the video output method according to mode 9 includes:
A video output method for converting image data into a YUV video signal including a luminance signal Y and a color difference signal UV, and outputting the video signal obtained by the conversion via a predetermined interface,
The pixel value is expressed in a floating-point format with a mantissa part of N bits (N is a natural number of 2 or more), a radix of X (X is a natural number of 2 or more), and an exponent part of M bits (M is a natural number of 2 or more). A mantissa conversion step of converting mantissa data in each pixel data of the first image data into the YUV video signal for the first image data composed of the pixel data;
A multiplexing step of multiplexing the data of the exponent part of each pixel data constituting the first image data into the color difference signal UV in the YUV video signal;
A video signal output step of outputting a YUV video signal composed of the luminance signal Y and the color difference signal UV in which the exponent data is multiplexed.
Thereby, the same operation and effect as those of the video output device of mode 1 can be obtained.

[Mode 10] In order to achieve the above object, a video processing method according to mode 10 includes:
Data of the mantissa part in each pixel data of image data composed of pixel data representing pixel values in floating point format is converted into a luminance signal Y and a color difference signal UV, and each color difference signal UV is converted to each pixel data. A video processing method for generating display data capable of displaying video in a predetermined display format based on a video signal obtained by multiplexing data of exponent parts,
A signal separation step of separating the color difference signal UV obtained by multiplexing the exponent data into the mantissa color difference signal UV and the exponent data signal;
A format conversion step of converting the luminance signal Y and the color difference signal UV obtained by the separation into a signal of a data format of a mantissa part before being converted into the YUV method;
A display data generation step of generating display data based on the data signal of the exponent part separated in the signal separation step and the signal in the data format of the mantissa part obtained by the conversion in the format conversion step. Features.
Thereby, the same operation and effect as those of the video processing apparatus according to mode 8 can be obtained.

[Mode 11] On the other hand, in order to achieve the above object, the video output program of mode 11
A video output program that converts image data into a YUV video signal including a luminance signal Y and a color difference signal UV, and outputs the video signal obtained by the conversion via a predetermined interface,
The pixel value is expressed in a floating-point format with a mantissa part of N bits (N is a natural number of 2 or more), a radix of X (X is a natural number of 2 or more), and an exponent part of M bits (M is a natural number of 2 or more). A mantissa conversion step of converting the mantissa data of each pixel data constituting the first image data into the YUV video signal for the first image data composed of the pixel data;
A multiplexing step of multiplexing the data of the exponent part of each pixel data constituting the first image data into the color difference signal UV in the YUV video signal;
And a program for causing a computer to execute a process including a video signal output step of outputting a video signal of a YUV system including the luminance signal Y and the color difference signal UV multiplexed with the exponent data. And
With such a configuration, when the program is read by the computer and the computer executes processing in accordance with the read program, the same operation and effect as those of the video output device of mode 1 can be obtained.

[Mode 12] On the other hand, in order to achieve the above object, the video processing program of mode 12
Data of the mantissa part in each pixel data of image data composed of pixel data representing pixel values in floating point format is converted into a luminance signal Y and a color difference signal UV, and each color difference signal UV is converted to each pixel data. A video processing program for generating display data capable of displaying video in a predetermined display format based on a video signal obtained by multiplexing exponential data,
A signal separation step of separating the color difference signal UV obtained by multiplexing the exponent data into the mantissa color difference signal UV and the exponent data signal;
A format conversion step of converting the luminance signal Y and the color difference signal UV obtained by the separation into a signal of a data format of a mantissa part before being converted into the YUV method;
A process comprising: a display data generation step for generating display data based on the exponent data signal separated in the signal separation step and the mantissa data format signal obtained by conversion in the format conversion step Including a program to be executed.

  With such a configuration, when the program is read by the computer and the computer executes processing in accordance with the read program, the same operations and effects as those of the video processing apparatus according to mode 8 are obtained.

  Embodiments of a video output device, a video output method, a video processing system, a video processing device, and a video processing method according to the present invention will be described below with reference to the drawings. 1 to 12 are diagrams showing embodiments of a video output device, a video output method, a video processing system, a video processing device, and a video processing method according to the present invention. Here, FIG. 1 is a block diagram showing a schematic configuration of the video processing system 1.

The schematic configuration of the video processing system 1 will be described below with reference to FIG.
As shown in FIG. 1, the video processing system 1 converts the HDR video signal into a YUV video signal based on a floating-point format HDR video signal and a synchronization signal from the HDR video signal source 10, and generates a video processing device. 30 based on a UVE _p multiplexed signal and a synchronizing signal in which exponent data is multiplexed with a luminance signal Y (hereinafter referred to as a Y signal) from the video output device 20 and a color difference signal UV. The video output device 20 and the video processing device 30 are connected via an interface (port) for inputting and outputting various signals.

  Here, the HDR video signal source 10 includes, for example, an imaging device that generates and outputs an HDR video signal from imaging data captured by an electronic camera or the like, a storage medium in which HDR video data is stored, and the like. In this embodiment, the HDR video signal source 10 uses a value representing each color element for each color element in the RGB (Red Green Blue) color space, the mantissa part is N bits (N is a natural number of 8 or more), and the radix. Includes information on Radiance format RGBE image data (hereinafter simply referred to as HDR image data) composed of pixel data expressed in floating-point format with 2 and exponent part M bits (M is a natural number of 8 or more). A signal (hereinafter referred to as an HDR video signal) is output.

Specifically, in the HDR image data, each pixel data includes an N-bit mantissa part (R _i [(N−1): 0], G _i [(N−1)) of each color element of R, G, B. : 0], B _i [(N−1): 0]) and data of the exponent (E [(M−1): 0]) common to all the color elements. It becomes the HDR image data in decimal format.
Further, a detailed configuration of the video output device 20 will be described with reference to FIG. Here, FIG. 2 is a block diagram showing a detailed configuration of the video output device 20.

As shown in FIG. 2, the video output device 20 includes an exponent part converter 200, a YUV411 converter 201, a packing index part data generator 202, a UVE _p multiplexer 203, a delay part 204, a pixel counter The controller 205 includes a Y signal port 206, a UV signal port 207, and a synchronization signal port 208.
Based on the data indicated by the HDR video signal in the floating-point format from the HDR video signal source 10, the exponent conversion unit 200 converts the number of bits of the exponent part of each HDR image data constituting the data into the output format of the YUV411 system. Convert based on.

Specifically, when the bit number of the exponent part of each pixel data in the HDR image data is 5 bits or more, the numerical value of the mantissa part is changed so that the exponent part can be expressed by 4 bits or less, and the exponent part Convert the number of bits to 4 bits.
In the HDR image data after conversion (hereinafter referred to as exponent part conversion format HDR image data), each pixel data has an N-bit mantissa part (R ′ _i [(N−1): 0] of each color element. , G ′ _i [(N−1): 0], B ′ _i [(N−1): 0]) and the exponent (E ′ [(M−1): 0) common to all colors. ]) Data.

The exponent conversion unit 200 outputs the mantissa data in each pixel data of the exponent conversion format HDR image data to the YUV411 conversion unit 201, while the exponent data is supplied to the packing exponent data generation unit 202. Output.
When the converted mantissa data is input from the exponent conversion unit 200, the YUV411 conversion unit 201 converts the converted data into a video signal in the YUV411 output format.

  Here, as described above, the YUV411 output format outputs all pixels of the Y signal, which is a luminance signal, and the U signal and V signal, which are color difference signals, have a low spatial frequency relative to the color of the human eye. In consideration of this, the amount of information to be output is reduced. Specifically, for the U signal and the V signal, one representative value is obtained from four pixels, and only this representative value is output. For example, if the mantissa part is 8-bit image data, 1 byte of U signal and 1 byte of V signal are output for 4 bytes of Y signal.

Therefore, the YUV411 conversion unit 201 uses the mantissa data (R ′ _i [(N−1): 0], G ′ _i [(N−1): 0]) in each pixel data of the exponent part conversion format HDR image data. , B ′ _i [(N−1): 0]) is converted into YUV411 data (Y [(N−1): 0], U [(N−1): 0], V [(N−1)). : 0]).
Then, the YUV411 conversion unit 201 outputs Y [(N−1): 0] in the converted mantissa data to the delay unit 204, while U [(N−1): 0] and V [(N−1): 0] is output to the UVE _p multiplexer 203.

When the exponent data after conversion (E ′ [m (m is a natural number of 4 or less): 0]) is input from the exponent conversion unit 200, the packing exponent data generation unit 202 in the horizontal scanning direction. Packing exponent data (E _p = {E _{n (n = 1, 2, 3,...)} , En _{+ 1} , E _{n + 2} , E _{n + 3} }). Specifically, since the exponent part conversion unit 200 converts the number of bits of the exponent part of each pixel to 4 bits or less, the packing exponent part data has a maximum of 4 bits × 4 = 16 bits (2 bytes). It becomes data.

Then, the YUV411 conversion unit 201 outputs the generated packing index part data to the UVE _p multiplexing unit 203.
The UVE _p multiplexing unit 203 receives the U signal and the V signal from the YUV 411 conversion unit 201 and the packing exponent data generation unit 202 from the packing exponent data generation unit 202 (hereinafter referred to as a packing exponent data signal). Then, based on the multiplex control signal input from the pixel counter / control unit 205, the UV signal and the packing index part data signal are multiplexed and output to the UV signal port 207.

Specifically, the U signal and the V signal are reduced to ¼ information amount as described above, so that the dummy data generated at the output timing of either the odd line or the even line can be substituted. The packing index part data signal is output. Therefore, when the U and V signals at the output timing of the odd lines are outputted, the packing exponential part data signals at the output timing of the even lines (hereinafter, referred to as E _p signal) is to be output.

Delay unit 204, a Y signal, and the UV signal is a buffer for synchronizing the E _p signal. Specifically, the delay amount generated by the packing index part data generation unit 202 and the UVE _p multiplexing unit 203 is corrected by delaying the Y signal.
The pixel counter / control unit 205 counts the line number and pixel number of the HDR image data based on the synchronization signal IN from the HDR video signal source 10. Based on the count result, a control signal for controlling the exponent part conversion unit 200 and the packing exponent part data generation unit 202 is generated, and a multiplex control signal for controlling the UVE multiplexing unit 203 is generated. Then, these generated control signals are input to each component. On the other hand, the Y signal after a delay, in order to synchronize the UV signal, E _p signal, generates a sync signal OUT obtained by delaying the synchronizing signal IN, and outputs it to the synchronization signal port 208.

The Y signal port 206 is a port having a data output width (bus width) of 8 bits, and outputs the Y signal input via the delay unit 204 to the video processing device 30.
The UV signal port 207 is a port having a data output width (bus width) of 8 bits, and is a multiplexed UV signal and E _p signal (hereinafter referred to as UVE _p multiplexed signal) input from the UVE _p multiplexing unit 203. Is output to the video processing device 30.
The synchronization signal port 208 outputs the synchronization signal OUT input from the pixel counter / control unit 205 to the video processing device 30.

Further, a detailed configuration of the video processing device 30 will be described with reference to FIG. Here, FIG. 3 is a block diagram showing a detailed configuration of the video processing device 30.
The video processing device 30 includes a Y signal port 300, a UV signal port 301, a synchronization signal port 302, a delay amount adjustment unit 303, a UVE _p multiple signal separation unit 304, an HDR video conversion unit 305, and an exponent part separation. 306, an exponent expansion unit 307, a pixel counter / control unit 308, and a VRAM (Video Random Access Memory) 309.

The Y signal port 300 is a port having an 8-bit data output width (bus width), and outputs the Y signal output from the video output device 20 and input to the delay amount adjustment unit 303.
The UV signal port 301 is a port having a data output width (bus width) of 8 bits, and outputs the UVE _p multiplexed signal output and input from the video output device 20 to the UVE _p multiplexed signal separation unit 304.

The synchronization signal port 302 outputs the synchronization signal OUT output and input from the video output device 20 to the pixel counter / control unit 308.
The delay amount adjusting unit 303 is a buffer for synchronizing the Y signal and the UVE _p multiplexed signal. Specifically, the delay amount generated in the UVE _p multiple signal demultiplexing unit 304 is corrected by delaying the Y signal.

The UVE _p multiplexed signal separation unit 304 separates the UVE _p multiplexed signal into a UV signal and an E _p signal, and synchronizes the UV signal, the E _p signal, and the synchronization signal OUT.
In addition, the UVE _p multiple signal separation unit 304 outputs the separated UV signal to the HDR video conversion unit 305 and outputs the separated E _p signal to the exponent part separation unit 306.
Based on the Y signal input from the delay amount adjustment unit 303 and the UV signal input from the UVE _p multiplex signal separation unit 304, the HDR video conversion unit 305 first performs interpolation processing on the UV signal and lacks it. The UV signal of the pixel portion thus generated is generated, and the U signal and the V signal are separated. Next, based on the Y signal and the interpolated U signal and V signal, the Y signal, U signal, and V signal are converted (returned) to an RGB signal.

The HDR video conversion unit 305 then converts the converted RGB data (R ′ _i [(N−1): 0], G ′ _i [(N−1): 0], B ′ _i [(N−1)). : 0]) is output to the VRAM 309.
The exponent part separation unit 306 stores packing exponent part data (E _p [4: 0]) having a configuration in which four pixels are combined into one based on the E _p signal input from the UVE _p multiple signal separation unit 304. The data is divided into four parts, and the divided exponent part data is output to the exponent part extension unit 307.

The exponent part extension unit 307 extends the 4-bit exponent part data to 8 bits (inserts 0 in the upper 4 bits), and outputs the expanded 8-bit exponent part data to the VRAM 309.
The pixel counter / control unit 308 counts the synchronization signal OUT input from the synchronization signal port 302, generates an address control signal that gives a pixel address (XY coordinates), and outputs this to the VRAM 309.

  Based on the address control signal from the pixel counter / control unit 308, the VRAM 309 stores the RGB data input from the HDR video conversion unit 305 and the exponent data input from the exponent expansion unit 307 in association with each other. .

Next, the operation of the video processing system 1 having the above configuration will be described with reference to FIGS.
First, the operation of the video output device 20 will be described with reference to FIGS.
Here, FIG. 4 is a diagram illustrating an example of the data configuration of the exponent part conversion format HDR image data after being converted by the exponent part converter. FIG. 5 is a diagram showing the output order of the Y signal. FIG. 6 is a diagram showing a U signal thinning method and an output method. FIG. 7 is a diagram illustrating a method of multiplexing and outputting the U signal and the V signal. FIG. 8 is a diagram illustrating an example of an internal configuration of the UVE _p multiplexing unit 203. Further, FIG. 9 (a) is a diagram showing the configuration of E _p signals, (b) are diagrams illustrating a method of outputting by multiplexing the UV signal and the E _p signal.

The HDR video signal source 10 outputs an HDR video signal in the Radiance format composed of pixel data having a mantissa part of 8 bits, a radix of 2 and an exponent part of 8 bits for each color of RGB together with a synchronization signal IN. 20 is output.
The HDR video signal output from the HDR video signal source 10 is input to the exponent conversion unit 200 of the video output device 20, while the synchronization signal IN is input to the pixel counter / control unit 205 of the video output device 20. .

When the synchronization signal IN is input to the pixel counter / control unit 205, the video output device 20 counts the number of pulses of the synchronization signal IN and controls the operations of the exponent part conversion unit 200 and the packing index part data generation unit 202. And a control signal for controlling the operation of the UVE _p multiplexing unit 203 are generated and output to the respective units.
The exponent conversion unit 200 converts an 8-bit exponent in the input HDR video signal into a 4-bit exponent based on a control signal input from the pixel counter / control unit 205, as shown in FIG. In addition, exponent part conversion format HDR image data (R ′ _i [7] having pixel data in which values representing RGB color elements are represented by an 8-bit mantissa part and a 4-bit exponent part common to these color elements. : 0], G ′ _i [7: 0], B ′ _i [7: 0], E ′ [3: 0]).

Then, the exponent part conversion unit 200 uses the mantissa data (R ′ _i [7: 0], G ′ _i [7: 0], B ′ _i [7: 0]) is output to the YUV411 conversion unit 201, while the exponent data (E ′ [3: 0]) is output to the packing exponent data generation unit 202.
When the mantissa data is input, the YUV 411 conversion unit 201 converts the data into the YUV 411 format and outputs it.

Hereinafter, description will be given by taking HDR image data having 640 pixels in the horizontal scanning direction (hereinafter referred to as horizontal direction) and 480 pixels in the vertical scanning direction (hereinafter referred to as vertical direction) as an example.
The YUV411 conversion unit 201 first converts the pixel data of R ′ _i , G ′ _i , and B ′ _i into luminance information (Y), the difference between the luminance and the red component (U), and the difference between the luminance and the blue component (V). Convert to pixel data. Since this conversion method is well-known, description is abbreviate | omitted.

  Next, regarding the pixel data (Y [7: 0]) of the Y signal in the pixel data converted into YUV, as shown in FIG. 5, the data of the pixel Y1 at the left end of the first line in the figure is sequentially applied. Output. Then, after outputting the data of the pixel Y640 at the right end of the line in the drawing, the data of the pixel Y1 at the left end of the second line is output. As described above, the pixel data of the Y signal is output to the delay unit 204 in order from the pixel data sequentially input from the exponent part conversion unit 200 in order.

Further, as shown in FIG. 6, the YUV 411 conversion unit 201, as shown in FIG. 6, for each of 4 pixel data of 2 pixels in the horizontal direction × 2 pixels in the vertical direction, the data of these 4 pixels (for example, u11, u12 , U21, u22), one representative value (for example, U1) is calculated. Then, only the obtained representative value is output to the UVE _p multiplexing unit 203. Similarly, for the V signal, only the representative value calculated from the four pixels is output to the UVE _p multiplexing unit 203. Therefore, the U signal and the V signal output from the YUV411 conversion unit 201 to the UVE _p multiplexing unit 203 are ¼ of the total data amount of the Y signal.

On the other hand, when the exponent data (E ′ [3: 0]) in the exponent conversion format HDR image data is input, the packing exponent data generation unit 202 receives a control signal input from the pixel counter / control unit 205. Based on the above, the packing index portion data (E _p = {E ′ _n , E ′ _{n + 1} , E ′) composed of E ′ [3: 0] for the four pixels for every four consecutive pixels in the horizontal direction. _{n + 2} , E ′ _{n + 3} }). Here, n is the pixel number of each line of the image. When the number of pixels in the horizontal direction is 640, n is a value in the range of “1-637”. Therefore, the packing index portion data is “E _p1 = {E ′ ₁ , E ′ ₂ , E ′ ₃ , E ′ ₄ }, E _p2 = {E ′ ₅ , E ′ ₆ , E ′ ₇ , E ′ ₈ },..., E _p159 = {E ′ ₆₃₃ , E ′ ₆₃₄ , E ′ ₆₃₅ , E ′ ₆₃₆ }, E _p160 = {E ′ ₆₃₇ , E ′ ₆₃₈ , E ′ ₆₃₉ , E ′ ₆₄₀ } "

The generated packing index part data signal (hereinafter referred to as an E _p signal) is sequentially output to the UVE _p multiplexing unit 203.
UVE _p multiplexing unit 203, and the U and V signals, when the E _p signal is input, based on the multiplex control signal input from the pixel counter control unit 205, and the U and V signals, E _p signal Are multiplexed and the multiplexed UVE _p multiplexed signal is output to the UV signal port 207.

  Here, in the YUV411 system, as described above, only one representative value signal is output for four pixels in the output of the U signal and the V signal. It is sufficient to output one pixel each during a period in which four pixels of the Y signal are output. Therefore, conventionally, for example, as shown in FIG. 7, U signal and V signal are alternately output (multiple output) at the odd line output timing, and dummy data is output at the even line output timing. . In some cases, dummy data is output during an odd line output period, and U and V signals are output during an even line output period.

  That is, in the conventional YUV411 system, the U signal and the V signal are multiplexed and output in the output period of the odd lines, so that the UV signal data corresponding to two lines of the Y signal in the output period of one line. Is output. In this way, since the UV signal only needs to be output for one odd-numbered line with respect to two Y-signal lines, conventionally, in order to synchronize the timing with the Y signal, dummy data is output during the even-line output period. It was output.

On the other hand, the video output device 20 according to the present embodiment has an exponent data signal E _{p in} the period for outputting dummy data (in the above example, the output period for the even lines in which dummy data was conventionally output). Output a signal. Thereby, the E _p signal and the UV signal are multiplexed.
Specifically, as shown in FIG. 8, the UVE _p multiplexing unit 203 includes a multiplexer 203a, and the multiplexer 203a outputs a signal based on the multiplexing control signal from the pixel counter / control unit 205. and by switching to one of the UV signal and the E _p signal, multiplexes the E _p signal to the UV signal.

On the other hand, the pixel counter / control unit 205 outputs a multiplex control signal indicating “1” when the pixel count is an odd line, and outputs a multiplex control signal indicating “0” when the pixel count is an even line. It has become.
Thus, the multiplexer 203a, when the E _p signal and UV signal is input, the period during which the multiplexing control signal "1" is inputted, outputs a UV signal, whereas, the input multiplexed control signal "0" During this period, the circuit is configured to output the E _p signal.

As a result, the E _p signal (data) having the data structure shown in FIG. 9A is output to the UV signal port 207 in the even line output period, as shown in FIG. 9B, and the UV signal is output to the odd line. Is output to the UV signal port 207 during the output period.
Note that the subscript numerals in E ′ ₁ E ′ ₂ , E ′ ₃ E ′ _4, etc. in FIG. 9B indicate pixel numbers. For example, if E ′ ₁ , pixel number 1 This is the exponent data of the numbered pixel.

On the other hand, in the delay unit 204, a delay amount generated in the packing exponential part data generating unit 202 and UVE _p multiplexing unit 203 being adapted to delay the Y signal, the Y signal is, UV signal or E _p signal and synchronization Then, it is output to the Y signal port 206.
As described above, the Y signal output to the Y signal port 206 is output from the Y signal port 206 to the video processing device 30. The UVE _p multiplexed signal output to the UV signal port 207 is output from the UV signal port 207 to the video processing device 30.

The output of the Y signal and the UVE _p multiplexed signal in this embodiment is performed at a frame rate of 30 [fps]. Specifically, it is performed based on a synchronization signal (synchronization signal OUT) generated by delaying the synchronization signal IN input from the HDR video signal source 10.
Here, the synchronization signal OUT is a vertical synchronization signal indicating the start of a frame of video data to be output, a horizontal synchronization signal indicating the start of a line, three signals of pixel synchronization signals, or a composite signal of these three signals.

Through the above processing procedure, the HDR video signal in the floating-point format from the HDR video signal source 10 is converted into the YUV format, and the exponent data is multiplexed with the UV signal and processed together with the Y signal. It is output to the device 30.
Next, based on FIGS. 10-12, operation | movement of the video processing apparatus 30 is demonstrated.
Here, FIG. 10 is a diagram illustrating a configuration of a demultiplexer 304 a included in the UVE _p multiplexed signal demultiplexing unit 304. FIG. 11 is a diagram illustrating a configuration of a UV signal separation circuit included in the UVE _p multiplexed signal separation unit 304. FIG. 12 is a diagram showing the flow of UV signal interpolation processing.

In the video processing device 30, the Y signal among the signals output from the video output device 20 is input to the delay amount adjustment unit 303 via the Y signal port 300. The UVE _p multiplexed signal is input to the UVE _p multiplexed signal separation unit 304, and the synchronization signal OUT is input to the pixel counter / control unit 308.
The pixel counter / control unit 308 generates an address control signal that gives a pixel address (XY coordinates) based on the input synchronization signal OUT, and outputs the address control signal to the VRAM 309. Also, based on the input synchronization signal OUT, a separation control signal that switches between “0” and “1” according to the line number is generated, and this is output to the UVE _p multiplexed signal separation unit 304.

The UVE _p multiplex signal separation unit 304 includes a demultiplexer 302a shown in FIG. 10, and the demultiplexer 302a separates the UVE _p multiplex signal into a UV signal and an E _p signal. Specifically, the UV signal and the E _p signals multiplexed, and the UV signal by being inputted from the pixel counter control unit 308, and outputs separately on the basis of the isolation control signal switched every aforementioned the line number Separated into E _p signal.

Further, the UVE _p multiplex signal demultiplexing unit 304 includes a UV signal demultiplexing circuit 302b shown in FIG. 11. In the UV signal demultiplexed by the demultiplexer 302a in the UV signal demultiplexing circuit 302b, four UV pixels are separated. The data for the remaining three pixels lost by selecting one representative value is interpolated, and the multiplexed UV signal is separated into a U signal and a V signal, and the U signal and V signal after the interpolation are separated. Is output to the HDR video conversion unit 305.

  As shown in FIG. 11, the UV signal separation circuit 302 b includes a line memory 1300 that accumulates an input UV signal for one line, a multiplexer 1301, a delay element 1302, and a multiplexer 1303. The UV signal is input to the multiplexer 1303 via the line memory 1300 or directly. When a UV signal corresponding to an odd-numbered line is output, the UV signal separation circuit 302b outputs this signal and stores it in the line memory 1300. On the other hand, when the UV signal corresponding to the even-numbered line is output, the odd-numbered signal stored in the line memory 1300 is output. As a result, the multiplexer 1301 always outputs UV signals corresponding to odd-numbered lines.

  In the UV signal separation circuit 302b, when odd-numbered (pixel unit) pixel data is input, the pixel data is output to the HDR video conversion unit 305 and delayed by the delay element 1302. When even-numbered (pixel unit) pixel data is input, the pixel data delayed by the delay element 1302 is output to the HDR video conversion unit 305. As a result, the U signal is separated from the V signal and output to the HDR video conversion unit 305. The pixel data output in this way has a pixel array as shown in FIG. In FIG. 12, the pixel data delayed and output by the line memory 1300 and the delay element 1302 are shown in gray as interpolation pixel data. Through the operation described above, the V signal is also separated from the U signal and output to the HDR video conversion unit 305.

Also, E _p signal separated with UV signal at the demultiplexer 302a is output to the exponent part separating section 306.
On the other hand, the delay amount adjustment unit 303 delays the Y signal input via the Y signal port 300 by the processing time in the UVE _p multiple signal demultiplexing unit 304, and synchronizes the output timing of the Y signal with the UV signal.

The HDR video conversion unit 305 includes a U signal and a V signal input from the UVE _p multiplex signal demultiplexing unit 304, a Y signal that is delayed in the delay amount adjustment unit 303, and is input in synchronization with the U signal and the V signal. Is converted into an RGB HDR video signal (mantissa part). The converted HDR video signal (R ′ _i [7: 0], G ′ _i [7: 0], B ′ _i [7: 0]) is output to the VRAM 309.

When the E _p signal is input from the UVE _p multiplex signal demultiplexing unit 304, the exponent demultiplexing unit 306 receives four pixels of exponent data (E ′ _n , E ′ _{n + 1} , E) included in the E _p signal. ' _{n + 2} , E' _{n + 3} ) is divided into data for each pixel. Then, the exponent part data (E ′ [3: 0]) of each divided pixel is output to the exponent part expansion unit 307.
The exponent expansion unit 307 expands the 4-bit exponent data of each pixel separated by the exponent separation unit 306 into 8-bit exponent data (E ′ [7: 0]). The exponent data is output to the VRAM 309. Specifically, the expansion of the exponent part is performed by setting the input 4-bit exponent part data as the lower 4 bits and inserting “0” as the upper 4 bits thereof.

The VRAM 309 operates on the basis of the address control signal input from the pixel counter / control unit 308, and mantissa data (R ′ _i [7: 0], G) of the HDR video signal input from the HDR video conversion unit 305. ' _i [7: 0], B' _i [7: 0]) and exponent part data (E '[7: 0]) input from the exponent part extension unit 307 are stored in association with each other. Therefore, the value of each color element is stored in the VRAM 309 with the mantissa part (R ′ _i [7: 0], G ′ _i [7: 0], B ′ _i [7: 0]) and the exponent part (E ′ [7 : 0]) HDR image data composed of pixel data expressed in a floating-point format is configured on the memory.

Note that the video processing device 30 may be configured to generate display data that can display the image in another format by processing the floating-point format HDR image data stored in the VRAM 309 as necessary. For example, the HDR image data in the floating point format is converted into the HDR image data in the fixed point format.
As described above, according to the video processing system 1 of the present embodiment, the video output device 20 converts the floating-point format HDR video signal obtained by the HDR video signal source such as an electronic camera into a YUV411 video signal. When the data is output to the video processing device 30, the exponent data can be multiplexed with the UV signal and output using the existing UV signal port as it is. For this reason, since the exponent data can be output without changing the conventional YUV output interface, the system can output the floating-point format HDR video signal to the video processing device 30 without increasing the cost. Can be configured.

  Further, in the video output device 20, packing index part data in which index data for four pixels are collected together can be generated, and this packing index part data can be multiplexed with the UV signal. In the dummy data output period at the time of output, the packing index part data can be output. As a result, the exponent data can be easily multiplexed with the UV signal and output.

Further, in the image processing apparatus 30, the UVE _p multiple signal packing exponential part data are multiplexed in the UV signal, and the UV signal, together can be separated into the E _p signal, the E _p signal to the exponent data of each pixel Can be separated. Further, the Y signal and the UV signal can be converted into RGB signals, and the mantissa data of the HDR image data can be generated (restored) and stored in the VRAM 309. Further, the separated exponent data can be expanded, and the exponent data of the HDR image data can be generated (restored) and stored in the VRAM 309. Thereby, the HDR image data in the floating point format can be formed on the memory of the VRAM 309.

As described above, the video processing apparatus 30 can convert the video signal output from the video output apparatus 20 by the YUV411 system into the HDR image data in the original format and store it in the VRAM 309. Therefore, the system 1 is used. In addition, a video system such as a security system using a surveillance camera or a security system can be easily configured.
In the above embodiment, the exponent part conversion unit 200 corresponds to any one of the exponent part conversion means of the forms 2 to 4, and the YUV411 conversion unit 201 is any one of the forms 1, 2, 3, 6 and 7. The UVE _p multiplexing unit 203 corresponds to the multiplexing unit of any one of the first, second, fifth and sixth or the multiplexing step of the ninth or eleventh aspect. Correspondingly, the packing index part data generation unit 202 corresponds to the packing index part data generation means of the form 2 or 3.

Further, in the above embodiment, the process of separating the UV signal and E _p in the UVE _p multiple signal demultiplexing unit 304 corresponds to the signal demultiplexing means of mode 6 or 8 or the signal demultiplexing step of mode 10 or 12, and UVE _The process of interpolating the deficient data of the UV signal in the _p multiplex signal separation unit 304 corresponds to the color difference signal interpolation unit of mode 7, and the HDR video conversion unit 305 is the format conversion unit of mode 6 or 8, or mode 10 or Corresponding to the 12 format conversion steps, the mantissa data converted by the HDR video conversion unit 305 and the exponent data separated and expanded through the exponent separation unit 306 and the exponent expansion unit 307 are stored in the VRAM 309. Then, the process of forming the floating-point format HDR image data on the memory is the display data generation means of form 6 or 8, or form 10 Ku corresponds to the display data generation step 12.

  In the above-described embodiment, the video output device 20 and the video processing device 30 are configured mainly by hardware. However, the present invention is not limited to this, and various processing functions such as conversion processing, multiplexing processing, and separation processing are performed by software. It is also possible to adopt a software-based configuration realized in In this case, the video output device 20 and the video processing device 30 are necessary for executing a processor for executing a program, a ROM storing a program for realizing each of the above-described processes, a program to be executed, and a program. The computer system includes a RAM for storing various data, a bus for transferring data between these components, and the like.

In the above embodiment, the HDR image data output from the HDR video signal source 10 has a mantissa part of 8 bits, a radix 2 and an exponent part of 8 bits. It may be less than 9 bits or 9 bits or more, the radix may be a numerical value other than 2, and the exponent may be 3 bits or less or 5 bits or more.
In the above embodiment, the present invention is applied to the YUV411 system, which is a kind of YUV system. However, the present invention is not limited to this. It is good also as composition to apply.

1 is a block diagram illustrating a schematic configuration of a video processing system 1. FIG. 2 is a block diagram illustrating a detailed configuration of a video output device 20. FIG. 2 is a block diagram illustrating a detailed configuration of a video processing device 30. FIG. It is a figure which shows an example of the data structure of the exponent part conversion format HDR image data after conversion by the exponent part conversion part. It is a figure which shows the output order of Y signal. It is a figure which shows the thinning-out method and output method of U signal. It is a figure which shows the method of multiplexing and outputting a U signal and a V signal. 3 is a diagram illustrating an example of an internal configuration of a UVE _p multiplexing unit 203. FIG. (A) is a diagram showing a structure of E _p signals, (b) are diagrams illustrating a method of outputting by multiplexing the UV signal and the E _p signal. It is a figure which shows the structure of the demultiplexer 304a which the UVE _p multiplex signal separation part 304 has. 3 is a diagram illustrating a configuration of a UV signal separation circuit included in a UVE _p multiple signal separation unit 304. FIG. It is a figure which shows the flow of the interpolation process of UV signal. It is a figure for demonstrating the structure of the video processing system using the conventional YUV system.

Explanation of symbols

1 is a video processing system, 10 is a HDR video signal source, 20 is a video output device, 30 is a video processing device, 200 is an exponent conversion unit, 201 is a YUV411 conversion unit, 202 is a packing exponent data generation unit, and 203 is UVE _p multiplexing sections 203, 206, and 300 are Y signal ports, 207 and 301 are UV signal ports, 304 are UVE _p multiplexed signal separating sections 304 and 305 are HDR video converting sections, 306 is an exponent separating section, and 307 is an exponent expanding section. 309 is VRAM

Claims (12)

A video output device that converts image data into a YUV video signal including a luminance signal Y and a color difference signal UV, and outputs the video signal obtained by the conversion via a predetermined interface,
The pixel value is expressed in a floating-point format with a mantissa part of N bits (N is a natural number of 2 or more), a radix of X (X is a natural number of 2 or more), and an exponent part of M bits (M is a natural number of 2 or more). Mantissa conversion means for converting the mantissa data of each pixel data constituting the first image data into the YUV video signal for the first image data composed of the pixel data;
Multiplexing means for multiplexing the data of the exponent part of each pixel data constituting the first image data into the color difference signal UV in the YUV video signal;
A video output device comprising: a video signal output means for outputting a YUV video signal comprising the luminance signal Y and the color difference signal UV multiplexed with the exponent data. Exponent part conversion means for converting the M-bit exponent part of each pixel data constituting the first image data into an m-bit exponent part (m is a natural number of M>m);
In the second image data, which is image data after the exponent part is converted by the exponent part conversion means, for each pixel data of a predetermined number of pixels continuous in a predetermined scanning direction, an exponent in the pixel data of the predetermined number of pixels Packing index part data generating means for generating packing index part data composed of parts of data,
The mantissa conversion means converts mantissa data of each pixel data constituting the second image data into the YUV video signal,
2. The video output apparatus according to claim 1, wherein the multiplexing unit multiplexes the packing index part data into a color difference signal UV in the YUV video signal. The first image data is color image data in RGB (Red, Green, Blue) format, and a value representing each of R, G, and B color elements constituting the pixel value has a mantissa part of the N bits. The radix is 2 and the exponent is represented in the M-bit floating point format, and the value indicated by the M-bit exponent is composed of pixel data having the same value in each color element,
The exponent part converting means converts the M-bit exponent part corresponding to each color element of each pixel data constituting the first image data into an exponent part of 4 bits or less having the same value in each color element. And
The mantissa conversion means converts mantissa data corresponding to each color element of each pixel data constituting the second image data into a YUV411 video signal that is one type of the YUV method,
In the second image data, the packing index part data generation unit is configured to include, for each pixel data of four pixels continuous in the horizontal scanning direction, packing index part data composed of index part data in the pixel data of the four pixels. The video output device according to claim 2, wherein: The first image data is expressed in a floating-point format in which values representing the R, G, and B color elements constituting the pixel value are 8 bits for the mantissa, 2 for the radix, and 8 bits for the exponent. And the value indicated by the 8-bit exponent part is composed of pixel data having the same value in each color element,
The exponent part conversion means converts the 8-bit exponent part corresponding to each color element of each pixel data constituting the first image data into a 4-bit exponent part having the same value in each color element. The video output apparatus according to claim 3.   The multiplexing means multiplexes the data of the exponent part with the color difference signal UV corresponding to either an odd line or an even line of an image. The video output device according to item. A video processing system comprising a video output device and a video processing device, the video processing device generating display data capable of displaying video in a predetermined display format based on a video signal output from the video output device. And
The video output device
The pixel value is expressed in a floating-point format with a mantissa part of N bits (N is a natural number of 2 or more), a radix of X (X is a natural number of 2 or more), and an exponent part of M bits (M is a natural number of 2 or more). A mantissa for converting mantissa data in each pixel data of the first image data into YUV video signal composed of the luminance signal Y and the color difference signal UV with respect to the first image data composed of the pixel data. Part conversion means;
Multiplexing means for multiplexing the data of the exponent part of each pixel data constituting the first image data into the color difference signal UV in the YUV video signal;
A YUV video signal comprising the luminance signal Y in the YUV video signal obtained by the conversion by the mantissa conversion means and the color difference signal UV in which the exponent data is multiplexed by the multiplexing means. Video signal output means for outputting,
The video signal output means outputs the luminance signal Y and the color difference signal UV to the video processing device independently of each other,
The video processing device includes:
A signal separation means for separating the color difference signal UV multiplexed with the exponent data input from the video output device into the mantissa color difference signal UV and the exponent data signal;
A format converting means for converting the luminance signal Y input from the video output device and the color difference signal UV obtained by the separation into a signal in a data format of a mantissa before being converted into the YUV format;
Display data generating means for generating display data based on the data signal of the exponent part separated by the signal separating means and the signal of the data format of the mantissa part obtained by conversion by the format converting means. Characteristic video processing system. The mantissa conversion means includes a dividing unit that divides the second image data into a plurality of data units in a pixel area composed of j pixels (j is a natural number of 2 or more) adjacent to each other, and each of the pixels The mantissa data in the pixel data corresponding to j pixels in a region is converted into one luminance signal Y for each pixel and k pixels (k is a natural number where k <j) for each pixel region. A conversion unit for converting into a YUV video signal composed of a corresponding color difference signal UV,
The video processing apparatus corresponds to (j−k) pixels thinned out for each pixel area when the mantissa data is converted into the YUV video signal by the mantissa conversion means. The video processing system according to claim 6, further comprising color difference signal interpolation means for interpolating the color difference signal UV. Data of the mantissa part in each pixel data of image data composed of pixel data representing pixel values in a floating point format is converted into a luminance signal Y and a color difference signal UV, and each pixel data is converted into each color difference signal UV. A video processing device for generating display data capable of displaying a video in a predetermined display format based on a video signal obtained by multiplexing the exponent data of
A signal separating means for separating the color difference signal UV in which the exponent data is multiplexed into the mantissa color difference signal UV and the exponent data signal;
Format conversion means for converting the luminance signal Y and the color difference signal UV obtained by the separation into a signal of a data format of a mantissa part before being converted into the YUV system;
Display data generating means for generating display data based on the data signal of the exponent part separated by the signal separating means and the signal of the data format of the mantissa part obtained by conversion by the format converting means. A video processing device. A video output method for converting image data into a YUV video signal including a luminance signal Y and a color difference signal UV, and outputting the video signal obtained by the conversion via a predetermined interface,
The pixel value is expressed in a floating-point format with a mantissa part of N bits (N is a natural number of 2 or more), a radix of X (X is a natural number of 2 or more), and an exponent part of M bits (M is a natural number of 2 or more). A mantissa conversion step of converting mantissa data in each pixel data of the first image data into the YUV video signal for the first image data composed of the pixel data;
A multiplexing step of multiplexing the data of the exponent part of each pixel data constituting the first image data into the color difference signal UV in the YUV video signal;
And a video signal output step of outputting a YUV video signal comprising the luminance signal Y and the color difference signal UV in which the exponent data is multiplexed. Data of the mantissa part in each pixel data of image data composed of pixel data representing pixel values in floating point format is converted into a luminance signal Y and a color difference signal UV, and each color difference signal UV is converted to each pixel data. A video processing method for generating display data capable of displaying video in a predetermined display format based on a video signal obtained by multiplexing data of exponent parts,
A signal separation step of separating the color difference signal UV obtained by multiplexing the exponent data into the mantissa color difference signal UV and the exponent data signal;
A format conversion step of converting the luminance signal Y and the color difference signal UV obtained by the separation into a signal of a data format of a mantissa part before being converted into the YUV method;
A display data generation step of generating display data based on the data signal of the exponent part separated in the signal separation step and the signal in the data format of the mantissa part obtained by the conversion in the format conversion step. A characteristic video processing method. A video output program that converts image data into a YUV video signal including a luminance signal Y and a color difference signal UV, and outputs the video signal obtained by the conversion via a predetermined interface,
The pixel value is expressed in a floating-point format with a mantissa part of N bits (N is a natural number of 2 or more), a radix of X (X is a natural number of 2 or more), and an exponent part of M bits (M is a natural number of 2 or more). A mantissa conversion step of converting the mantissa data of each pixel data constituting the first image data into the YUV video signal for the first image data composed of the pixel data;
A multiplexing step of multiplexing the data of the exponent part of each pixel data constituting the first image data into the color difference signal UV in the YUV video signal;
And a program for causing a computer to execute a process including a video signal output step of outputting a video signal of a YUV system including the luminance signal Y and the color difference signal UV multiplexed with the exponent data. A video output program. Data of the mantissa part in each pixel data of image data composed of pixel data representing pixel values in floating point format is converted into a luminance signal Y and a color difference signal UV, and each color difference signal UV is converted to each pixel data. A video processing program for generating display data capable of displaying video in a predetermined display format based on a video signal obtained by multiplexing exponential data,
A signal separation step of separating the color difference signal UV obtained by multiplexing the exponent data into the mantissa color difference signal UV and the exponent data signal;
A format conversion step of converting the luminance signal Y and the color difference signal UV obtained by the separation into a signal of a data format of a mantissa part before being converted into the YUV method;
A process comprising: a display data generation step for generating display data based on the exponent data signal separated in the signal separation step and the mantissa data format signal obtained by conversion in the format conversion step A video processing program comprising a program for causing the program to be executed.

Images