JP2009115910A - Synchronous converter, synchronous converting method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To suppress image distortion due to overtaking in a simple method. <P>SOLUTION: When the image state is determined to be a specified state (YES in any of S103-S106), an image signal composed of two consecutive frames is generated to thereby perform α blending for interpolating the frames (S107). On the other hand, when the image state is determined not to be a specified state (still state, other states) (NO in S101 or YES in S101, and then NO in all of S103-S106), an image signal of the same frame is continuously read twice from a frame buffer 14 and output to thereby perform the same frame interpolation processing for interpolating the frames (S102). <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a synchronous conversion device that writes an input image signal to a frame buffer in synchronization with an input synchronization frequency, and reads and outputs the image signal from the frame buffer in synchronization with an output synchronization frequency different from the input synchronization frequency. Is.

  When the input synchronization frequency of the image signal to be written to the frame buffer is lower than the output synchronization frequency, the reading speed is faster than the writing speed to the frame buffer. This phenomenon is called “overtaking operation”). Due to this overtaking operation, there is a problem that the upper and lower images of the image to be read are disturbed.

In order to deal with such problems, methods as described in Patent Documents 1 and 2 have been proposed.
That is, Patent Document 1 predicts the risk of an overtaking operation by comparing a write address and a read address, and corrects the read timing according to the prediction result, thereby continuously capturing images of the same frame twice. Describes a method of reading out and interpolating frames.

Patent Document 2 describes a method of searching for a motion vector from a video before frame rate conversion, generating an interpolated image in which the position of the image is moved based on images of previous and subsequent frames, and inserting them between frames. Has been.
JP 2001-67060 A Japanese Patent Application Laid-Open No. 11-112939

  However, although the method described in Patent Document 1 described above is a simple method, in an image with regular movement (such as panning, tilt, zoom, rotation, and object movement), the movement stops for a moment. There is a problem that unnaturalness stands out.

  On the other hand, although the method described in Patent Document 2 described above is effective in avoiding the unnaturalness of a moving image, a motion vector is searched from a pixel unit, and a new interpolation pixel is generated based on the searched motion vector. This is not a simple method because the hardware scale increases. Also, if the motion vector search result is incorrect, there is a problem that noise is generated in the image and the image quality is lowered.

  The present invention has been made in view of these problems, and an object of the present invention is to suppress image disturbance caused by an overtaking operation by a simple method.

  The synchronous conversion device according to claim 1 of the present invention, which has been made to achieve the above object, writes an input image signal to a frame buffer in synchronization with an input synchronization frequency and outputs synchronization different from the input synchronization frequency. The image signal is read out from the frame buffer and output in synchronization with the frequency.

  Specifically, this synchronous conversion device generates a first interpolation means for interpolating a frame by redundantly reading out the image signal of the same frame from the frame buffer and an image signal in which a plurality of consecutive frames are mixed. And a second interpolation means for interpolating the frame.

  Then, the interpolation switching means determines whether or not the image signal read from the frame buffer represents a stationary state. If it is determined that the image signal represents a stationary state, the first interpolation means interpolates the frame. When it is determined that the frame does not represent a stationary state, the second interpolation means interpolates the frame.

  According to such a synchronous conversion device, it is possible to easily suppress image disturbance caused by an overtaking operation by performing image interpolation suitable for the image state. That is, when the output synchronization frequency is higher than the input synchronization frequency, it is necessary to interpolate an output frame (image) in order to suppress image disturbance caused by the overtaking operation. Here, only by the method of interpolating frames by reading out the same frame image signals from the frame buffer, unnaturalness such as the momentary stoppage of motion is noticeable. On the other hand, only the method of interpolating frames by generating an image signal in which a plurality of consecutive frames are mixed can make the motion natural, but the resolution is reduced by mixing (the image becomes blurred). .

  Therefore, in the synchronous conversion device of the present invention, when it is determined that the image signal represents the still state, the frame is interpolated by reading out the same frame image signal from the frame buffer, and the still state is determined. If it is determined that it does not represent, the frame is interpolated by generating an image signal in which a plurality of consecutive frames are mixed.

In this way, by selecting an image interpolation method suitable for the image state, it is possible to make the unnatural movement and the decrease in resolution inconspicuous.
Here, the determination as to whether or not the image signal read from the frame buffer represents a still state is performed, for example, as described in claim 2, the image signal read from the frame buffer and the image in the previous frame. This can be done based on whether the difference from the signal is less than a predetermined threshold.

  In the synchronous conversion device according to claim 3, the interpolation switching means does not represent regular motion even when it is determined that the image signal read from the frame buffer does not represent a stationary state. If it is determined, the first interpolation means interpolates the frame.

  According to such a synchronous conversion device, it is possible to make the unnaturalness of movement and the decrease in resolution inconspicuous. In other words, even if it does not represent a stationary state, if it does not represent a regular movement, the unnaturalness of the movie will not be noticeable, so as with the stationary state, the same frame image signal is duplicated from the frame buffer. Thus, the resolution can be prevented from being lowered.

  Here, the determination as to whether or not it represents regular movement is performed by dividing the image into a plurality of blocks and detecting the movement direction of each block, for example, as described in claim 4. The determination may be made based on the arrangement pattern of the block movement direction. In this way, regular movement can be easily determined.

  Next, the synchronous conversion method according to claim 5 writes the input image signal in the frame buffer in synchronization with the input synchronization frequency, and also outputs the image from the frame buffer in synchronization with the output synchronization frequency different from the input synchronization frequency. This is a method of reading and outputting a signal. In this synchronous conversion method, the step of determining whether or not the image signal read from the frame buffer represents a still state and the same frame from the frame buffer if it is determined to represent the still state. Interpolating the frames by reading out the image signals of the same, and interpolating the frames by generating an image signal in which a plurality of consecutive frames are mixed when it is determined that the frames do not represent a stationary state. .

According to such a synchronous conversion method, an effect similar to the effect described for the synchronous conversion device of claim 1 can be obtained.
Next, the program according to claim 6 writes the input image signal to the frame buffer in synchronization with the input synchronization frequency, and outputs the image signal from the frame buffer in synchronization with the output synchronization frequency different from the input synchronization frequency. This is for causing a computer to function as a synchronous conversion device that reads and outputs. The program interpolates the frame by generating a first interpolation means for interpolating the frame by reading out the image signal of the same frame from the frame buffer and generating an image signal in which a plurality of consecutive frames are mixed. It is determined whether the second interpolation means and the image signal read from the frame buffer represent a stationary state. If it is determined that the image signal represents a stationary state, the first interpolation means interpolates the frame. If it is determined that it does not represent a stationary state, the computer is caused to function as an interpolation switching means for causing the second interpolation means to interpolate the frame.

  According to such a program, it is possible to cause the computer to function as the synchronous conversion device according to the first aspect, thereby obtaining the above-described effects.

Embodiments to which the present invention is applied will be described below with reference to the drawings.
[1. overall structure]
FIG. 1 is a block diagram illustrating a schematic configuration of a frame rate conversion apparatus as a synchronous conversion apparatus according to an embodiment.

  This frame rate conversion device displays a captured image having a predetermined input vertical synchronization frequency on the monitor 25 as a display image having a predetermined output vertical synchronization frequency. In the present embodiment, as an example, a case will be described in which the vertical synchronization frequency (input vertical synchronization frequency) of the captured video (PAL video) is 50 Hz and the vertical synchronization frequency (output vertical synchronization frequency) of the display video is 60 Hz.

  As shown in FIG. 1, the image signal of the captured video input to the frame rate conversion device is converted from an analog signal to a digital signal by an analog / digital converter (A / D) 11 and decoded (decoded) by a decoder circuit 12. After that, the scaler (down) 13 reduces the size in accordance with the monitor 25. Then, it is written in the frame buffer 14 in synchronization with the input vertical synchronization frequency (50 Hz).

  The image signal written in the frame buffer 14 is read from the frame buffer 14 in synchronization with the output vertical synchronization frequency (60 Hz). Then, through the input / output phase detection unit 15, the motion detection unit 16, the motion direction detection unit 17, and the specific image detection unit 18, the frame rate is selected by one conversion unit selected from the two types of frame rate conversion units 21 and 22. Converted. Thereafter, the scaler (up) 23 enlarges the size in accordance with the monitor 25, and the display unit 24 displays it on the monitor 25.

[2. Explanation of processing]
Next, processing performed by the frame rate conversion apparatus according to the present embodiment will be described.
The frame rate conversion apparatus according to the present embodiment performs different frame rate conversion processing depending on the image state (the type of motion of the video) in order to prevent video disturbance due to the overtaking operation. That is, for an image state with regular motion (panning, tilt, zoom, rotation, and object movement), the frame is interpolated by generating an image signal in which two consecutive frames are mixed. On the other hand, for other image states (still state or irregular motion state), the frame is interpolated by reading out and outputting the image signal of the same frame twice consecutively (overlapping) as necessary. .

  Here, whether or not the captured video is in an image state with regular motion is determined based on the motion direction of each block obtained by dividing the entire image into a plurality of blocks. Specifically, as shown in FIG. 2, the entire image is divided into 5 × 5 blocks, and each block is further divided into 5 × 5 sub-blocks.

  Then, as shown in FIG. 3, the difference between the images of two consecutive frames (the target image and the image one frame before) is calculated for each pixel, and the absolute value of the calculated difference is a correlation threshold value (correlation exists). Pixels that are equal to or greater than a threshold value set in advance to determine / no are extracted as moving pixels (moving pixels).

  Subsequently, as shown in FIG. 4, the relationship between the number of pixels with respect to the luminance (0 to 255) of each pixel is obtained for each sub-block using only the pixels determined to be moving pixels in the images of two consecutive frames as samples. Is created (in other words, the peculiar content is extracted). That is, one histogram is created for each sub-block.

  Subsequently, as shown in FIG. 5, the motion direction of the central sub-block (hereinafter also referred to as “motion-direction detection target sub-block”) in each block of the target image is represented by the motion direction of the block including the sub-block. Detect as. Specifically, from the histogram of all sub-blocks (25 sub-blocks in the corresponding block) created based on the image one frame before and the image two frames before, the target image, the image one frame before, The most similar to the histogram of the sub-blocks of motion direction detection created based on the above is detected. Then, the direction from the subblock with the highest degree of similarity toward the subblock that is the target of motion direction detection is the motion direction of the subblock that is the target of motion direction detection (that is, the motion direction of the block that includes the subblock).

  In this way, the movement direction of each block constituting the image is obtained, and regular movement such as panning, tilt, zoom, rotation, and object movement is detected based on the arrangement pattern of the movement direction of each block.

  That is, as shown in FIG. 6, in the panning and tilt images, all the movement directions of the blocks are the same direction. However, when a still image (stationary object) is mixed in a part of the image, the movement direction of the block in that part is in a stationary state (represented by a black circle in the drawing).

  Also, as shown in FIG. 7, in a zoomed (reduced or enlarged) image, the movement direction of the blocks is radial (a direction toward the center or a direction opposite to the center) centering on the stationary block. However, when a still image is mixed in a part of the image, the movement direction of the block in that part is in a stationary state.

  Further, in the rotation image, the movement direction of the block is a constant rotation direction around the stationary block. However, when a still image is mixed in a part of the image, the movement direction of the block in that part is in a stationary state.

On the other hand, in the image of object movement, the movement direction of some blocks is one of the above-described movement directions of panning, tilting, zooming, and rotation, and the other blocks are stationary.
Here, the movement direction of the block is digitized as shown in FIG. That is, the left direction is “0”, the lower left direction is “2”, the lower direction is “4”, the lower right direction is “6”, the right direction is “8”, the upper right direction is “10”, and the upper direction is “12”. ”, The upper left direction is“ 14 ”, and the stationary state is“ 16 ”. Also, the direction between the left direction and the lower left direction is “1”, the direction between the lower left direction and the lower direction is “3”, the direction between the lower direction and the lower right direction is “5”, and the lower right direction. The direction between the right direction and the right direction is “7”, the direction between the right direction and the upper right direction is “9”, the direction between the upper right direction and the upper direction is “11”, the upper direction and the upper left direction The direction between the left and right directions is “13”, and the direction between the upper left direction and the left direction is “15”.

  Furthermore, the names of the blocks in the image are defined as shown in FIG. That is, the names of the blocks in the top row are “B00”, “B01”, “B02”, “B03”, and “B04” in order from the left. In addition, the names of the blocks in the second row from the top are “B10”, “B11”, “B12”, “B13”, and “B14” in order from the left. Similarly, the names of the blocks in the third to fifth rows from the top are also "B20", ..., "B24", "B30", ..., "B34", "B40", ..., "B44".

  And IMG [24: 0] = {B4 [4: 0], B3 [4: 0], B2 [4: 0], B1 [4: 0], B0 [4: 0]} Thus, as exemplified below, a conditional expression for determining a motion such as panning, tilt, zoom, rotation, and object movement can be set.

(1) Conditional expressions for panning and tilt detection (conditional expressions 1 to k)
Conditional expression 1 (the whole image moves to the left)
IMG [24: 0] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,0,0,0}
Conditional expression 2 (the whole image moves downward)
IMG [24: 0] = {4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4}
Conditional expression 3 (the whole image moves in the lower right direction)
IMG [24: 0] = {6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6}
(2) Conditional expression for zoom detection (conditional expression k + 1 to m)
Conditional expression k + 1 (reduction)
IMG [24: 0] = {14, 13, 12, 11, 10, 15, 14, 12, 10, 9, 0, 0, 16, 8, 8, 1, 1, 2, 4, 6, 7, 2, 3,4,5,6}
Conditional expression k + 2 (enlarge)
IMG [24: 0] = {6, 5, 4, 3, 2, 7, 6, 4, 2, 1, 8, 8, 16, 0, 0, 9, 10, 12, 14, 15, 10, 11, 12, 13, 14}
(3) Conditional expression for rotation detection (conditional expressions m + 1 to n)
Conditional expression m + 1 (Rotate left)
IMG [24: 0] = {10, 9, 8, 7, 6, 11, 10, 8, 6, 5, 12, 12, 16, 4, 4, 13, 14, 0, 2, 3, 14, 15, 0, 1, 2}
Conditional expression m + 2 (rightward rotation)
IMG [24: 0] = {2,1,0,15,14,3,2,0,14,13,4,4,16,12,12,5,6,8,10,11,6 7, 8, 9, 10}
(4) Conditional expression for detecting object movement (conditional expression n + 1 to z)
Conditional expression n + 1 (the object moves to the right)
IMG [24: 0] = {16, 16, 16, 16, 16, 16, 8, 8, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16}
Conditional expression n + 2 (object approaches the camera)
IMG [24: 0] = {16, 6, 4, 2, 16, 16, 8, 16, 0, 16, 16, 10, 12, 14, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16}
Conditional expression n + 3 (object rotates)
IMG [24: 0] = {16, 1, 0, 15, 16, 16, 4, 16, 12, 16, 16, 7, 8, 9, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16}
The conditional expressions exemplified here are only a part of the conditional expressions, and there are conditional expressions other than those exemplified.

[3. Specific contents of frame rate conversion processing]
Next, specific contents of the frame rate conversion process will be described. The processing described below can be configured to be executed by a computer incorporated in the frame rate conversion apparatus according to a program stored in advance.

[3-1. Rate conversion processing]
First, the rate conversion process executed for each pixel of the captured video will be described with reference to the flowchart of FIG.

  When the rate conversion process is started, first, in S101, whether or not the motion detection for the pixel to be processed (hereinafter referred to as “target pixel”) is “Yes” (whether the target pixel is a moving pixel). Or not). The specific processing contents of whether or not the motion detection is “present” will be described later (FIG. 11).

  If it is determined in S101 that the motion detection for the target pixel is “Yes”, the process proceeds to S102, and the image signal of the same frame is read out twice and output as necessary. After performing the same frame interpolation process for interpolating the process, the process returns to S101. The specific contents of the same frame interpolation process will be described later (FIG. 12).

On the other hand, if it is determined in S101 that the motion detection for the target pixel is not “Yes” (“No”), the process proceeds to S103.
In S103 to S106, it is determined whether or not the image state is a specific state. If it is determined that the image state is not a specific state, the process proceeds to S102. If it is determined that the image state is a specific state, the process proceeds to S107. . Here, the specific state is a state that is determined as an image having regular movement (specifically, panning, tilt, zoom (enlargement and reduction), rotation, and object movement). The image state is determined by an image type detection process (FIG. 16) described later.

That is, first, in S103, it is determined whether or not the image state is a panning tilt.
If it is determined in S103 that the image state is not panning tilt, the process proceeds to S104, and it is determined whether or not the image state is zoom.

If it is determined in S104 that the image state is not zoom, the process proceeds to S105 to determine whether or not the image state is rotation.
If it is determined in S105 that the image state is not rotation, the process proceeds to S106, and it is determined whether or not the image state is object movement.

  If it is determined in S106 that the image state is not an object movement (that is, if it is determined that the image state is not a specific state), the process proceeds to S102, and the same frame interpolation processing is performed as in the case of the stationary state pixels. After performing, it returns to S101.

  On the other hand, when the image state is determined to be panning tilt in S103, the image state is determined to be zoom in S104, the image state is determined to be rotated in S105, or the image state is determined in S106. If it is determined that is an object movement, the process proceeds to S107. In S107, an α blend process for interpolating frames is performed by generating an image signal obtained by mixing two consecutive frames, and the process returns to S101. The specific content of the α blend process will be described later (FIG. 13).

[3-2. Motion detection processing]
Next, the motion detection process for performing the motion detection determination in S101 of the rate conversion process described above will be described with reference to the flowchart of FIG.

  When this motion detection process is started, first, in S201, a pixel data difference is obtained for the target pixel. Specifically, as described above, the difference between the images of two consecutive frames (the target image and the image one frame before) is calculated for each pixel (FIG. 3).

Subsequently, in S202, it is determined whether or not the absolute value of the difference for the target pixel calculated in S201 is greater than or equal to the correlation threshold value.
If it is determined in S202 that the absolute value of the difference for the target pixel is equal to or greater than the correlation threshold value, the process proceeds to S203, and motion detection for the target pixel is stored as “present”. Thereafter, the process returns to S201.

  On the other hand, if it is determined in S202 that the absolute value of the difference for the target pixel is not equal to or greater than the correlation threshold value, the process proceeds to S204, and the motion detection for the target pixel is stored as “none”. Thereafter, the process returns to S201.

[3-3. Same frame interpolation processing]
Next, the same frame interpolation process executed in S102 of the rate conversion process described above will be described with reference to the flowchart of FIG.

  When the same frame interpolation process is started, first, in S301, it is determined whether or not an overtaking timing (a timing at which an overtaking operation occurs unless an extra frame is interpolated) has arrived.

If it is determined in S301 that it is not the overtaking timing, the process proceeds to S302, the target pixel is output as it is, and the same frame interpolation processing is ended.
On the other hand, if it is determined in S301 that it is the overtaking timing, the process proceeds to S303, and after outputting the pixel one frame before, the same frame interpolation processing is terminated. That is, pixels in the same frame are read out twice in succession.

[3-4. α blend processing]
Next, the α blend process executed in S107 of the rate conversion process described above will be described with reference to the flowchart of FIG.

In this α blend process, an output pixel is obtained by the following equation (S401).
Output pixel = pixel of frame one frame before × α + pixel of target frame × (1−α)
Here, the ratio α is a ratio that takes into account the previous frame and the target frame, and increases as the timing of the output frame is closer to the timing of the previous frame. A specific method for calculating the ratio α will be described later (FIG. 14).

That is, an interpolated pixel is generated by mixing pixels in two consecutive frames at a mixing ratio corresponding to the degree of phase shift between the writing timing and the display timing.
[3-5. Rate detection processing]
Next, the rate detection process for calculating the ratio α used in the α blend process described above will be described with reference to the flowchart of FIG.

When this rate detection process is started, first, in S501, it is determined whether or not the vertical synchronization signal Vsync of the output video has been detected. If it is determined that the detection has been detected, the process proceeds to S502.
In S502, the scanning line on the input side is referred to. That is, the number of lines of the input video in the frame buffer 14 at the time when the vertical sync signal Vsync of the output video is detected is referred to.

  Subsequently, in S503, a value obtained by dividing the number of lines referred to in S502 by the total number of lines of the input video is calculated as a ratio α, and the process returns to S501. A specific method for calculating the total number of lines of the input video will be described later (FIG. 15).

[3-6. Input video total line number detection process]
Next, the input video total line number detection process for calculating the total number of lines of the input video used in S503 of the above-described rate detection process will be described with reference to the flowchart of FIG.

  When the input video total line number detection process is started, first, in S601, it is determined whether or not the horizontal synchronization signal Hsync of the input video has been detected. If it is determined that the input video total line number detection processing has been detected, the process proceeds to S602.

In S602, it is determined whether or not the vertical synchronization signal Vsync of the input video is detected.
If it is determined in S602 that the vertical synchronizing signal Vsync of the input video is not detected, the process proceeds to S603, and 1 is added to the value of the line counter for counting the number of lines. Thereafter, the process returns to S601.

On the other hand, if it is determined in S602 that the vertical synchronizing signal Vsync of the input video has been detected, the process proceeds to S604 and the value of the line counter is stored as the total number of lines.
Subsequently, in S605, the value of the line counter is reset (to 0). Thereafter, the process returns to S601.

[3-7. Image type detection process]
Next, image type detection processing executed for each captured video image (in units of frames) will be described with reference to the flowchart of FIG.

When the image type detection process is started, first, in S701, it is determined whether or not the vertical synchronization signal Vsync of the output video has been detected. If it is determined that the detection has been detected, the process proceeds to S702.
In step S <b> 702, a panning / tilt detection process for detecting a panning or tilt image based on conditional expressions 1 to k is performed. The specific contents of the panning tilt detection process will be described later (FIG. 17).

In step S703, it is determined whether the image state is undetermined.
If it is determined in S703 that the image state is undetermined, the process proceeds to S704, and zoom detection processing for detecting a zoom (enlargement / reduction) image based on conditional expressions k + 1 to m is performed. The specific contents of the zoom detection process will be described later (FIG. 18).

In step S705, it is determined whether the image state is undetermined.
If it is determined in S705 that the image state is undetermined, the process proceeds to S706, and a rotation detection process for detecting a rotation image based on the conditional expressions m + 1 to n is performed. The specific contents of the rotation detection process will be described later (FIG. 19).

In step S707, it is determined whether the image state is undetermined.
If it is determined in S707 that the image state is undetermined, the process proceeds to S708, and an object movement detection process for detecting an object movement image based on the conditional expressions n + 1 to z is performed. The specific contents of the object movement detection process will be described later (FIG. 20).

In step S709, it is determined whether the image state is undetermined.
If it is determined in S709 that the image state is undetermined (not regular movement), the process proceeds to S710, and the image state is stored as another moving screen. Thereafter, the process returns to S701.

On the other hand, if it is determined in S703, S705, S707, or S709 that the image state is not yet determined, the process returns to S701.
[3-8. Panning tilt detection process]
Next, the panning tilt detection process executed in S702 of the above-described image type detection process will be described with reference to the flowchart of FIG.

In this panning tilt detection process, it is determined whether IMG [24: 0] satisfies any of conditional expressions 1 to k for panning and tilt detection (S801 to S803).
If it is determined that IMG [24: 0] corresponds to any of conditional expressions 1 to k, the process proceeds to S804, and the image state is stored as a panning tilt screen.

On the other hand, when it is determined that IMG [24: 0] does not correspond to any of the conditional expressions 1 to k, the panning tilt detection process is terminated while the image state is undetermined.
[3-9. Zoom detection processing]
Next, the zoom detection process executed in S704 of the above-described image type detection process will be described with reference to the flowchart of FIG.

In this zoom detection process, it is determined whether IMG [24: 0] corresponds to any of zoom detection conditional expressions k + 1 to m (S901 to S903).
If it is determined that IMG [24: 0] corresponds to any of conditional expressions k + 1 to m, the process proceeds to S904, and the image state is stored as a zoom screen.

On the other hand, when it is determined that IMG [24: 0] does not correspond to any of conditional expressions k + 1 to m, the zoom detection process is terminated while the image state is undetermined.
[3-10. Rotation detection process]
Next, the rotation detection process executed in S706 of the above-described image type detection process will be described with reference to the flowchart of FIG.

In this rotation detection process, it is determined whether IMG [24: 0] corresponds to any of rotation detection conditional expressions m + 1 to n (S1001 to S1003).
If it is determined that IMG [24: 0] corresponds to any of conditional expressions m + 1 to n, the process proceeds to S1004, and the image state is stored as a rotation screen.

On the other hand, if it is determined that IMG [24: 0] does not correspond to any of conditional expressions m + 1 to n, the rotation detection process is terminated while the image state is undetermined.
[3-11. Object movement detection process]
Next, the object movement detection process executed in S708 of the above-described image type detection process will be described with reference to the flowchart of FIG.

In this object movement detection process, it is determined whether IMG [24: 0] corresponds to any of conditional expressions n + 1 to z for object movement detection (S1101 to S1103).
If it is determined that IMG [24: 0] corresponds to any of conditional expressions n + 1 to z, the process proceeds to S1104, and the image state is stored as an object movement screen.

On the other hand, if it is determined that IMG [24: 0] does not correspond to any of the conditional expressions n + 1 to z, the object movement detection process is terminated while the image state is undetermined.
[4. effect]
As described above, when it is determined that the image state is a specific state (YES in any of S103 to S106), the frame rate conversion apparatus of the present embodiment is an image obtained by mixing two consecutive frames. Α blend processing for interpolating frames by generating signals is performed (S107). On the other hand, if it is determined that the image state is not a specific state (still state or irregular motion state), the frame is interpolated by reading out the image signal of the same frame twice from the frame buffer 14 and outputting it. The same frame interpolation process is performed (S102). In addition, when still images are mixed in a part of an image with regular movement, only the still image portion is subjected to the same frame interpolation processing, and the other portions are subjected to α blend processing. Become.

  According to such a frame rate conversion apparatus, it is possible to easily suppress the video disturbance caused by the overtaking operation by performing the frame interpolation suitable for the image state. In other words, with only the same frame interpolation processing, unnaturalness such as a momentary stoppage of movement is conspicuous. On the other hand, only the α blend process can make the movement natural, but the resolution is lowered by mixing (the image is blurred). On the other hand, if a frame interpolation method suitable for the image state is selected as in the present embodiment, unnatural motion and a decrease in resolution can be made inconspicuous.

  Further, even if it does not represent a stationary state, if it does not represent a regular motion, the same frame interpolation processing is performed as in the stationary state, so that a reduction in resolution can be prevented. .

  Furthermore, since the determination as to whether or not it represents regular movement is made based on the arrangement pattern in the movement direction of each block obtained by dividing the entire image into a plurality of blocks, regular movement is determined. A simple determination can be made.

[5. Correspondence with Claims]
In the frame rate conversion apparatus of the present embodiment, the computer of the frame rate conversion apparatus that executes the processes of S101 and S103 to S106 in the rate conversion process (FIG. 10) corresponds to the interpolation switching unit. The computer of the frame rate conversion apparatus that executes the process of S102 corresponds to the first interpolation means, and the computer of the frame rate conversion apparatus that executes the process of S107 corresponds to the second interpolation means.

[6. Other forms]
As mentioned above, although one Embodiment of this invention was described, it cannot be overemphasized that this invention can take a various form.

  For example, in the frame rate conversion apparatus according to the above embodiment, the same frame interpolation processing is performed as in the stationary state as long as it does not represent a regular motion even if it does not represent the stationary state. However, the present invention is not limited to this, and the α blend process may be performed except in a stationary state.

It is a block diagram which shows schematic structure of the frame rate conversion apparatus of embodiment. It is explanatory drawing of a motion direction detection object subblock. It is explanatory drawing of a motion detection. It is explanatory drawing of specific content extraction. It is explanatory drawing of a motion direction detection. It is explanatory drawing which shows the example of a detection pattern of a panning and a tilt. It is explanatory drawing which shows the example of a detection pattern of zoom, rotation, and an object movement. It is explanatory drawing of digitization of a moving direction. It is explanatory drawing of the name definition of each block. It is a flowchart of a rate conversion process. It is a flowchart of a motion detection process. It is a flowchart of the same frame interpolation process. It is a flowchart of an alpha blend process. It is a flowchart of a rate detection process. It is a flowchart of input video total line number detection processing. It is a flowchart of an image type detection process. It is a flowchart of a panning tilt detection process. It is a flowchart of a zoom detection process. It is a flowchart of a rotation detection process. It is a flowchart of an object movement detection process.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 11 ... Analog-digital converter, 12 ... Decoder circuit, 13, 23 ... Scaler, 14 ... Frame buffer, 15 ... Input-output phase detection part, 16 ... Motion detection part, 17 ... Motion direction detection part, 18 ... Specific image detection part, 21, 22 ... frame rate conversion unit, 24 ... display unit, 25 ... monitor

Claims (6)

A synchronous conversion device that writes an input image signal to a frame buffer in synchronization with an input synchronization frequency, and reads and outputs the image signal from the frame buffer in synchronization with an output synchronization frequency different from the input synchronization frequency. ,
First interpolation means for interpolating a frame by redundantly reading out image signals of the same frame from the frame buffer;
A second interpolation means for interpolating frames by generating an image signal in which a plurality of consecutive frames are mixed;
It is determined whether the image signal read from the frame buffer represents a stationary state. If it is determined that the image signal represents a stationary state, the first interpolation means interpolates the frame, An interpolation switching means for causing the second interpolation means to interpolate a frame when it is determined that it does not represent;
A synchronous conversion device comprising: When the difference between the image signal read from the frame buffer and the image signal in the previous frame is less than a predetermined threshold value, the interpolation switching means determines that it represents a stationary state. The synchronous conversion device according to claim 1, wherein: When the interpolation switching means determines that the image signal read from the frame buffer does not represent a stationary state, but determines that the image signal does not represent regular motion, the first interpolation means The synchronous conversion device according to claim 1, wherein the frame is interpolated. The interpolation switching means divides the image into a plurality of blocks, detects the movement direction of each block, and determines whether or not the movement represents a regular movement based on the arrangement pattern of the movement direction of each block in the image. The synchronous conversion device according to claim 3, wherein: A synchronous conversion method for writing an input image signal to a frame buffer in synchronization with an input synchronization frequency, and reading and outputting the image signal from the frame buffer in synchronization with an output synchronization frequency different from the input synchronization frequency. ,
Determining whether the image signal read from the frame buffer represents a stationary state;
When it is determined that it represents a stationary state, the frame is interpolated by reading out the same frame image signal from the frame buffer, and when it is determined that it does not represent a stationary state, a plurality of consecutive Interpolating the frames by generating an image signal mixed with the frames;
A synchronous conversion method comprising: A computer as a synchronous conversion device that writes an input image signal to a frame buffer in synchronization with an input synchronization frequency, and reads and outputs the image signal from the frame buffer in synchronization with an output synchronization frequency different from the input synchronization frequency. A program for functioning,
First interpolation means for interpolating a frame by redundantly reading out image signals of the same frame from the frame buffer;
A second interpolation means for interpolating frames by generating an image signal in which a plurality of consecutive frames are mixed;
It is determined whether the image signal read from the frame buffer represents a stationary state. If it is determined that the image signal represents a stationary state, the first interpolation means interpolates the frame, A program for causing a computer to function as an interpolating switching means for causing the second interpolating means to interpolate a frame when it is determined that it does not represent.