JP2007060706A - Motion vector detecting apparatus and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To expand search area without increasing transfer rate of search data when motion vector is detected. <P>SOLUTION: When motion vector is detected per a reference block by calculating image data of a reference block and calculating image data of an inspection block in search area, variable setting of the search area is performed, based on motion vectors detected in a plurality of continuous reference blocks. Blocks which are continuous in horizontal direction or vertical direction are preferable for a plurality of the reference blocks. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a motion vector detection device and a motion vector detection method in an image encoding device compliant with, for example, MPEG (Moving Picture Image Coding Experts Group).

  The MPEG system is an encoding system that compresses image data by combining DCT (Discrete Cosine Transform), motion compensation prediction, and variable length encoding.

  FIG. 13 shows the configuration of an image encoding device compliant with the MPEG system. In this figure, image data is supplied to the input terminal T1. This image data is input to the motion vector detection circuit 21 and the subtraction circuit 22. The motion vector detection circuit 21 obtains motion vectors between the current frame and a reference frame (for example, the previous frame) using the input image data, and provides the motion vector to the motion compensation circuit 23.

  The image data of the reference frame is also stored in the frame memory 24. This image data is supplied to the motion compensation circuit 23. The motion compensation circuit 23 performs motion compensation of the image data sent from the frame memory 24 using the motion vector sent from the motion vector detection circuit 21. The output of the motion compensation circuit 23 is sent to the subtraction circuit 22 and the addition circuit 25.

  The subtraction circuit 22 subtracts the current frame image data supplied from the input terminal T1 and the motion compensated reference frame image data supplied from the motion compensation circuit 23 to obtain prediction error data, and the DCT circuit 26 To supply. The DCT circuit 26 subjects this prediction error data to DCT processing and sends it to the quantizer 27. The quantizer 27 quantizes the output of the DCT circuit 26 and sends it to the variable length coding circuit 28. The variable length coding circuit 28 performs variable length coding on the output of the quantizer 27 and outputs it from the output terminal T2.

  The output of the quantizer 27 is also supplied to the inverse quantizer 29. Then, the inverse quantization process is performed here, and the output thereof is subjected to the inverse DCT process in the inverse DCT circuit 30, is returned to the original prediction error data, and is given to the addition circuit 25.

  The adder circuit 25 adds the prediction error data to the output data of the motion compensation circuit 23 to obtain the image data of the current frame. The obtained image data is stored in the frame memory 24 as image data of the next reference frame.

  A block matching method is known as a motion vector detection method in such an image encoding device. In the block matching method, a screen is divided into small rectangular areas (blocks), and motion is detected for each block. The block size includes 8 horizontal pixels × 8 vertical pixels (hereinafter abbreviated as 8 × 8), 16 × 16, and the like. Next, the block matching method will be described with reference to FIG.

  In FIG. 14, an M × N reference block RB is set in the reference frame 41. In addition, a test block SB having the same size as the reference block RB is set in the search frame 42. The inspection block SB is moved around a predetermined search range 43 of ± m × ± n around the same position as the reference block RB. Then, the degree of coincidence between the reference block RB and the inspection block SB is calculated, and the inspection block having the highest degree of coincidence is set as a matching block, and a motion vector is obtained from the matching block.

  That is, when the matching degree of the inspection block SBk at the position shifted (u, v) from the inspection block SB0 at the same position as the reference block RB is the highest, the motion vector of the reference block RB is (u, v). And At this time, the inspection block having the smallest sum of absolute value differences for each pixel at the same position in the reference block RB and the inspection block SB, the sum of squares of the differences for each pixel, and the like is set as the inspection block with the highest degree of coincidence.

  In the MPEG system, one sequence of moving images is divided into GOPs (Group of Pictures) made up of a plurality of frames (pictures) for encoding. The GOP includes an intra-frame encoded image (I picture), an inter-frame encoded image (P picture) predicted from an already encoded temporally previous frame, and two frames that are already encoded temporally. And an inter-frame encoded image (B picture) predicted from the image.

  For example, in FIG. 15, first, motion detection is performed using P3 which is a P picture as a reference frame and I0 which is an I picture as a search frame. Next, motion detection in both directions is performed using B1, which is a B picture, as a reference frame and I0 and P3 as search frames. Next, motion detection in both directions is performed using B2, which is a B picture, as a reference frame and I0 and P3 as search frames.

  As shown in FIG. 16, it is generally desirable that the search range necessary for motion detection increases in proportion to the frame interval between the reference frame and the search frame. Here, a case where the block size is 16 × 16 will be described. For example, if the search range is ± 16 in both the horizontal and vertical directions when separated by one frame, it is desirable that the search range be ± 32 when separated by 2 frames and ± 48 when separated by 3 frames.

  However, when the search range is expanded in proportion to the frame interval in this way, the amount of hardware necessary for motion detection also increases by 4 and 9 times, respectively, when one frame is separated. That is, a very large amount of hardware is required to perform motion detection separated by three frames, such as motion detection using P3 as a reference frame and I0 as a search frame.

  Therefore, telescopic search is known as a method for expanding the search range without increasing the hardware amount. This is to cover a large search range by providing an offset at the center of the search range for each reference block while the search range is always ± 16. Next, telescopic search will be described with reference to FIG.

  As shown in FIG. 17, for example, when a search is performed in search frame 3 that is three frames away from the reference frame, the motion vector in the reference block is first searched within ± 16 in search frame 1 one frame away. Find MV1. Next, in a search frame 2 that is two frames away, a search is performed within a search range of ± 16 with MV1 as the center of the search range to obtain a motion vector MV2. At this time, the search range viewed from the reference block in the reference frame is ± 32. Finally, in search frame 3 that is three frames away, a search is performed in a search range of ± 16 with MV2 as the center of the search range to obtain a motion vector MV3. In this way, the motion vector finally three frames away is obtained for the search range ± 48. In this case, the hardware amount required is only a hardware amount that covers a search range of ± 16.

  FIG. 18 shows a comparison of search data transfer amounts for the case where the search range is fixed and the case of telescopic search.

  When the search range is fixed, as shown in (1) of this figure, since the search range overlaps 32 × 48 in the reference blocks adjacent in the horizontal direction, for example, RB0 and RB1, the search range does not overlap 16 × 48 Only the new data need be transferred.

  On the other hand, in the case of the telescopic search, in the search frames 2 and 3, since the search range is different for each reference block, ± 16 search ranges (48 × 48) are set to 256 (48 × 48) for each reference block. = 16 × 16) It is necessary to transfer with a clock.

  That is, in order to perform a telescopic search within a search range of ± 16, a search data transfer rate of (48 × 48) / (16 × 48) = 3 times is required. This value further increases as the search range increases. For example, in the case of ± 32, the transfer rate required for the telescopic search is (80 × 80) / (80 × 16) = 5 times, and the data The transfer becomes very difficult. If the pixel data is 8 bits and the pixel clock is 13.5 MHz, in this case, a transfer rate of (80 × 80/256) × 13.5 MHz × 1 byte = 337.5 Mbyte / sec is required. This large transfer rate is a major obstacle to realizing hardware.

  Further, in the conventional telescopic search, as shown in FIG. 19, after setting the search range of ± 16 after determining that the position has moved largely diagonally to the upper right, if the correct motion vector is a vector with small motion, A correct motion vector cannot be detected within the search range of ± 16. In this case, since the motion of the image is small, the image quality deteriorates visually.

  As described above, the conventional telescopic search has a drawback that a very large data transfer rate is required to transfer search data. Also, if the correct motion vector is determined to have moved greatly and the correct motion vector is a small motion vector, the correct motion vector cannot be detected, and the image motion is small, resulting in a large visual degradation. had.

  The present invention has been made in view of such problems, and provides a motion vector detection device and a motion vector detection method capable of expanding a search range without increasing a search data transfer rate.

  The present invention relates to a motion vector between a reference block of a reference frame and a search block of a search frame separated by N frames (N> M) based on a motion vector of the reference block of the reference frame and a search block of a search frame separated by M frames. Detecting a motion vector distribution between the reference frame and the search frame separated by M frames based on a motion vector of a reference block of the reference frame and a search block of a search frame separated by M frames And setting means for setting a search range in search frames separated by N frames based on the detected motion vector distribution, and in the search range set by the setting means, the reference block of the reference frame and the N frames Working with search blocks in remote search frames And having a motion detection means for detecting a vector.

  Further, the present invention is based on a motion vector between a reference block of a reference frame and a search block of a search frame that is M frames away from the reference block of the reference frame and a search block of a search frame that is N frames (N> M) apart. A motion vector detection method for detecting a motion vector, comprising: a motion between a reference frame and a search frame separated by M frames based on a motion vector between a reference block of the reference frame and a search block of a search frame separated by M frames A vector distribution is detected, a search range in the search frame separated by N frames is set based on the detected motion vector distribution, and the reference block of the reference frame in the set search range is separated by N frames. Detect motion vector with search block of search frame And wherein the door.

  According to the present invention, the motion vector search range can be expanded without increasing the search data transfer rate.

  In addition, by making the search range variable while always covering a small movement part, if there is an error in determining the mode of the search range, or the movement of some reference blocks in a slice moves with other reference blocks Even in the case where the directions are different, it is possible to cope with a small movement that is easily noticeable in image quality degradation.

  Embodiments of the present invention will be described below with reference to the drawings. [1] First embodiment of motion vector detection circuit, [2] Second embodiment of motion vector detection circuit, [3] Motion A third embodiment of the vector detection circuit, [4] a fourth embodiment of the motion vector detection circuit, and [5] an imaging recording apparatus provided with the motion vector detection circuit according to the present invention will be described in detail in the order. .

[1] First Embodiment of Motion Vector Detection Circuit FIG. 1 is a block diagram showing a first embodiment of a motion vector detection circuit to which the present invention is applied. This motion vector detection circuit includes a reference frame memory 1, a search frame memory 2, a motion detection circuit 3, and a motion vector distribution detection circuit 4.

  The reference frame memory 1 stores an image of the reference frame. The search frame memory 2 stores the image of the search frame memory, and sets the search range of the inspection block according to the search data read control signal c sent from the motion vector distribution detection circuit 4.

  The motion detection circuit 3 obtains the motion vector d of the reference block from the image data a of the reference block transferred from the reference frame memory 1 and the image data b of the inspection block transferred from the search frame memory 2. This motion vector d is obtained by the block matching method as in the conventional example. That is, the motion vector d is obtained from the position of the test block that minimizes the sum of absolute value differences for each pixel at the same position in the reference block and the test block and the sum of squares of the differences for each pixel (hereinafter referred to as residual). The motion detection circuit 3 outputs the obtained motion vector d together with the residual e at that time.

  The motion vector distribution detection circuit 4 generates a search data read control signal c based on the motion vector d and the residual e sent from the motion detection circuit 3 and supplies them to the search frame memory 2.

  FIG. 2 is a diagram for explaining the search range of the motion detection circuit 3 described above. (A) of this figure is a case where the operation of moving the reference block in the horizontal direction from the left end to the right end is advanced from the upper end to the lower end, and (b) is the reference block in the vertical direction of the image from the upper end to the lower end. This is a case where the moving operation is advanced from the left end to the right end. In the following description, a horizontal reference block is referred to as a horizontal slice, and a vertical reference block is referred to as a vertical slice.

  In the present invention, the distribution of motion vectors between the base frame and the reference frame is obtained for each horizontal slice or vertical slice, and the search range in the distant frame is set according to the distribution state. In the following description, a case will be described in which a motion vector distribution is obtained for each horizontal slice shown in FIG.

  In FIG. 2A, when the image size is 720 × 576 and the reference block size is 16 × 16, 576/16 = 36 horizontal slices are formed. In addition, 720/16 = 45 reference blocks per horizontal slice. Here, the motion vector between the reference block and the test block one frame before is detected for each horizontal slice, and the distribution is obtained. Then, the search range of the horizontal slice in the reference frame separated by two or more frames is set according to the distribution state.

  For example, when the number of motion vectors facing upward among 45 motion vectors in a certain horizontal slice exceeds a predetermined threshold, the search range is set upward.

  As described with reference to FIG. 18, in the telescopic search, in the search frames 2 and 3, since the search range is different for each reference block, ± 16 search ranges (48 × 48) for each reference block. Must be transferred with 256 (= 16 × 16) clocks. On the other hand, in this embodiment, since the search range is set for each horizontal slice, the search range is shared between reference blocks adjacent in the horizontal direction in the horizontal slice. The read rate of the search image data is the same as that in the case where the normal search range is shown.

[2] Second Embodiment of Motion Vector Detection Circuit FIG. 3 is a block diagram showing a second embodiment of a motion vector detection circuit to which the present invention is applied. Here, the same number is attached | subjected to the part corresponding to FIG.

  This motion vector detection circuit includes a first motion detection circuit 3-A and a second motion detection circuit 3-B. The motion vector d1 and the residual e1 output from the first motion detection circuit 3-A or the motion vector d2 and the residual e2 output from the second motion detection circuit 3-B are used to determine the motion vector for each slice. A motion vector distribution detection circuit 4 ′ that detects the distribution and generates a search data read control signal c is provided. Further, the final motion vector f is obtained from the motion vector d1 and the residual e1 output from the first motion detection circuit 3-A and the motion vector d2 and the residual e2 output from the second motion detection circuit 3-B. A comparison circuit 5 is provided.

  The first motion detection circuit 3-A and the second motion detection circuit 3-B can perform motion detection in different search ranges. The search range of each motion detection circuit is ± 16 in both the horizontal and vertical directions.

  When the search range is fixed, since there are generally many images that move greatly in the horizontal direction, the horizontal search range is set to be larger than the vertical search range as shown in FIG. In this figure, the range designated 3-A on the left side of the Y axis is the search range of the first motion detection circuit 3-A, and the range designated 3-B on the right side of the Y axis is the second motion detection. This is the search range of the circuit 3-B. In this case, the total search range of the first motion detection circuit 3-A and the second motion detection circuit 3-B is horizontal ± 32 and vertical ± 16, and the reference vector corresponding to the center of the search range is ( X, Y) = (0, 0).

  On the other hand, when the search range is variable in units of horizontal slices, the search range of the first motion detection circuit 3-A is fixed and the search range of the second motion detection circuit 3-B is changed. As a result, as shown in FIG. 4B, the search range can be expanded to horizontal ± 48 and vertical ± 48 (details will be described later).

  The motion detection operation of the motion vector detection circuit shown in FIG. 3 will be described below with reference to FIG. In FIG. 5, before performing motion detection from P3 to I0 separated by 3 frames, motion detection from B1 to I0 separated by 1 frame is performed using the first motion detection circuit 3-A, and the residual e1 The result of the motion vector d1 is stored in the motion vector distribution detection circuit 4 ′. Also, motion detection from B-2 to I0 separated by two frames is performed using the second motion detection circuit 3-B. The search range at this time is horizontal ± 16 and vertical ± 16 for both the first motion detection circuit 3-A and the second motion detection circuit 3-B, and the reference vector corresponding to the center of the search range is (X , Y) = (0, 0).

  Next, the first motion detection circuit 3-A performs motion detection from B2 to I0 separated by 2 frames, and the second motion detection circuit 3-B performs motion from B-1 to I0 separated by 1 frame. Perform detection.

  Then, motion detection from P3 to I0, which is 3 frames away, is obtained for each horizontal slice using the result of motion detection from B1 to I0 previously obtained and stored in the motion vector distribution detection circuit 4 ′. Change the search range.

  In the above description, the motion detection of the image separated by 3 frames is performed by changing the search range for each slice using the result of the motion detection of the image separated by 1 frame. The motion detection of the images separated by N frames or fields (N and M are integers satisfying N> M ≧ 1) is performed by changing the search range for each slice using the motion detection results of the images.

  6 to 8 are diagrams showing the search range modes of the motion vector detection circuit shown in FIG. 3, and FIG. 9 shows the operation when the motion vector detection circuit detects the motion from P3 to I0. It is a flowchart.

  As shown in FIG. 9, first, motion detection between frames is performed in step S1. This corresponds to the motion detection between one frame from B1 to I0 already described with reference to FIG.

  Next, in step S2, a motion vector distribution for motion detection between one frame is created for one slice. Specifically, the horizontal ± 16 and vertical ± 16 (32 × 32) motion vector search range is divided into small blocks of 4 × 4 size, and the number of motion vectors detected for each small block is calculated. Count. At this time, a motion vector having a residual larger than a certain threshold is not counted because it has low reliability.

  Next, in step S3, it is determined whether or not it is a still mode (normal mode). If Count (area 1) is equal to or greater than the threshold Th1, the normal mode is determined. Here, a function called Count (designated area) is a function that returns the total number of motion vectors included in the designated area.

  Area 1 is a ± 4 area in which the central portion of FIG. 6B is shaded. This area corresponds to an area where the movement is small. That is, when the majority of the motion vectors of one slice created in step 1 are motion vectors with small motion, the normal mode is determined.

  If the normal mode is determined in step S3, then the normal mode is set in step S4, and the search range corresponding to the normal mode is set in step S10.

  FIG. 6A shows the search range in the still mode. This search range is the same as the case where the search range is fixed as shown in FIG. The center vectors of the search range are SMV1 = (mvx1, mvy1) = (− 16, 0) and SMV2 = (mvx2, mvy2) = (16, 0), respectively.

  The case where it is determined in step S3 that the mode is the still mode has been described. If it is determined in step S3 that the mode is not the still mode, it is next determined in step S5 whether the mode is the horizontal or vertical movement mode. This determination is performed by examining whether or not the sum of the motion vectors included in each area 2-1 to 2-4 shown in FIGS. 7 (2) to (5) is equal to or greater than a predetermined threshold Th2.

  If the total of motion vectors included in area 2-1 is equal to or greater than Th2, mode 2-1 is selected. If the total of motion vectors included in area 2-2 is equal to or greater than Th2, mode 2-2 is selected. If the sum of the motion vectors included in is greater than or equal to Th2, it is determined as mode 2-3, and if the sum of the motion vectors included in area 2-4 is greater than or equal to Th2, it is determined as mode 2-4.

  If it is determined in step S5 that the mode is horizontal or vertical movement mode, that is, any of modes 2-1 to 2-4, the mode is set in step S6, and a search corresponding to the mode is performed in step S10. Set the range.

  FIG. 7A shows the search range of modes 2-1 to 2-4. In this figure, the ± 16 search range in the center is the search range of the first motion detection circuit 3-A, and the adjacent 2-1 to 2-4 are the first in the modes 2-1 to 2-4. 2 is a search range of the motion detection circuit 3-B. The center vector of each search range is as follows.

Mode 2-1 ... SMV1 = (0,0), SMV2 = (0,16)
Mode 2-2 ... SMV1 = (0,0), SMV2 = (16,0)
Mode 2-3... SMV1 = (0, 0), SMV2 = (0, -16)
Mode 2-4... SMV1 = (0, 0), SMV2 = (-16, 0)
If it is determined in step S5 that the mode is not the horizontal or vertical movement mode, it is next determined in step S7 whether the mode is the oblique movement mode. This determination is performed by checking whether or not the total of motion vectors included in each area 3-1 to 3-4 shown in FIGS. 8 (2) to 8 (5) is equal to or greater than a predetermined threshold Th3.

  If the total of motion vectors included in area 3-1 is equal to or greater than Th3, mode 3-1 is selected. If the total of motion vectors included in area 3-2 is equal to or greater than Th3, mode 3-2 is selected, and area 3-3 is determined. If the sum of the motion vectors included in is greater than or equal to Th3, it is determined as mode 3-3, and if the sum of the motion vectors included in area 3-4 is greater than or equal to Th3, it is determined as mode 3-4.

  If it is determined in step S7 that the mode is an oblique movement mode, that is, any of modes 3-1 to 3-4, then that mode is set in step S8, and a search range corresponding to that mode is set in step S10. Set.

  FIG. 8A shows the search range of modes 3-1 to 3-4. In this figure, the search range of ± 16 in the center is the search range of the first motion detection circuit 3-A, and the outer 3-1 to 3-4 are the first in the modes 3-1 to 3-4. 2 is a search range of the motion detection circuit 3-B. The center vector of each search range is as follows.

Mode 3-1 ... SMV1 = (0,0), SMV2 = (16,16)
Mode 3-2... SMV1 = (0, 0), SMV2 = (16, -16)
Mode 3-3... SMV1 = (0, 0), SMV2 = (-16, -16)
Mode 3-4... SMV1 = (0, 0), SMV2 = (-16, 16)
If it is determined in step S7 that the mode is not the oblique movement mode, that is, if it is not determined in any of the still mode, the horizontal, vertical mode, or the diagonal movement mode, the normal mode is set in step S9. In step S10, a search range corresponding to the normal mode is set.

  After the search range is set in step S10, next, in step S11, motion detection from P3 to I0 separated by three frames is performed for the corresponding one horizontal slice. The results from the first motion detection circuit 3-A and the second motion detection circuit 3-B are input to the comparison circuit 5 in FIG. 3, and a motion vector having a small residual is finally output as the motion vector f. The

  Next, in step S12, it is determined whether or not the processing for one frame has been completed. If not, the process returns to step S2 to repeat the processing. If the processing for one frame is finished, the processing is finished.

  Thus, according to the present embodiment, the search range is made variable while always covering the periphery of (0, 0) (here, ± 16 in both horizontal and vertical directions). Therefore, even if there is an error in determining the motion mode, or when the motion of some reference blocks in the slice is different from the motion direction of other reference blocks, the image quality is likely to be noticeable and the motion is small. It is possible to deal with it.

  In this embodiment, there are three types of motion modes (stationary, horizontal, vertical movement, and diagonal movement), but it may be configured to have more fine modes. In this embodiment, an example in which a motion vector between three frames is obtained based on a vector distribution between one frame has been shown. However, in reality, N frames or fields (N , M is an integer satisfying N> M ≧ 1).

[3] Third Embodiment of Motion Vector Detection Circuit FIG. 10 shows a search range in a third embodiment of a motion vector detection circuit to which the present invention is applied. The configuration of the apparatus is almost the same as in FIG. However, in the present embodiment, the first motion detection circuit 3-A performs a conventional telescopic search within a search range of ± 16 for both horizontal and vertical, and the second motion detection circuit 3-B is always (0, 0). Because the search range is fixed within the search range of ± 16 for both horizontal and vertical centering on the center, the search necessary for the conventional telescopic search and the search range fixed search is used instead of the motion vector distribution detection circuit 4 'in FIG. A circuit for generating a data read control signal is provided.

  With this configuration, the second motion detection circuit 3-B can detect the correct motion vector even if it is determined that the motion has moved greatly and the correct motion vector is a vector with small motion.

[4] Fourth Embodiment of Motion Vector Detection Circuit FIG. 11 is a block diagram showing a fourth embodiment of a motion vector detection circuit to which the present invention is applied. Here, the same number is attached | subjected to the part corresponding to FIG. The configuration of this image encoding apparatus is basically the same as that of the motion vector detection circuit shown in FIG. However, in this apparatus, the motion vector distribution detection circuit 4 ″ does not use the motion vector d and the residual e output from the motion detection circuit 3, but receives a motion signal g from the outside. In this case, the search range in the motion detection circuit 3 is not variable for each slice, but is the same for all slices.

[5] Imaging / Recording Device Equipped with Motion Vector Detection Circuit According to the Present Invention FIG. 12 shows an imaging / recording device equipped with a motion vector detection circuit according to the present invention. In this figure, the encoding circuit 12 is basically configured as shown in FIG. 13, but the internal motion vector detection circuit is configured as shown in FIG. 1, FIG. 3, or FIG. Yes. The encoding circuit 12 is supplied with a video signal and a motion signal g from the camera unit 11. The motion signal g has been described with reference to FIG. When the motion vector detection circuit in the encoding circuit 12 is configured as shown in FIG. 1 or FIG. 3, the motion vector distribution detection circuit includes the motion vector and residual sent from the motion detection circuit, and the camera unit. 11 is used to generate a search data read control signal for the search frame memory.

The video signal encoded by the encoding circuit 12 undergoes a predetermined recording modulation process in the modulation circuit 13 and is recorded in the recording device 14. As the recording device, a hard disk, a magneto-optical disk, a DVD (Digital Video Disk) -RAM (Random Access Memory), or the like is used.
In addition, this invention is not limited to the said embodiment, A various deformation | transformation is possible based on the meaning of this invention. For example, the unit for changing the search range may be the right half and the left half of the horizontal slice. In this case, the reading rate of the search data slightly increases at the portion where the search range at the center of the horizontal slice changes, but it can be absorbed in time in the horizontal blanking period. By reducing the unit for making the search range variable in this way, it is possible to cope with fine movements although the search data read rate increases. The present invention can also be applied when detecting motion between fields.

It is a block diagram which shows 1st Embodiment of the motion vector detection apparatus to which this invention is applied. It is a figure explaining the search range of the motion detection circuit in FIG. It is a block diagram which shows 2nd Embodiment of the motion vector detection circuit to which this invention is applied. It is a figure explaining the search range of the motion detection circuit in FIG. It is a figure which shows the example of the motion detection of the motion vector detection circuit of FIG. It is a figure which shows the still mode of the motion detection circuit in FIG. It is a figure which shows the horizontal and vertical movement modes of the motion detection circuit in FIG. It is a figure which shows the diagonal movement mode of the motion detection circuit in FIG. 4 is a flowchart showing an operation of the motion vector detection circuit of FIG. 3. It is a figure which shows the search range in 3rd Embodiment of the motion vector detection apparatus to which this invention is applied. It is a block diagram which shows 4th Embodiment of the motion vector detection circuit to which this invention is applied. It is a block diagram which shows the structure of the imaging recording device provided with the motion vector detection circuit which concerns on this invention. It is a block diagram which shows the structure of the image coding apparatus based on an MPEG system. It is a figure for demonstrating the block matching method. It is a figure which shows the example of the motion detection in MPEG. It is a figure which shows the relationship between a frame space | interval and a desirable search range. It is a figure which shows the conventional telescopic search. It is a figure which compares the data transfer amount of a normal search and a telescopic search. It is a figure which shows a mode that a correct motion vector cannot be detected in a telescopic search.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Reference | standard frame memory, 2 ... Search frame memory, 3, 3-A, 3-B ... Motion detection circuit, 4, 4 ', 4 "... Motion vector distribution detection circuit

Claims (4)

Motion for detecting a motion vector between a reference block of a reference frame and a search block of a search frame separated by N frames (N> M) based on a motion vector of a reference block of the reference frame and a search block of a search frame separated by M frames In the vector detection device,
Based on a motion vector between a reference block of the reference frame and a search block of a search frame separated by M frames, a motion vector distribution between the reference frame and the search frame separated by M frames is detected, and the detected motion vector distribution is detected. And setting means for setting a search range in search frames separated by N frames,
Motion detection means for detecting a motion vector between a reference block of the reference frame and a search block of the search frame separated by N frames in a search range set by the setting means;
A motion vector detection apparatus comprising:   The motion vector detection apparatus according to claim 1, further comprising two or more independent motion detection means. Motion for detecting a motion vector between a reference block of a reference frame and a search block of a search frame separated by N frames (N> M) based on a motion vector of a reference block of the reference frame and a search block of a search frame separated by M frames A vector detection method,
Detecting a motion vector distribution between the reference frame and the search frame separated by M frames based on the motion vector of the reference block of the reference frame and a search block of the search frame separated by M frames;
Based on the detected motion vector distribution, setting a search range in the search frames separated by N frames,
Detecting a motion vector between a reference block of the reference frame and a search block of a search frame separated by N frames in the set search range;
A motion vector detection method characterized by the above.   4. The motion vector detection method according to claim 3, wherein the motion vector is detected by two or more independent motion detection means.

Images