JPH08140087A - Method and device for detecting movement and recorder with built-in camera - Google Patents

Abstract

PURPOSE: To speed up the movement detection speed of an object, to reduce operation amount and to obtain an excellent image at the time of a decoding by obtaining the motion vector of a device itself by a movement detection sensor when a search area is set in a previous frame image and setting the middle point of a search range in the negative direction of the motion vector information of the device itself between the present frame and the previous frame. CONSTITUTION: A coding means 3 determines a search area based on the lens positional information Sit from a photographing means 2 and the movement information CA wee from a movement information detection means 1 for the video signal Vi from the photographing means 2, performs a compression coding by performing a full search block matching processing within this search area and supplies the compressed video signal ComVi for which this compression. coding is performed to a recording means 4. The recording means 4 records the compressed video signal ComVi from the coding means 3 on a recording medium. Thus, high image quality video can be obtained at the time of a decoding without increasing circuit scale and power consumption by performing the motion detection.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention relates to a camera-integrated V
The present invention relates to a motion detection method and device suitable for application to TR and the like, and a camera-integrated recording device.

[0002]

2. Description of the Related Art Conventionally, various compression techniques have been used for recording and transmitting video signals. Of these various compression techniques, record the video signal of the current frame,
Alternatively, there is a technique of performing a so-called motion detection process of detecting a motion between the constituent elements of the video signal of the previous frame and the constituent elements of the video signal of the current frame when transmitting, and recording or transmitting only the detected motion information. .

FIG. 7 is a block diagram showing an example of a main part of a motion detecting device for compressing information by performing a motion detecting process. Hereinafter, motion detection will be described with reference to FIG.

In FIG. 7, 100 is an input terminal to which a video signal of a current frame to be recorded is supplied, and 102 is an input to which a video signal of a frame immediately before the video signal of a current frame to be recorded is supplied. It is a terminal.

The video signal of the current frame supplied through the input terminal 100 is supplied to the current frame memory 101,
It is stored in the current frame memory 101 by a control signal from the controller 112. In addition, the input terminal 102
The video signal of the previous frame supplied via the reference frame memory 103 is supplied to the reference frame memory 103, and is stored in the reference frame memory 103 by a control signal from the controller 112.

A control signal from the controller 112 sets a block of 8 pixels × 8 pixels (hereinafter referred to as a block of interest) in the video signal of the current frame stored in the current frame memory 101, and in the subsequent processing. The video signal in this block of interest is read out.

On the other hand, in response to a control signal from the controller 112, the address moving circuit 113 sequentially changes the read address of the reference frame memory 103 on a pixel-by-pixel basis, so that the current frame memory 101 in the storage space of the reference frame memory 103. The video signal is sequentially read in a block unit of 8 pixels × 8 pixels from a search area in a predetermined range set corresponding to the target block set in the storage space.

The video signal of the block of interest read from the current frame memory 101 and the reference frame memory 103
The video signals of the blocks in the search area read from are supplied to the adding circuit 104, respectively.

In the adder circuit 104, the video signal of the block in the search area read from the reference frame memory 103 is subtracted for each pixel from the video signal of the target block read from the current frame memory 101. . The subtraction result is supplied to the absolute value circuit 105. The absolute value circuit 105 obtains the absolute value of the addition result from the addition circuit 104. The output of the absolute value circuit 105 is the addition circuit 106.
Is supplied to.

In the adder circuit 106, the output from the absolute value circuit 105 and the output from the delay circuit 107 (the addition result at the time of reading the previous pixel) are added, and the added output is supplied to the memory 108. The memory 108 stores difference absolute value sum data for each block from the start of processing, and the minimum value detection circuit 109 calculates difference absolute value sum data obtained for each block in the search area stored in the memory 108. Of the absolute value of the difference absolute value as the detection result is detected by the motion vector detection circuit 110.
Supply to each.

The motion vector detection circuit 110 calculates a motion vector from the minimum absolute difference value sum data from the minimum value detection circuit 109, and records the calculated motion vector information through an output terminal 111 (not shown). It is supplied to the system and the controller 112, respectively. Controller 1
Reference numeral 12 supplies a write / read control signal to the memory 108 and controls the address moving circuit 113 in the next frame based on the motion vector information from the motion vector detecting circuit 110.

FIG. 8 is an explanatory diagram showing how motion vector information is obtained by full search block matching performed using the apparatus shown in FIG.

In the figure, Sa1 is a search area set in the previous frame image corresponding to the target block in the current frame image, Blk (f) is the target block in the current frame, and Blk (f-1) is Block of previous frame, Vec1 is block of previous frame Blk (f-1)
And motion vector information obtained from the sum of absolute differences between the blocks Blk (f) of the current frame.

That is, the target block B of the current frame image
In the search area Sa1 provided corresponding to lk (f), the read position is shifted by one pixel and the data is read, and the block Blk (f-
The motion vector information calculated from the smallest value among the values obtained as a result of obtaining the sum of absolute differences between the data of 1) and the data of the block of interest Blk (f) and calculating in the search area Sa1 Block Blk (f-
1) and the motion vector between the block of interest Blk (f) of the current frame.

The process described above is generally called full search block matching. By this process, a motion vector is detected, and the motion vector information of the current frame obtained by this detection is recorded or transmitted to record or transmit the motion vector information. The amount of information to be transmitted can be reduced.

Regarding full search block matching, Japanese Patent Application Laid-Open No. 54-124927 and US Pat.
No. 97720.

[0017]

By the way, when the motion vector information obtained by the above-mentioned motion detection is recorded or transmitted, in order to obtain a high quality image at the time of reproduction or reception, the image to be recorded or transmitted is to be recorded. In addition to surely extracting the motion vector information from the signal, it is necessary to enlarge the search area set corresponding to the target block of the current frame.

When the search area is enlarged, a plurality of systems for performing the above-described full search block matching calculation are provided, and each of the divided search areas is divided by the plurality of systems for performing the full search block matching calculation. It is necessary to perform full search block matching calculation for each area.

However, if a large number of systems for performing full search block matching operations are provided for the purpose of improving image quality, the circuit scale becomes large and power consumption increases. By the way, the motion vector detection circuit
It accounts for more than 80% of the main circuit system.

Therefore, there is a disadvantage in that it is impossible to simultaneously achieve high image quality at the time of decoding and low power consumption of the circuit for motion detection.

It is also assumed that motion detection processing is performed on a video signal picked up by a camera-integrated VTR and the motion vector information obtained by this motion detection processing is recorded on a video tape cassette set in the VTR section. In terms of the purpose of carrying and using the camera-incorporated VTR, it is very disadvantageous that the circuit scale becomes large, for example, the case becomes large, and the power consumption becomes large. Since the camera-integrated VTR is driven by a battery, it is extremely disadvantageous, for example, the continuous use time is shortened. However, for this reason, high image quality cannot be avoided. This is because high image quality is always desired by users.

The present invention has been made in consideration of the above points, and it is possible to realize both high image quality at the time of decoding, simplification of a circuit for motion detection, and low power consumption. An attempt is made to propose a possible motion detecting method and apparatus, and a camera-integrated VTR.

[0023]

According to a first aspect of the present invention, a search area is set in a previous frame image corresponding to a target area image in a current frame image, and a reference area is sequentially moved in the search area. At the same time, the image of the reference area is read out, the absolute value of the difference from the attention area image is obtained, the difference is accumulated to obtain the difference accumulated value for each reference area, and the difference accumulated value is determined to be the smallest value. When a search area is set in the previous frame image, the motion vector of the device itself is obtained by the motion detection sensor 1 and the motion detection information between the current frame and the previous frame is obtained. This is a motion detection method in which the midpoint of the search range is set in the negative direction of the motion vector information of the device itself.

A second invention is the same as the first invention,
At the time of search, this is a motion detection method in which the search is sequentially performed from the center of the search area toward the outside.

According to a third aspect of the invention, a search area is set in the previous frame image corresponding to the attention area image in the current frame image obtained by the video camera, and the reference area is sequentially moved within the search area. At the same time, the image of the reference area is read out, the absolute value of the difference from the attention area image is obtained, the difference is accumulated to obtain the difference accumulated value for each reference area, and the difference accumulated value is determined to be the smallest value. In the case of setting a search area in the previous frame image, a motion detection method that sets the search range based on the position information of the lens of the video camera is used. is there.

A fourth invention is the same as the above-mentioned third invention,
At the time of search, this is a motion detection method in which the search is sequentially performed from the center of the search area toward the outside.

According to a fifth aspect of the invention, a search area is set in the previous frame image corresponding to the attention area image in the current frame image obtained by the video camera, and the reference area is sequentially moved in the search area. At the same time, the image of the reference area is read out, the absolute value of the difference from the attention area image is obtained, the difference is accumulated to obtain the difference accumulated value for each reference area, and the difference accumulated value is determined to be the smallest value. When a search area is set in the previous frame image, the motion vector of the device itself is obtained by the motion detection sensor 1 and the motion detection information between the current frame and the previous frame is obtained. This is a motion detection method in which the midpoint of the search range is set in the negative direction of the motion vector information of the device itself, and the search range is set based on the position information of the lens of the video camera.

A sixth invention is the same as the fifth invention, except that
At the time of search, this is a motion detection method in which the search is sequentially performed from the center of the search area toward the outside.

A seventh aspect of the invention is to provide a photographing means 2 and a motion information detecting means 1 which is provided integrally with the photographing means 2 and detects the movement of the photographing means 2 itself, and an attention area image in the current frame image. Correspondingly, a search area is set in the previous frame image, the reference area is sequentially moved within the search area, the image of the reference area is read, and the absolute value of the difference from the attention area image is obtained. Differences are accumulated to obtain a difference accumulated value for each reference area, and the smallest difference accumulated value is used as motion vector information, and the motion information from the motion information detecting means 1 and the photographing means are used. 2 and the encoder 3 for setting the search range based on the lens position information from the motion information detector 1 and the lens position information from the imager 2. A.

An eighth invention is based on the seventh invention,
During the search, the motion detection device sequentially searches from the center of the search area toward the outside.

The ninth aspect of the present invention writes in the current frame memory 31 for storing the video signal of the current frame, the reference frame memory 34 for storing the video signal of the reference frame, the current frame memory 31 and the reference frame memory 34. The read address data is supplied to the write address generating means 37 for supplying the address data and the current frame memory 31, the image data of the attention area is read from the current frame memory 31, and the head of the reference area in the reference frame memory 34 is read. The read address supplied to the reference frame memory 34 based on the control means 40 for generating the address data, the head read address data from the control means 40 and the motion information of the device itself from the sensor 6 for detecting the motion of the device. Readings that generate data Difference generation for obtaining the difference between the address generation means 39, the image data of the attention area read from the current frame memory 31 and the image data of the reference area in the search area read from the reference frame memory 34. Means 32, 41, accumulating means 42, 43 for accumulating difference data from the difference acquiring means 32, 41 for each reference area in the search area to obtain difference sum data, and each reference area in the search range Minimum value detecting means 44, 45 for detecting the minimum value of the difference sum data of (1) and motion vector detecting means 46 for obtaining motion vector information from the minimum difference sum data detected by the minimum value detecting means 44, 45. It is a motion detection device having.

A tenth aspect of the invention is the motion detecting device according to the ninth aspect of the invention, in which, during a search, the search is sequentially performed from the center of the search area toward the outside.

An eleventh aspect of the invention is an image pickup section 50 comprising an optical system 50a, an image pickup element 50b, and image pickup signal processing circuits 50c and 50d, and a motion detection means 6 for detecting the movement of the image pickup section 50. From the encoding means 5, the encoding means 5 which encodes the video signal from the image pickup section 50 based on the motion information from the motion detection means 6 so as to cancel the motion component of the image pickup section 50. Error correction code adding means 7 for adding an error correction code to the encoded output of
A modulation means 11 for modulating the output from the error correction code adding means 7, recording means 12, 13 for recording the modulated output from the modulation means 11 on a recording medium, and a recording signal reproduced from the recording medium. Reproduction means 13, 14 and demodulation means 1 for demodulating the outputs from the reproduction means 13, 14
5, 16 and error correction processing means 17 for performing error correction processing on the outputs from the demodulation means 15 and 16, and decoding for decoding the output from the error correction processing means 17 to obtain the original video signal. It is a camera-integrated recording device having means 18.

In a twelfth aspect based on the eleventh aspect, the variable apex angle prism 51 is driven in the optical system 50a, and the variable apex angle prism 51 is driven based on the motion information from the motion detecting means 56 and 65. Drive means 57 to 64, 6 for
6 to 73 is a camera-integrated recording device.

In a thirteenth aspect based on the eleventh aspect, the encoding means 5 comprises a current frame memory 31 for storing a video signal of a current frame, and a reference frame memory 34 for storing a video signal of a reference frame. Write address generating means 3 for supplying write address data to the current frame memory 31 and the reference frame memory 34.
7 and a control means 40 for supplying read address data to the current frame memory 31, reading image data of the attention area from the current frame memory 31, and generating head read address data of the reference area in the reference frame memory 34. , Read address generating means 39 for generating read address data to be supplied to the reference frame memory 34 based on the head read address data from the control means 40 and the motion information from the motion detecting means 6, and the current frame memory. Difference acquisition means 32, 41 for obtaining a difference between the image data of the attention area read from 31 and the image data of the reference area in the search area read from the reference frame memory 34, and the difference acquisition means 32. , 41 for each reference area in the search area Minimum value detecting means for detecting the accumulation means 42, 43 for obtaining a difference sum data cumulatively, the minimum value of the difference sum data for each reference area within the search range 44,4
5 and a motion vector detecting means 46 for obtaining motion vector information from the minimum difference sum data detected by the minimum value detecting means 44, 45.

A fourteenth aspect of the invention is the camera-integrated recording device according to the thirteenth aspect of the invention, in which a search is performed sequentially from the center of the search area toward the outside during the search.

[0037]

According to the first aspect of the invention described above, a search area is set in the previous frame image corresponding to the attention area image in the current frame image, and the reference area is sequentially moved in the search area. , The image of the reference area is read, the absolute value of the difference from the area of interest image is obtained, the difference is accumulated to obtain the difference cumulative value for each reference area, and the smallest difference cumulative value is obtained. When the search area is set in the previous frame image when using as the motion vector information, the motion vector of the device itself is obtained by the motion detection sensor 1, and the motion vector information of the device itself between the current frame and the previous frame is obtained. The midpoint of the search range is set in the negative direction of. As a result, the motion vector component of the apparatus itself can be canceled, a search area corresponding to the motion of the subject can be set, and the motion of the subject can be detected accurately at high speed.

According to the above-mentioned second invention, in the first invention, at the time of search, the search is sequentially performed from the center of the search region toward the outside. Thereby, the detection speed can be improved.

According to the third aspect of the invention described above, a search area is set in the previous frame image corresponding to the attention area image in the current frame image obtained by the video camera, and reference is made in the search area. While sequentially moving the area, the image of the reference area is read out, the absolute value of the difference from the area of interest image is obtained, the difference is accumulated to obtain the difference accumulated value for each reference area, and the difference accumulated value is calculated. When the search area is set in the preceding frame image when the smallest one is used as the motion vector information, the search range is set based on the position information of the lens of the video camera. by this,
By setting a search area corresponding to the movement of the subject, the movement of the subject can be detected accurately at high speed.

According to the above-mentioned fourth invention, in the third invention, at the time of search, the search is sequentially performed from the center of the search area toward the outside. Thereby, the detection speed can be improved.

According to the fifth aspect of the invention described above, a search area is set in the previous frame image corresponding to the attention area image in the current frame image obtained by the video camera, and reference is made in the search area. While sequentially moving the area, the image of the reference area is read out, the absolute value of the difference from the area of interest image is obtained, the difference is accumulated to obtain the difference accumulated value for each reference area, and the difference accumulated value is calculated. When the search area is set in the previous frame image when the smallest one among the above is used as the motion vector information, the motion vector of the device itself is obtained by the motion detection sensor 1, and the motion vector between the current frame and the previous frame is obtained. The midpoint of the search range is set in the negative direction of the motion vector information of the device itself, and the search range is set based on the position information of the lens of the video camera. By this, the search area corresponding to the movement of the subject is set, and high speed,
Moreover, the movement of the subject can be accurately detected.

According to the sixth aspect of the invention described above, in the fifth aspect of the invention, the search is performed sequentially from the center of the search area toward the outside. Thereby, the detection speed can be improved.

According to the seventh aspect of the invention described above, the encoding means 3 sets a search area in the previous frame image corresponding to the attention area image in the current frame image, and refers to the search area in the search area. While sequentially moving the area, the image of the reference area is read out, the absolute value of the difference from the area of interest image is obtained, the difference is accumulated to obtain the difference accumulated value for each reference area, and the difference accumulated value is calculated. In addition to the motion vector information having the smallest value, the motion information from the motion information detecting means 1 which is provided integrally with the photographing means 2 and detects the movement of the photographing means 2 itself, the lens from the photographing means 2 The search range is set based on the position information, or the motion information from the motion information detecting means 1 and the lens position information from the photographing means 2. This cancels the motion component of the photographing means itself, sets the search area corresponding to the lens position of the photographing means, sets the search area corresponding to the movement of the subject, and moves the subject accurately at high speed. Can be detected.

According to the eighth invention described above, in the seventh invention, during the search, the search is sequentially performed from the center of the search area toward the outside. Thereby, the detection speed can be improved.

According to the ninth aspect of the invention described above, the video signal of the current frame is stored in the current frame memory 31, the video signal of the reference frame is stored in the reference frame memory 34, and the write address generating means 37 is used to store the current signal. Write address data is supplied to the frame memory 31 and the reference frame memory 34, and read address data is supplied to the current frame memory 31 by the control means 40.
The image data of the area of interest is read from the current frame memory 31, the head read address data of the reference area in the reference frame memory 34 is generated, and the read address generating means 39 causes the head read address data from the control means 40 and the device. Image data of the attention area read from the current frame memory 31 by generating read address data to be supplied to the reference frame memory 34 based on the motion information of the device itself from the sensor 6 that detects the motion of The difference between the image data of the reference area in the search area read from the reference frame memory 34 is obtained by the difference acquisition means 32, 41, and the difference data from the difference acquisition means 32, 41 is referred to in the search area. Accumulating means 42 for accumulating the sum of differences for each area,
43, the minimum value of the difference sum data for each reference area in the search range is detected by the minimum value detecting means 44, 45, and from the minimum difference sum data detected by the minimum value detecting means 44, 45. The motion vector information is obtained by the motion vector detecting means 46. Accordingly, the motion vector information can be detected accurately at high speed.

According to the tenth invention described above, the ninth invention
In the invention described above, in the search, the search is sequentially performed from the center of the search area toward the outside. by this,
The detection speed can be improved.

According to the eleventh invention described above, the optical system 5
0a, the image pickup element 50b, and the processing circuit 50 for the image pickup signal
The motion detection means 6 detects the motion of the image pickup section 50 composed of c and 50d, and the encoding means 5 detects the image signal from the image pickup section 50 based on the motion information from the motion detection means 6 to obtain the image. The encoding is performed so as to cancel the motion component of the unit 50, the error correction code adding means 7 adds an error correction code to the encoded output from the encoding means 5, and the output from the error correction code adding means 7 is output. The modulation means 11 modulates, the modulated output from the modulating means 11 is recorded on a recording medium by the recording means 12 and 13, and the recording signal is reproduced from the recording medium by the reproducing means 13 and 14,
The outputs from the reproducing means 13 and 14 are demodulated means 15 and 1,
6, demodulation is performed, the output from the demodulation means 15 and 16 is subjected to error correction processing by the error correction processing means 17, and the output from the error correction processing means 17 is decoded by the decoding means 18 to obtain the original video. Get the signal. With this, the motion vector information is detected quickly and accurately,
The detection result can be recorded, and a good reproduced image can be obtained during reproduction.

According to the twelfth invention described above, the first invention
In the first aspect of the invention, a variable vertical angle prism 51 is used in the optical system 50a, and the variable vertical angle prism 51 is driven by the driving means 57 based on the motion information from the motion detecting means 56, 65.
It drives by ~ 64 and 66-73. As a result, the output of the camera shake correction sensor can be used for motion detection.

According to the thirteenth invention described above, the first invention
In the invention of claim 1, the video signal of the current frame is stored in the current frame memory 31, the video signal of the reference frame is stored in the reference frame memory 34, and the write address generating means 37 causes the current frame memory 31 and the reference frame memory. Supply write address data to 34,
The control unit 40 supplies the read address data to the current frame memory 31, and the current frame memory 3 is read.
1, the image data of the attention area is read, the head read address data of the reference area in the reference frame memory 34 is generated, and the read address generating means 39 is used.
Thus, based on the head read address data from the control means 40 and the motion information of the device itself from the sensor 6 that detects the motion of the device, the reference frame memory 3 is obtained.
4, the difference between the image data of the attention area read from the current frame memory 31 and the image data of the reference area in the search area read from the reference frame memory 34 is generated. Is obtained by the difference obtaining means 32, 41, the difference data from the difference obtaining means 32, 41 is accumulated for each reference area in the search area, and difference sum data is obtained by the accumulating means 42, 43 within the search range. The minimum value of the difference sum data for each reference area is detected by the minimum value detecting means 44, 45, and the minimum value detecting means 44,
The motion vector detecting means 46 obtains motion vector information from the minimum difference sum data detected by 45. Accordingly, the motion vector information can be detected accurately at high speed.

According to the fourteenth invention described above, the first invention
In the third aspect of the invention, at the time of search, the search is sequentially performed from the center of the search area toward the outside. Thereby, the detection speed can be improved.

[0051]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the motion detecting method and apparatus according to the present invention and a camera-integrated VTR will be described below in detail with reference to FIGS.

In the description of one embodiment of the motion detecting method and apparatus and the VTR with camera according to the present invention, the following item description will be described at the head of each item, and each item will be described in the following order.

* First embodiment A. Configuration and operation explanation of one embodiment of the motion detecting method and apparatus of the present invention and a camera-integrated VTR (see FIG. 1) B. Description of Motion Detection Method in this Example (See FIG. 2) C. Configuration of a camera-integrated VTR to which an embodiment of the motion detection method and apparatus of the present invention and a camera-integrated VTR are applied, and its operation description (see FIG. 3) D. Internal structure and operation of the compression coding circuit of the camera-integrated VTR shown in FIG. 3 (see FIG. 4) * Second embodiment The configuration and the operation of the main part of another example of the motion detection method and device and the camera-integrated VTR of the present invention (FIG. 5).
Reference) * Third Example F. Description of search for explaining still another example of the motion detection method and apparatus and the VTR with camera according to the present invention (see FIG. 6)

[First Embodiment]

A. The structure and operation of one embodiment of the motion detecting method and apparatus and the camera-integrated VTR of the present invention (see FIG. 1)

FIG. 1 is a block diagram showing an embodiment of a motion detecting method and apparatus according to the present invention and a camera-integrated VTR.

[Connection and Configuration] In FIG. 1, reference numeral 1 is a motion information detecting means including an angular velocity sensor such as a piezoelectric vibrating gyro, 2 is a photographing means such as a video camera, and 3 is a video signal Vi from the photographing means 2. , Motion information detection means 1
Encoding means for performing full search block matching processing within the search area set based on the motion information CAvec from the camera and the lens (zoom, focus) position information Sit from the photographing means Sit, and 4 are the encoding means. Three
Is a recording means for recording the compressed video signal ComVi from the recording medium onto a recording medium such as a magnetic tape, a disk, or a semiconductor memory.

Here, the condition is that the motion information detecting means 1 is integrated with the photographing means 2. This is because the movement information detecting means 1 detects the movement of the photographing means 2. Further, the motion information detecting means 1, the photographing means 2, the encoding means 3 and the recording means 4 can constitute a camera integrated VTR, a camera integrated disk recorder, a camera integrated semiconductor memory recorder and the like. Further, it is possible that the motion information detecting means 1 and the photographing means 2 constitute a video camera, and the encoding means 3 and the recording means 4 constitute a VTR or a disc recorder. Further, it is possible that the motion information detecting means 1, the photographing means 2 and the encoding means 3 constitute a video camera, and the recording means 4 constitutes a VTR or a disc recorder.

In any case, the motion information detecting means 1 and the photographing means 2 may be integrated. In the case of a so-called component type system such as a video camera and a VTR or a disk recorder, a cable is connected between the video camera and the VTR or the disk recorder, or an infrared ray or F
The connection may be made in a cordless state with M waves.

The encoding means 3 uses the motion information CAvec from the motion information detecting means 1 and the lens position information Sit from the photographing means 2 to determine the search area based on any one of the following three methods. Then, the video signal Vi is compression-encoded by the full search block matching process within this search area.

In the first method, the search area is set to the video signal Vi of the current frame from the photographing means 2 and the target block is set, and the search area is set to the video signal Vi of the previous frame corresponding to the target block. At this time, the center of the search area is set in the negative direction of the motion information CAvec between the previous frame and the current frame of the photographing means 2 itself from the motion information detecting means 1 to determine the search area.

In the second method, the search area is set to the video signal Vi of the current frame from the photographing means 2 and the target block is set, and the search area is set to the video signal Vi of the previous frame corresponding to the target block. At this time, a motion vector is predicted by using the lens position information Sit from the photographing means 2 with the screen center as a convergence point, and the search area is determined based on the predicted motion vector. For example, when the position of the objective lens is far from the photographing means, the area of the image of the subject becomes large, and when the position of the objective lens is close to the photographing means, the area of the image of the subject becomes small. That is, the subject expands or contracts toward the outside or the center of the image. Therefore,
By determining the size and position of the search area based on the position information of the objective lens, the full search block matching process can be performed in the most suitable state. For example, a method of setting a large search area at the center during zooming and a small search area at the other side can be considered.

The third method is a method in which the first method and the second method are combined.

As described above, the motion vector information CAvec due to the motion of the photographing means 2 itself is canceled by the first method, the second method, or the method combining the first and second methods, and based on the lens position. By setting, even if a narrow search area is set as a result, motion information can be reliably extracted.

[Operation] The image signal Vi of the subject from the photographing means 2 is supplied to the encoding means 3. On the other hand, the motion information detecting means 1 detects the motion of the photographing means 2 and supplies the motion information CAvec obtained as a result to the encoding means 3. Further, the lens position information Sit such as zoom and focus is supplied from the photographing means 2 to the encoding means 3.

The encoding means 3 determines the search area for the video signal Vi from the photographing means 2 based on the lens position information Sit from the photographing means 2 and the motion information CAvec from the motion information detecting means 1, and this search area Full search block matching processing is performed in the inside to perform compression coding, and the compression coded compressed video signal ComVi is supplied to the recording means 4. The recording means 4 records the compressed video signal ComVi from the encoding means 3 on a recording medium.

As described above, in the encoding means 3,
Motion information C between the previous frame and the current frame of the photographing means 2
By canceling Avec, a search area narrowing down the target is set, and the lens position information Sit is set.
The search area is set based on this, and full search block matching processing is performed in this search area, and as a result, motion detection is performed. Therefore, high image quality during decoding can be achieved without increasing the circuit scale and power consumption. You can get a video.

B. Description of the motion detection method in this example (see FIG. 2)

FIG. 2 is an explanatory diagram for explaining the motion detecting method and apparatus according to the present invention and the motion detecting method used in one embodiment of the camera-integrated VTR. The first method will be described with reference to FIG.

In FIG. 2, Sa1 is set by the same method as the search area Sa1 in FIG. 8 used as a conventional example, that is, by the method of setting without considering the motion vector information of the photographing means 2 itself as in this example. As described above, the search area Sa2 is a search area set based on the motion vector information of the photographing means 2 itself. Further, CAvec is a motion vector of the image capturing unit 2 between the current frame and the previous frame, VEC is a motion vector between the previous frame and the current frame, vec1 is a motion vector component of the image capturing unit 2 between the current frame and the previous frame, v
ec2 is the motion vector component of the subject between the current frame and the previous frame, Blk (f) is the target block of the current frame image, and Blk (f-1) is the block of the previous frame image.

As shown in FIG. 2, when a subject is photographed by the photographing means 2 and the video signal obtained by the photographing is coded by the coding means 3, it is used as a motion vector component between the current frame and the previous frame. Includes the motion vector component of the image capturing unit 2 itself and the motion vector component of the subject.

Therefore, the search area Sa1 set without considering such a thing cannot be said to be the optimum search area for the target block Blk (f) of the current frame image. In other words, the attention block Blk () of the current frame image is taken into consideration without considering that the motion vector component between the current frame and the previous frame includes the motion vector component of the photographing unit 2 itself and the motion vector component of the subject. Regarding f), if a search area for surely obtaining the motion vector of the subject is provided, it means that the search area Sa1 must be made wider.

However, in this example, considering that the motion vector component between the current frame and the previous frame includes the motion vector component of the photographing means 2 itself and the motion vector component of the subject, the current frame image is considered. When setting the search area Sa2 on the previous frame image for the target block Blk (f) of
The search area Sa2 is set so that the midpoint of the search area Sa2 is set in the negative direction from 1. The negative direction referred to here is a direction opposite to the direction of the motion vector CAvec of the photographing means 2 itself, that is, the search area Sa in the figure.
This is the direction of the search area Sa2 viewed from 1.

In this way, when setting the search area, the motion information CAvec of the photographing means 2 itself is detected by the motion information detecting means 1, and this motion vector CAve is detected.
By setting the midpoint of the search area Sa2 in the negative direction with respect to c, it is possible to equivalently cancel the motion vector component vec1 of the imaging means 2 itself between the current frame and the previous frame, and thereby the search area Sa2. Sa2
The setting of can be optimized, and the motion vector of the subject can be reliably extracted even if the area size is reduced accordingly.

This can be easily understood by assuming a case where the subject has no movement. Now, let us consider a case in which the subject being photographed by the photographing means 2 has no movement at all, and its position is the center of the image obtained by the photographing. If the photographing means 2 is not moved at all, the motion vector information cannot be obtained because the subject does not move at all even if the motion detection is performed between the current frame and the previous frame.

However, if the photographing means 2 is moved to the left side, the subject located at the center of the image obtained by photographing is positioned on the right side of the image due to the movement of the photographing means 2. Therefore, when the motion detection is performed here, the motion vector information is obtained even though the subject itself does not move. That is, the photographing means 2
The motion vector information based on its own motion should be canceled at the time of motion detection. Therefore, in this example, the search area is set as described above, and the motion vector component of the photographing means 2 itself is canceled.

Briefly, when the photographing means 2 is moved to the right, a search area is set on the left side of the image, and when the photographing means 2 is moved to the left, a search area is set on the right side of the image. However, when the photographing means 2 is moved upward, the search area is set on the lower side of the image, and when the photographing means 2 is moved downward, the search area is set on the upper side of the image.

C. The structure and operation of a camera-integrated VTR to which one embodiment of the motion detection method and apparatus and the camera-integrated VTR of the present invention are applied (see FIG. 3).

FIG. 3 is a block diagram showing an example of the structure of a camera-integrated VTR to which an embodiment of the motion detecting method and apparatus and the camera-integrated VTR of the present invention are applied.

[Connection and Configuration] In FIG. 3, reference numeral 50a is an optical lens including an objective lens, a motor for moving the objective lens for focusing, a motor for moving the objective lens for zooming, a dichroic mirror, a filter, and the like. The system 50d receives light from the optical system 50a, converts the received light into an electric signal, and outputs the CC.
D (Charge Coupled Device),
50c is an A / D converter for converting the output from the CCD 50b into digital data, and 50d is digital data from the A / D converter 50c including knee correction, white balance adjustment, γ correction, black balance adjustment,
This is a camera section signal processing circuit that obtains digital video data by performing processing such as adding a detail signal.

When the focusing motor and the zooming motor are driven, the lens position information Sit which is the position information of the objective lens is supplied to the compression encoding circuit 5. Here, the lens position information Sit may be obtained by the number of rotations of the focusing motor and the zooming motor, or the position of the objective lens may be obtained by a position sensor (a combination of a photodiode and a photodetector). Is also good.

Reference numeral 5 denotes a compression encoding circuit which compresses the video data Vi from the camera signal processing circuit 50d by obtaining the motion vector information by the above-described full search block matching processing. At this time, the lens position information Sit and the motion vector information C from the motion detection sensor 6
Based on Avec, the search area in the full search block matching process is determined as described above. The video data (motion vector information) compressed by the motion detection processing in the compression encoding circuit 5 is supplied to the error correction code adding circuit 7. Here, as the motion detection sensor 6, for example, a horizontal and vertical angular velocity sensor such as a piezoelectric vibration gyro is used.

Here, the basic search area is, for example, 5
11 pixels × 511 pixels (in case of half pitch).

Reference numeral 8 is a microphone for collecting sound, and 9 is an AD for converting a sound signal obtained as a result of collecting sound by the microphone 8 into digital sound data.
The converter 10 is an audio recording signal processing circuit for performing various recording processes such as pre-emphasis and interleaving on the digital audio data from the AD converter 9.

Reference numeral 7 generates an inner code error correction code and an outer code error correction code for the compressed video data ComVi from the compression encoding circuit 5 and the digital audio data from the audio recording signal processing circuit 10 to generate compressed video data ComVi, digital. An error correction code addition circuit for constructing a product code by voice data, an inner code error correction code and an outer code error correction code, and a channel coding circuit 11 for digitally modulating the recording data from the error correction code addition circuit 7. Is.

Reference numeral 12 is a recording amplification circuit for amplifying the digital modulation data from the channel coding circuit 11 for recording. 13 is tape transport, video
It is composed of a mechanism for pulling the tape cassette into the tape transport, a loading mechanism for pulling out the magnetic tape from the video tape cassette and winding the magnetic tape around a rotating drum (not shown), and a rotating head.

A magnetic tape 14 is mounted on the rotary drum of the tape transport 13 so that a video tape
A reproducing / amplifying circuit for amplifying an RF signal reproduced from the magnetic tape of the cassette for reproduction. A data extracting circuit 15 reproduces a clock signal from the RF signal from the reproducing / amplifying circuit 14 and extracts data by the clock signal. Is a channel coding circuit for demodulating the digital modulation data from the data extraction circuit 15 to obtain original data, that is, video data, audio data, an inner code error correction code and an outer code error correction code.

Reference numeral 17 denotes an error correction processing circuit for performing error correction processing on the video data and the audio data from the channel coding circuit 16 based on the inner code error correction code and the outer code error correction code from the channel coding circuit 16, respectively. 18 is an error correction processing circuit 18
The compression / decoding circuit for obtaining the original video data by decoding the compressed video data ComVi after the error correction processing from the reference numeral 19 is a DA for converting the video data from the compression / decoding circuit 18 into an analog video signal. It is a converter.

Reference numeral 20 is an audio reproduction signal processing circuit for performing reproduction processing such as de-emphasis and de-interleaving on the audio data after the error correction processing from the error correction processing circuit 17, and 21 is the audio reproduction signal processing. It is a DA converter that converts audio data from the circuit 20 into an analog audio signal.

Reference numeral 22 is a user interface including a recording key for the image pickup section 50, a stop key, a zooming key, a playback key for the tape transport 13, a fast forward key, a rewind key, and the like, and 23 is operation information from the user interface 22. The tape transport control signal TTcon to the tape transport 13 and the camera control signal CAco to the photographing unit 50 based on the control signal CON consisting of
It is a system controller that supplies n respectively.

[Operation During Recording] The light from the optical system 50a is photoelectrically converted by the CCD 50b and converted into an electric signal A
The signal is supplied to the -D converter 50c, converted into digital data by the AD converter 50c, and then supplied to the camera section signal processing circuit 50d. The digital data supplied to the camera unit signal processing circuit 50d is supplied to the compression encoding circuit 5 as digital video data after being subjected to various processes such as the various white balance adjustments described above. In the compression encoding circuit 5, 8 pixels × 8 pixels (or 16 pixels × 16 pixels) are included in the video data of the current frame of the video data Vi from the camera unit signal processing circuit 50d.
The target block of (pixel) is set, and the search area for performing the full search block matching process is set in the video data of the previous frame corresponding to the target block set in the video data of the current frame.

At this time, as described above, the search area is set by setting the midpoint of the search area in the negative direction with respect to the motion vector between the current frame and the previous frame of the apparatus (camera integrated type VTR in this example) itself. It is done to set. Then, in this search area, full search block matching processing is performed, and the compressed video data ComVi compressed as a result is supplied to the error correction code adding circuit 7.

On the other hand, the sound collected by the microphone 8 is supplied as an audio signal to the AD converter 9, converted into digital audio data by the AD converter 9, and then supplied to the audio recording signal processing circuit 10. It The digital audio data that has been subjected to the recording processing such as the pre-emphasis or interleave in the audio recording signal processing circuit 10 is supplied to the error correction code adding circuit 7.

The compressed video data ComVi and the digital audio data supplied to the error correction code addition circuit 7 are supplied to the channel coding circuit 11 after being added with the inner code error correction code and the outer code error correction code, and this channel coding is performed. After the digital modulation processing is performed in the circuit 11, the recording amplification circuit 12
Is supplied to the magnetic head mounted on the rotary drum of the tape transport 13 via. Then, the digital modulation data supplied to the magnetic head is recorded by the magnetic head so as to form an inclined track on the magnetic tape.

[Operation During Reproduction] When the reproduction key of the user interface 22 is pressed, a control signal CON indicating reproduction is supplied from the user interface 22 to the system controller 23. The system controller 23 supplies the tape transport control signal TTcon to the tape transport 13 based on the control signal CON from the user interface 22. As a result, the tape transport 13 scans the recording signal recorded on the magnetic tape by the magnetic head mounted on the rotary head and reproduces the recording signal.

The reproduced signal is supplied to the data extraction circuit 15 via the recording / amplification circuit 14, where the clock signal is extracted, and the data is extracted by the clock signal. The extracted data is supplied to the channel decoding circuit 16. Channel decoding circuit 1
The data supplied to No. 6 is demodulated and compressed video data C
omVi, digital voice data, inner code error correction code, and outer code error correction code. These data are supplied to the error correction processing circuit 17. The compressed video data ComVi and the digital audio data supplied to the error correction circuit 17 are subjected to error correction processing by an inner code error correction code and an outer code error correction code, respectively. At this time, it goes without saying that error concealment processing is performed on uncorrectable data.

The compressed video data ComVi after error correction from the error correction processing circuit 17 is decoded in the compression decoding circuit 18 to be the original digital video data,
After being converted into an analog video signal in the D-A converter 19, it is output through the output terminal 19a. on the other hand,
The error-corrected digital audio data from the error correction processing circuit 17 is subjected to reproduction processing such as de-emphasis and de-interleave processing in the audio reproduction signal processing circuit 20 and then converted into an analog audio signal in the DA converter 21. After the conversion, it is output via the output terminal 21a.

D. Internal structure and operation of the compression coding circuit of the camera-integrated VTR shown in FIG. 3 (see FIG. 4)

FIG. 4 is a block diagram showing an internal configuration example of the compression coding circuit 5 of the camera-integrated VTR shown in FIG.

[Connection and Configuration] In FIG. 4, reference numeral 30 is an input terminal to which the video signal Vi from the camera section signal processing circuit 50d shown in FIG. 3 is supplied. This input terminal 30 is connected to the data input terminal of the current frame memory 31 and the delay circuit 3
3, the data input terminals of the reference frame memory 34 are respectively connected, the data output terminals of the current frame memory 31 are connected to the addition side input terminals of the adder circuit 32, and the data output terminals of the reference frame memory 34 are added. 32 is connected to the input terminal on the subtraction side. Then, the output terminal of the adder circuit 32 is connected to the input terminal of the absolute value circuit 41. The absolute value circuit 41 obtains the absolute value of the addition result from the addition circuit 32.

Then, the output terminal of the absolute value circuit 41 is connected to one of the addition side input terminals of the addition circuit 42, the output terminal of the addition circuit 42 is connected to the data input terminal of the memory 44, and the delay circuit is connected. Adder circuit 42 via 43
Connect to the input terminal on the other addition side of.

Reference numeral 35 is an input terminal to which the write clock signal WCLK from the system controller 23 shown in FIG. 3 is supplied, and 36 is the system controller 2 shown in FIG.
3 is an input terminal to which the synchronizing signal SYNC is supplied, and these input terminals 35 and 36 are respectively connected to the input terminals of the write address generating circuit 37, and one output of the write address generating circuit 37 for one write address signal. The terminal is connected to the input terminal for the write address signal of the current frame memory 31, and the write address generating circuit 37
The other output terminal for the write address signal is connected to the input terminal for the write address signal of the reference frame memory 34.

Reference numeral 38 is an input terminal to which the motion vector data CAvec of the device itself from the motion detection sensor 6 shown in FIG. 3 is supplied, and this input terminal 38 is connected to the input terminal of the address decoder 39b of the read address generating circuit 39. Then, the output terminal for the read address signal of the address decoder 39b is connected to the input terminal for the read address signal of the reference frame memory 34, and the data output terminal of the counter 39a is connected to the data input terminal of the address decoder 39b. Reference numeral 40 denotes a controller which supplies the read clock signal RCLK and the read start address data Sadd to the counter 39a, and supplies the read clock signal RCLK and the block address signal Badd for designating the target block to the current frame memory 31.

Reference numeral 45 denotes a minimum value detection circuit, which is used for all blocks (for example, 256 blocks) in one search area stored in the memory.
The minimum difference absolute value sum data among the respective difference absolute value sum data is detected. Reference numeral 46 denotes a motion vector based on the minimum value difference absolute value sum data from the minimum value detection circuit 45, and the motion vector data vec is used as compressed video data ComVi to output the error shown in FIG. The correction code adding circuit 7 and the controller 40 are supplied.

Here, the controller 40 supplies the read start address data Sadd, that is, the head address data of each block to the counter 39a based on the motion vector data vec from the motion vector detection circuit 46.

Further, the read address generating circuit 39 described above.
In order to shift the search area, the position indicated by the start address of the block indicated by the read start address data Sadd from the controller 40 (for example, the upper left corner on the screen) is opposite to the direction indicated by the motion vector information CAvec at that time. In order to shift the position by the same amount in the direction, the read start address data Sadd from the controller 40 and the count value data obtained by sequentially counting from the read start address data by the counter 39a are sequentially moved to the motion vector information CAvec.
Decoding is performed by the decoder 39b according to the above.

For example, when the head position of the block indicated by the read start address data Sadd supplied from the controller 40 is at the upper left corner of the search area Sa1 shown in FIG. 2 (that is, the first block of the search area). ), The counter 39a of the read address generation circuit 39 sequentially counts and obtains count value data indicating each part of the block of the search area Sa1. Here, when the motion vector information CAvec is supplied, the read address generation circuit 39a The decoder 39b of 39 outputs all the outputs of the counter 39a to the motion vector information CA.
The read address data is obtained by decoding the value corresponding to vec. Then, for example, when the count value from the counter 39a is a value indicating the upper left corner of the search area Sa1, the value is decoded and decoded into read address data indicating the upper left corner of the search area Sa2. Then, similarly, the count value sequentially supplied from the counter 39b is decoded into a value indicating each part of the corresponding search area Sa2.

[Modification] Read address generation circuit 39
May be composed of only the counter 39a, the input terminal 38 may be connected to the controller 40, and the controller 40 may give the read start address signal Sadd according to the motion vector data CAvec to the counter 39a.

Further, without using the address decoder 39b, for example, the count value obtained in the counter 39a and the motion vector data CAvec may be added or subtracted. Specifically, for example, when the motion vector data CAvec is composed of the output from the horizontal angular velocity sensor and the output from the vertical angular velocity sensor, the following is performed.

If the output from the horizontal angular velocity sensor indicates that the device is moving to the left, the same movement amount to the right, that is, the horizontal count value of the counter 39a (horizontal address). If the output from the angular velocity sensor in the horizontal direction indicates that the device is moving to the right, the value corresponding to the amount of movement is added to, and if the output is moving to the right, the same amount of movement to the left, that is, the horizontal of the counter 39a. If the output from the vertical angular velocity sensor indicates that the device is moving upward, subtract it from the count value (horizontal address) in the direction, and the value is the same as in the lower direction. Only the amount of movement, that is, the value corresponding to the amount of movement is added to the vertical count value (vertical address) of the counter 39a, and the output from the vertical angular velocity sensor causes the device to move downward. If the show that Iteiru, by the same amount of movement upward, i.e., the vertical count value of the counter 39a may be subtracted value corresponding to the amount of movement from the (vertical address).

The search area may be made large when the subject is enlarged by zooming or the like, and may be made small when the subject is reduced.

[Operation] The video data Vi from the camera unit signal processing circuit 50d of the image pickup unit 50 shown in FIG. 3 is supplied to the current frame memory 34 through the input terminal 30, and the write address from the write address generation circuit 37. Written by signal. On the other hand, the video data Vi of the previous frame delayed by one frame by the delay circuit 33 is
It is supplied to the reference frame memory 34 and written by the write address signal from the write address generation circuit 37.

Here, the current frame image data Vi (f) is stored in the current frame memory 31, and the previous frame image data Vi (f-1) is stored in the reference frame memory 34.
Will be remembered respectively.

The controller 40 uses the current frame memory 3
1 is supplied with a read clock signal RCLK and a block read address signal Badd for sequentially reading video data corresponding to the target block. by this,
The video data corresponding to the target block is sequentially read from the current frame memory 31 and supplied to the adding circuit 32.
On the other hand, the controller 40 causes the counter 39a to read the read clock signal RCLK and the read start address signal S.
Supply add. As a result, the counter 39a starts counting from the value indicated by the read start address signal and supplies the count value data to the address decoder 39b.

The address decoder 39b has an input terminal 38.
The count value data supplied from the counter 39a is decoded on the basis of the motion vector data CAvec supplied via to obtain a read address signal, and the read address signal is supplied to the reference frame memory 34.
As a result, the video data of the corresponding block in the search area is read from the reference frame memory 34 and supplied to the adding circuit 32.

In the adder circuit 32, the video data of the block in the search area read from the reference frame memory 34 is subtracted in pixel units from the video data of the target block read from the current frame memory 31. The addition result of the video data of the previous frame and the video data of the current frame is supplied to the absolute value circuit 41, converted into absolute difference value data in the absolute value circuit 41, and further supplied to the addition circuit 42.

The absolute difference value data supplied to the adder circuit 42 is supplied to the memory 44 and stored in the memory 44. Next, the difference absolute value data corresponding to the next pixel supplied to the adding circuit 42 is added to the difference absolute value data corresponding to the previous pixel which is delayed and held in the delay circuit 43, and the difference absolute value sum is added. After being converted into data, it is supplied to the memory 4 and stored in the memory 4. That is, the difference between the pixel data of the target block of the current frame and the pixel data of the block of the previous frame is calculated, the difference data is sequentially added to obtain the difference absolute value sum data, and the difference absolute value sum data is stored in the memory 44. By performing the processing such as storing in each block, the difference absolute value sum data of, for example, 256 blocks can be finally stored in the memory 44.

When the search processing by all blocks in the search area is completed, the minimum value detection circuit 45 has the smallest difference absolute value among the difference absolute value sum data of the 256 blocks stored in the memory 44. The sum data is selected and the difference absolute value sum data is supplied to the motion vector detection circuit 45. The motion vector detection circuit 46 obtains motion vector data vec based on the difference absolute value sum data from the minimum value detection circuit 45, and this motion vector data ve
c is supplied as compressed video data ComVi to the error correction code adding circuit 7 and the controller 40 shown in FIG. 3 via the output terminal 47, respectively. Then, the same processing is performed for each block of interest on a frame-by-frame basis to convert the video data Vi into the compressed video data ComVi.

[Effects of the First Embodiment] As described above, in the present embodiment, the video signal obtained by shooting with the video camera is compression-encoded by the full search block matching process and transmitted or recorded. At this time, the search area is determined based on the motion vector information of the video camera itself and the position information of the lens, and full search block matching processing is performed using that search area. The power consumption can be reduced and the image quality at the time of decoding can be improved.

[Second Embodiment]

E. The motion detecting method and apparatus according to the present invention, and the configuration and operation of the main part of another example of the camera-integrated VTR (see FIG. 5)

FIG. 5 is a block diagram showing the essential portions of another example of the motion detecting method and apparatus according to the present invention and the camera-integrated VTR. In the image pickup section 50 of the camera-integrated VTR shown in FIG. It is an example of composition of an image pick-up part when using a vertical angle prism. Further, in this example, a horizontal angular velocity sensor 65 and a vertical angular velocity sensor 56 are used as the motion detection sensor 6 shown in FIG.
The parts other than the imaging unit 50 are the same as those in FIG. 3, and therefore, the illustration and description of the other parts are omitted.

The camera-integrated VTR shown in this figure can correct camera shake. Shake is a movement of a photographed image vertically and horizontally as the person moves his / her body regardless of his / her will when he / she shoots a subject with a video camera. In order to avoid such a phenomenon, a prism 51 as shown in FIG. 5 is used, the change in angular velocity due to camera shake is detected by the angular velocity sensors 56 and 65, and the prism 51 is driven based on the detected information. Light from the subject is refracted by the prism 51 so that the light from the subject does not move on the imaging surface of the CCD 52.

[Connection and Configuration] As shown in the figure, the prism 51 includes two plate glasses 51a and 51b, which are
The flat glass plates 51a and 51b are joined together by a transparent bellows-shaped spring portion 51c having a diameter substantially equal to each other.
The inside is filled with a liquid 51d having a refractive index substantially equal to that of glass. The prism 51 has a horizontal axis 51e.
Also, the vertical shaft 51f is attached to the housing of the video camera or a dedicated mounting member so as to be rotatable in the directions of the arrows near the respective shafts 51e and 51f.

Reference numeral 52 is a CCD for photoelectrically converting the light from the prism 51 to obtain an electric signal, 53 is an AD converter for converting the electric signal from the CCD 52 into a digital signal, and 54 is a digital signal from the AD converter 53. The camera unit signal processing circuit performs various signal processing such as white balance adjustment, black balance adjustment, knee correction, γ correction, and addition of a detail signal on the signal.

The system of the vertical axis 51f of the prism 51 includes an angular velocity sensor 56 for detecting the angular velocity in the vertical direction, and an AD converter 57, AD for converting the detection output from the angular velocity sensor 56 into a digital signal. A digital integration circuit 58 that integrates the digital signal from the converter 57, a DA converter 59 that converts the integrated output from the digital integration circuit 58 into an analog signal, the output of the DA converter 59 and the apex angle sensor 64. An adder circuit 60 for adding the detection outputs, an adder circuit 61 for adding the output of the adder circuit 60 and the output of the braking coil 63,
An actuator 6 that drives the prism 51 in the vertical direction about the vertical axis 51f based on the output of the adder circuit 61.
2, a braking coil 63, and an apex angle sensor 64 for detecting the vertical position of the vertical shaft 51f.

The system of the horizontal axis 51e of the prism 51 is an angular velocity sensor 65 for detecting an angular velocity in the horizontal direction, and an AD converter 66, AD for converting the detection output from the angular velocity sensor 65 into a digital signal. A digital integrating circuit 67 for integrating the digital signal from the converter 66, a DA converter 68 for converting the integrated output from the digital integrating circuit 67 into an analog signal, an output of the DA converter 68 and an apex angle sensor 73. An adder circuit 69 for adding the detection outputs, an adder circuit 70 for adding the output of the adder circuit 69 and the output of the braking coil 72,
An actuator 7 that drives the prism 51 horizontally based on the output of the adder circuit 70 around the horizontal axis 51e.
1, a braking coil 72, and an apex angle sensor 73 for detecting the horizontal position of the horizontal shaft 51e.

74 obtains motion vector information based on the detection outputs from the angular velocity sensors 56 and 65, and supplies this motion vector information CAvec to the compression coding circuit 5 shown in FIG. The information output circuit 74. The input terminal for the horizontal detection signal of the motion information output circuit 74 is connected to the connection point between the digital integration circuit 58 and the AD converter 57, and the input terminal for the vertical detection signal of the motion information output circuit 74. Is connected to the connection point between the digital integration circuit 67 and the AD converter 66.

Here, the motion information output circuit 74, for example, based on the detection output of the horizontal angular velocity and the detection output of the vertical angular velocity, the read start address data Sadd in the address decoder 39b shown in FIG. The motion vector information CAvec indicating the shift amount is obtained.

[Operation] The detection output from the angular velocity sensor 56 is integrated in the digital integration circuit 58 and then D-
It is converted into an analog signal in the A converter 59, and is added to the detection output from the apex angle sensor 64 in the adding circuit 60. The addition output of the addition circuit 60 is the addition circuit 6
At 1, the output is added to the output from the braking coil 63 and supplied to the actuator 62 as a drive signal. As a result, the prism 51 is driven in the vertical direction, and the vertical movement of the image due to camera shake is corrected.

On the other hand, the detection output from the angular velocity sensor 65 is integrated in the digital integration circuit 67 and then D-A
It is converted into an analog signal in the converter 68, and is added to the detection output from the apex angle sensor 73 in the adding circuit 69. The addition output of the addition circuit 69 is the addition circuit 70.
At, the output is added to the output from the braking coil 72 and is supplied to the actuator 71 as a drive signal. As a result, the prism 51 is driven in the horizontal direction, and the horizontal movement of the image due to camera shake is corrected.

Further, along with this, the motion information output circuit 7
4, the detection outputs from the angular velocity sensors 56 and 65 are respectively supplied, the motion information output circuit 74 obtains the motion vector information CAvec based on these detection outputs, and the motion vector information CAvec is output via the output terminal 75. The address decoder 39b shown in FIG. As a result, in the compression encoding circuit 5 shown in FIG. 4, the most suitable search area is determined for the target block of the current frame, the motion vector of the subject is detected quickly and reliably, and the compression encoding is favorably performed. Can be processed.

The applicant of the present invention has previously filed the above-mentioned image stabilization (see Japanese Patent Application No. 4-63905).

[Effects of the Second Embodiment] As described above, in the present embodiment, the video signal obtained by shooting with the video camera is compression-encoded by the full search block matching process and transmitted or recorded. At this time, the search area is determined by using the outputs of the angular velocity sensors 56 and 65 used for camera shake correction, and the full search block matching process is performed using the search area. The power consumption can be reduced and the image quality at the time of decoding can be improved. Further, in this case, since the angular velocity sensors 56 and 65 are not provided for determining the search area, it is possible to minimize the number of components of the camera-integrated VTR and the video camera, and to enhance the functionality.

[Third Embodiment]

F. Description of search for explaining still another example of the motion detection method and apparatus and the VTR with camera according to the present invention (see FIG. 6)

FIG. 6 is an explanatory diagram for explaining a method of searching the search area.

Normally, in the search area Sa3,
The search is performed by linearly shifting the block by one pixel from the upper left corner to the right, then lowering it by one pixel, and then moving to the left. However, as shown by the solid arrow in the figure, the address operation is performed. Alternatively, the blocks may be shifted by one pixel in the clockwise direction (or counterclockwise direction) from the center toward the outside to perform the search.

In this case, if the method for limiting the search area according to the first and second embodiments is used together,
If the search process is terminated faster when a value equal to or less than the set value in the search area Sa3 appears, encoding can be performed at higher speed. The certain set value referred to here is a set value corresponding to the sum of absolute differences described above.

[0140]

According to the first invention described above, a search area is set in the previous frame image in correspondence with the attention area image in the current frame image, and the reference area is sequentially moved in the search area. At the same time, the image of the reference area is read out, the absolute value of the difference from the attention area image is obtained, the difference is accumulated to obtain the difference accumulated value for each reference area, and the difference accumulated value is determined to be the smallest value. When setting a search area in the previous frame image when the following is used as the motion vector information, the motion vector of the device itself is used as the motion detection sensor 1.
Since the middle point of the search range is set in the negative direction of the motion vector information of the device itself between the current frame and the previous frame, the motion vector component of the device itself is canceled and the motion of the subject is dealt with. It is possible to set the search area and detect the motion of the subject at high speed and accurately, thereby increasing the motion detection speed of the subject, reducing the amount of calculation, and obtaining a good image during decoding. effective.

According to the second aspect of the invention described above, in the first aspect of the invention, during the search, the search is sequentially performed from the center of the search area toward the outside, so that the detection speed is improved. Therefore, there is an effect that the processing speed can be further improved.

According to the third invention described above, a search area is set in the previous frame image corresponding to the attention area image in the current frame image obtained by the video camera, and reference is made in the search area. While sequentially moving the area, the image of the reference area is read out, the absolute value of the difference from the area of interest image is obtained, the difference is accumulated to obtain the difference accumulated value for each reference area, and the difference accumulated value is calculated. When the search area is set in the previous frame image when the smallest one is used as the motion vector information, the search range is set based on the position information of the lens of the video camera. ,
By setting a search area that corresponds to the movement of the subject, the movement of the subject can be detected accurately at high speed. This makes the movement detection speed of the subject high, reduces the amount of calculation, and reduces the circuit scale. There is an effect that the power consumption of the circuit is reduced and a good image can be obtained at the time of decoding.

According to the fourth aspect of the invention described above, in the third aspect of the invention, during the search, the search is sequentially performed from the center of the search area toward the outside, so that the detection speed is improved. Therefore, there is an effect that the processing speed can be further improved.

According to the fifth aspect of the invention described above, a search area is set in the previous frame image corresponding to the attention area image in the current frame image obtained by the video camera, and reference is made in the search area. While sequentially moving the area, the image of the reference area is read out, the absolute value of the difference from the area of interest image is obtained, the difference is accumulated to obtain the difference cumulative value for each reference area, and the difference cumulative value is calculated. When the search area is set in the previous frame image when the smallest one among the above is used as the motion vector information, the motion vector of the device itself is obtained by the motion detection sensor 1, and the motion vector between the current frame and the previous frame is obtained. Since the middle point of the search range is set in the negative direction of the motion vector information of the device itself, and the search range is set based on the position information of the video camera lens, the movement of the subject Set the corresponding search area, high-speed,
In addition, it is possible to accurately detect the movement of the subject, thereby increasing the movement detection speed of the subject, reducing the amount of calculation, reducing the circuit scale, and reducing the power consumption of the circuit.
There is an effect that a good image can be obtained at the time of decoding.

According to the sixth invention described above, in the fifth invention, at the time of the search, the search is sequentially performed from the center of the search area toward the outside, so that the detection speed is improved. Therefore, there is an effect that the processing speed can be further improved.

According to the seventh aspect of the invention described above, the encoding means 3 sets a search area in the previous frame image in correspondence with the attention area image in the current frame image, and makes a reference in the search area. While sequentially moving the area, the image of the reference area is read out, the absolute value of the difference from the area of interest image is obtained, the difference is accumulated to obtain the difference accumulated value for each reference area, and the difference accumulated value is calculated. In addition to the motion vector information having the smallest value, the motion information from the motion information detecting means 1 which is provided integrally with the photographing means 2 and detects the movement of the photographing means 2 itself, the lens from the photographing means 2 Since the search range is set based on the position information or the movement information from the movement information detecting means 1 and the lens position information from the photographing means 2, the movement component of the photographing means itself is set. Then, the search area corresponding to the lens position of the photographing means is set, the search area corresponding to the movement of the subject is set, and the movement of the subject can be detected accurately at high speed. There is an effect that the motion detection speed can be increased, the calculation amount can be reduced to reduce the circuit scale, the power consumption of the circuit can be reduced, and a good image can be obtained at the time of decoding.

According to the eighth invention described above, in the seventh invention, at the time of the search, the search is sequentially performed from the center of the search area toward the outside, so that the detection speed is improved. Therefore, there is an effect that the processing speed can be further improved.

According to the ninth aspect of the invention described above, the video signal of the current frame is stored in the current frame memory 31, the video signal of the reference frame is stored in the reference frame memory 34, and the write address generating means 37 causes the current signal to be stored. Write address data is supplied to the frame memory 31 and the reference frame memory 34, and read address data is supplied to the current frame memory 31 by the control means 40.
The image data of the area of interest is read from the current frame memory 31, the head read address data of the reference area in the reference frame memory 34 is generated, and the read address generating means 39 causes the head read address data from the control means 40 and the device. Image data of the attention area read from the current frame memory 31 by generating read address data to be supplied to the reference frame memory 34 based on the motion information of the device itself from the sensor 6 that detects the motion of The difference between the image data of the reference area in the search area read from the reference frame memory 34 is obtained by the difference acquisition means 32, 41, and the difference data from the difference acquisition means 32, 41 is referred to in the search area. Accumulating means 42 for accumulating the sum of differences for each area,
43, the minimum value of the difference sum data for each reference area in the search range is detected by the minimum value detecting means 44, 45, and from the minimum difference sum data detected by the minimum value detecting means 44, 45. Since the motion vector information is obtained by the motion vector detecting means 46, the motion vector information can be detected at high speed and accurately.
There are effects that the motion detection speed of the object is increased, the amount of calculation is reduced, the circuit scale is reduced, the power consumption of the circuit is reduced, and a good image can be obtained at the time of decoding.

According to the tenth invention described above, the ninth invention
In the invention described above, at the time of search, the search is sequentially performed from the center of the search area toward the outside,
The detection speed can be improved, which has the effect of further improving the processing speed.

According to the eleventh invention described above, the optical system 5
0a, the image pickup element 50b, and the processing circuit 50 for the image pickup signal
The motion detection means 6 detects the motion of the image pickup section 50 composed of c and 50d, and the encoding means 5 detects the image signal from the image pickup section 50 based on the motion information from the motion detection means 6 to obtain the image. The encoding is performed so as to cancel the motion component of the unit 50, the error correction code adding means 7 adds an error correction code to the encoded output from the encoding means 5, and the output from the error correction code adding means 7 is output. The modulation means 11 modulates, the modulated output from the modulating means 11 is recorded on a recording medium by the recording means 12 and 13, and the recording signal is reproduced from the recording medium by the reproducing means 13 and 14,
The outputs from the reproducing means 13 and 14 are demodulated means 15 and 1,
6, demodulation is performed, the output from the demodulation means 15 and 16 is subjected to error correction processing by the error correction processing means 17, and the output from the error correction processing means 17 is decoded by the decoding means 18 to obtain the original video. Since the signal is obtained, the motion vector information is detected at high speed and accurately,
The detection result can be recorded, and a good reproduced image can be obtained even at the time of reproduction, thereby increasing the motion detection speed of the subject, reducing the amount of calculation, and downsizing the main body,
There is an effect that the power consumption of the circuit is reduced, the continuous use time is extended when a battery or the like is used, and a good image can be obtained at the time of decoding.

According to the twelfth invention described above, the first invention
In the first aspect of the invention, a variable vertical angle prism 51 is used in the optical system 50a, and the variable vertical angle prism 51 is driven by the driving means 57 based on the motion information from the motion detecting means 56, 65.
It drives by ~ 64 and 66-73. As a result, the output of the camera shake correction sensor can be used for motion detection.

According to the thirteenth invention described above, the first invention
In the invention of claim 1, the video signal of the current frame is stored in the current frame memory 31, the video signal of the reference frame is stored in the reference frame memory 34, and the write address generating means 37 causes the current frame memory 31 and the reference frame memory. Supply write address data to 34,
The control unit 40 supplies the read address data to the current frame memory 31, and the current frame memory 3 is read.
1, the image data of the attention area is read, the head read address data of the reference area in the reference frame memory 34 is generated, and the read address generating means 39 is used.
Thus, based on the head read address data from the control means 40 and the motion information of the device itself from the sensor 6 that detects the motion of the device, the reference frame memory 3 is obtained.
4, the difference between the image data of the attention area read from the current frame memory 31 and the image data of the reference area in the search area read from the reference frame memory 34 is generated. Is obtained by the difference obtaining means 32, 41, the difference data from the difference obtaining means 32, 41 is accumulated for each reference area in the search area, and difference sum data is obtained by the accumulating means 42, 43 within the search range. The minimum value of the difference sum data for each reference area is detected by the minimum value detecting means 44, 45, and the minimum value detecting means 44,
Since the motion vector information is obtained by the motion vector detecting means 46 from the minimum difference sum data detected by 45, the motion vector information can be detected at high speed and accurately, whereby the motion detection speed of the subject. Is faster, the amount of calculation is reduced to reduce the circuit scale, the power consumption of the circuit is reduced, and a good image can be obtained at the time of decoding.

According to the fourteenth invention described above, the first invention
In the third aspect of the invention, since the search is sequentially performed from the center of the search area toward the outside during the search, the detection speed can be improved.
Further, there is an effect that the processing speed can be improved.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of a motion detecting method and apparatus according to the present invention and a camera-integrated VTR.

FIG. 2 is a camera-integrated VT to which an embodiment of the motion detection method and apparatus and the camera-integrated VTR of the present invention is applied.
It is a block diagram which shows the structural example of R.

FIG. 3 is a configuration diagram showing an example of the internal configuration of a compression encoding circuit of the camera-integrated VTR shown in FIG.

FIG. 4 is an explanatory diagram for explaining a motion detection process of an embodiment of a motion detection method and apparatus according to the present invention and a camera-integrated VTR.

FIG. 5 is a configuration diagram showing a configuration example of a main part of another example of the embodiment of the motion detection method and apparatus according to the present invention and the camera-integrated VTR.

FIG. 6 is an explanatory diagram for explaining a search used for explaining still another example of the embodiment of the motion detecting method and apparatus according to the present invention and the camera-integrated VTR.

FIG. 7 is an explanatory diagram showing an example of a conventional motion detection device.

FIG. 8 is an explanatory diagram for explaining a conventional motion detection device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 motion information detection means 2 photographing means 3 encoding means 4 recording means 5 compression encoding circuit 6 motion detection sensor 7 error correction code addition circuit 11 channel coding circuit 12 recording amplification circuit 13 tape transport 14 reproduction amplification circuit 15 data extraction circuit 16 channel decoding circuit 17 error correction processing circuit 18 compression decoding circuit 31 current frame memory 32, 42 adder circuit 34 reference frame memory 37 write address generating circuit 39 read address generating circuit 40 controller 41 absolute value circuit 43 delay circuit 44 memory 45 Minimum value detection circuit 46 Motion vector detection circuit 50 Imaging section 50a Optical system 50b, 52 CCD 50c, 53, 57 A-D converter 50d Camera section signal processing circuit 50d 51 Prism 56, 65 Speed sensors 58,67 digital integrator circuit 59,68 D-A converter 60,61,69,70 adder circuit 62,71 actuator 63 and 72 brake coil 64 and 73 apex angle sensor

Continuation of front page (51) Int.Cl. ⁶ Identification number Office reference number FI Technical display location H04N 5/92 7/18 K H04N 5/92 H

Claims (14)

[Claims] 1. A search region is set in a previous frame image corresponding to a region of interest image in a current frame image, reference regions are sequentially moved within the search region, and an image in the reference region is read out. , A difference absolute value with respect to the attention area image is obtained, the difference is accumulated to obtain a difference accumulated value for each reference area, and the smallest of the difference accumulated values is used as motion vector information. In the method, when setting the search area in the previous frame image,
A motion detection method characterized in that the motion vector of the device itself is obtained by a motion detection sensor, and the midpoint of the search range is set in the negative direction of the motion vector information of the device itself between the current frame and the previous frame. 2. The motion detection method according to claim 1, wherein during the search, the search is sequentially performed from the center of the search area toward the outside. 3. A search area is set in the previous frame image corresponding to the attention area image in the current frame image obtained by the video camera, and the reference area is sequentially moved in the search area, and The image of the reference area is read out, the absolute value of the difference from the area of interest image is obtained, the difference is accumulated to obtain the difference accumulated value for each reference area, and the smallest difference accumulated value is moved. A motion detection method using vector information, where a search area is set in the previous frame image,
A motion detection method characterized in that the search range is set based on position information of a lens of a video camera. 4. The motion detection method according to claim 3, wherein during the search, the search is sequentially performed from the center of the search area toward the outside. 5. A search area is set in the previous frame image corresponding to the attention area image in the current frame image obtained by the video camera, and the reference area is sequentially moved in the search area, The image of the reference area is read out, the absolute value of the difference from the area of interest image is obtained, the difference is accumulated to obtain the difference accumulated value for each reference area, and the smallest difference accumulated value is moved. A motion detection method using vector information, where a search area is set in the previous frame image,
The motion vector of the device itself is obtained by the motion detection sensor, and the midpoint of the search range is set in the negative direction of the motion vector information of the device itself between the current frame and the previous frame, and based on the position information of the lens of the video camera. A motion detection method characterized in that the search range is set. 6. The motion detection method according to claim 5, wherein during the search, the search is sequentially performed from the center of the search area toward the outside. 7. A photographing means, a motion information detecting means which is provided integrally with the photographing means and detects a movement of the photographing means itself, and a region of interest in the current frame image corresponding to a region of interest image in the previous frame image. Set the search area to, and within the search area,
While sequentially moving the reference area, the image of the reference area is read, the absolute value of the difference from the area of interest image is obtained, the difference is accumulated to obtain a difference cumulative value for each reference area, and the difference cumulative value is calculated. With the smallest of the motion vector information, the motion information from the motion information detecting means,
A motion including a coding unit for setting the search range based on the lens position information from the photographing unit, or the motion information from the motion information detecting unit and the lens position information from the photographing unit. Detection device. 8. The motion detection apparatus according to claim 7, wherein during the search, the search is sequentially performed from the center of the search area toward the outside. 9. A current frame memory for storing a video signal of a current frame, a reference frame memory for storing a video signal of a reference frame, and a write address generation for supplying write address data to the current frame memory and the reference frame memory. Means for supplying read address data to the current frame memory, reading image data of a region of interest from the current frame memory, and generating head read address data of a reference region in the reference frame memory; Read address generating means for generating read address data to be supplied to the reference frame memory based on the head read address data from the device and motion information of the device itself from the sensor for detecting the motion of the device, and read from the current frame memory. Difference acquisition means for obtaining a difference between the image data of the attention area and the image data of the reference area in the search area read from the reference frame memory, and the difference data from the difference acquisition means in the search area. Accumulation means for accumulating difference sum data for each reference area, minimum value detection means for detecting the minimum value of the difference sum data for each reference area within the search range, and minimum value detection means for detecting the minimum value. And a motion vector detecting means for obtaining motion vector information from the minimum difference sum data. 10. The motion detection apparatus according to claim 9, wherein during the search, the search is sequentially performed from the center of the search area toward the outside. 11. An image pickup section comprising an optical system, an image pickup element, and a processing circuit for an image pickup signal, a motion detecting means for detecting a movement of the image pickup section, and a motion detection of a video signal from the image pickup section. Coding means for coding so as to cancel the motion component of the image pickup unit based on motion information from the means, and error correction code adding means for adding an error correction code to the coded output from the coding means. Modulating means for modulating the output from the error correction code adding means, recording means for recording the modulated output from the modulating means on a recording medium, reproducing means for reproducing a recording signal from the recording medium, and the reproducing means. Demodulating means for demodulating the output from the demodulating means, error correcting processing means for performing error correction processing on the output from the demodulating means, and decoding the output from the error correcting processing means. A camera-integrated recording device, comprising: a decoding means for obtaining an original video signal. 12. The camera according to claim 11, wherein the optical system includes a variable apex angle prism, and drive means for driving the variable apex angle prism based on the motion information from the motion detecting means. Integrated recording device. 13. The encoding means includes: a current frame memory for storing a video signal of a current frame; a reference frame memory for storing a video signal of a reference frame; and write address data in the current frame memory and the reference frame memory. And write address generating means for supplying the read address data to the current frame memory to read the image data of the region of interest from the current frame memory and generate head read address data of the reference region in the reference frame memory. Means, read address generating means for generating read address data to be supplied to the reference frame memory based on the head read address data from the control means and the motion information from the motion detecting means, and read from the current frame memory. Was issued Difference acquisition means for obtaining a difference between the image data of the attention area and the image data of the reference area in the search area read from the reference frame memory, and the difference data from the difference acquisition means is referred to in the search area. Accumulation means for accumulating difference sum data for each area, minimum value detection means for detecting the minimum value of the difference sum data for each reference area within the search range, and minimum value detection means 12. The camera-integrated recording device according to claim 11, comprising motion vector detection means for obtaining motion vector information from the minimum difference sum data. 14. The camera-integrated recording apparatus according to claim 13, wherein during the search, the search is sequentially performed from the center of the search area toward the outside.