JPH06205360A - Image pickup recording device - Google Patents

Abstract

PURPOSE:To execute picture compression processing suitable for each mode by provid ing two compression means and selecting a compression content depending on the image pickup mode at consecutive photographing and at single photographing. CONSTITUTION:Whether the mode is the single photographing mode or the consecutive photographing mode is discriminated based on a consecutive shot discrimination signal and in the case of the single shot mode, an input picture signal is inputted to a coder 14 via a DC transformation device 4 and a coefficient quantization device 5, in which in-frame DCT compression is executed. Furthermore, in the case of the consecutive shot mode, the same processing is executed as the single shot mode for a 1st picture and an output of the quantization device 5 is inversely converted and the result is inputted to a frame memory 19. When 2nd and succeeding pictures are inputted, the presence of a motion is discriminated by the comparison with the preceding picture in the unit of macro blocks. When no motion is in existence, a motion compensation discrimination device 17 executes inter-frame difference DCT compression and when a motion is large, in-frame DCT compression is executed as no correlation and when a motion is small, its motion vector is searched and inter-frame DCT compression is executed by motion compensation.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image capturing / recording apparatus, and more particularly to an image capturing / recording apparatus for digitally recording a still image of a subject that has been single-shot or continuously shot.

[0002]

2. Description of the Related Art There is, for example, a digital still camera as an image recording device for recording a still image of a single-shot or continuous-shot subject as a digital signal in a memory card such as an IC card. FIG. 3 shows a functional block diagram of such a conventional image pickup recording apparatus. In FIG.
The lens 38 is a so-called optical lens, and this lens 3
The subject (not shown) imaged at 8 is photoelectrically converted by the image pickup device 39 and taken in as an electric signal.

The analog image information captured as an electric signal is converted into a digital signal by the A / D converter 40. The digital signal captured in this way is
The frame controller 46 is temporarily operated by the memory controller 41.
And is fetched for one frame. This captured image signal is converted into a luminance signal and a color difference signal by the signal processing circuit 42.

Such an image pickup recording device capable of recording a still image on the memory card 45 usually has an output terminal 48 for outputting a video signal of a subject, and an LCD display in which the video signal from the terminal 48 is incorporated, for example. Or displayed on an external display device. One of the image signals sent to the output terminal 48 has the color difference signal modulated into a chroma signal by the signal processing circuit 42 and added to the luminance signal to be added to the D / A converter 47.
Is output to. Then, it is converted into an NTSC analog signal and output from the output terminal 48.

The other image signal recorded in the IC memory card is input to the compression encoding processing block 43 and image compression is performed. The compression-coded image signal is stored in the IC memory card 45 through the interface circuit 44.

FIG. 2 is a functional block diagram showing details of the conventional compression encoding circuit 43. In FIG. 2, the luminance signal and the color difference signal processed by the signal processing circuit 42 are input from the input terminal 25, respectively subjected to the processing described below, and output from the output terminal 34. This will be explained using an example.

The digitized luminance signal input from the input terminal 25 is temporarily stored in the frame memory 26. Then, from this frame memory 26, horizontal 8 pixels x
Processing is performed in block units as block data of a total of 64 pixels of 8 pixels in the vertical direction. The retrieved block is
By the two-dimensional discrete cosine transformer (hereinafter DCT) 27,
The cosine transform is applied to each of the horizontal direction and the vertical direction, and as a result, an 8 × 8 DCT coefficient matrix, that is, 2
Dimensional spatial frequency is converted.

In the case of a natural image, the converted coefficients are obtained by utilizing the fact that the converted coefficients in the horizontal direction and the vertical direction are large in the low frequency range and are small in the high frequency range, and the human visual characteristics are utilized. The compression rate is increased. Therefore, weighting is performed on each coefficient, and processing is performed so that more coefficients are "0". Specifically, the coefficient quantizer 28 performs vector operation of the 8 × 8 weighted data extracted from the set quantization coefficient table 49 and the spatial frequency coefficient output from the DCT 27,
The coefficient is scaled and output. The quantized coefficient thus obtained is subjected to an encoding process in terms of an AC component (hereinafter referred to as an AC component) and a DC component (hereinafter referred to as a DC component) due to the nature of the signal.

The AC component is a zigzag scan converter 29.
As a result, the Matrices are scanned in a zigzag order from the lowest spatial frequency component. By this zigzag scanning, the quantization coefficient of the spatial frequency is converted by the coefficient quantizer 28 so that “0” s are continuously arranged because the data of the higher frequency has “0” s. The bit stream converted in this way is used by the zero run length circuit 3
At 0, "0" is packed and the compression rate is further increased.
The "0" -compressed data bit stream is further enhanced by the Huffman encoder 31 by allocating the data having the highest appearance probability to the one having the shortest code length based on the data appearance probability.

The DC component differs from the AC component in the nature of its signal. The DC component is large as a data value due to the nature of the image signal, but if the image signal is divided into blocks of 8 × 8 pixels for compression encoding, adjacent blocks have a very high correlation. Therefore, by transmitting the data difference between the blocks, the compression rate can be further increased, and the difference is transmitted by the DPCM circuit 35. The Huffman encoder 36 assigns a code length to the DC component thus converted into the DPCM in accordance with the appearance probability in the same manner as the AC component.

In this way, the respective outputs of the Huffman encoders 31 and 36 are input to the multiplexer 32, multiplexed as serial data and input to the buffer 33. Since the bit stream data obtained in this way is variable length coded, its data amount changes depending on the image signal. Therefore, when the amount of data accumulated in the buffer is large, the amount of data is controlled to be constant by applying feedback such as increasing the quantization step of the coefficient quantizer 28. . The image signal thus obtained from the image pickup device 34 is compressed and stored in the storage medium.

[0012]

However, in the compression method of the image recording apparatus according to the prior art as described above, the compression is performed without distinguishing between the single shooting mode and the continuous shooting mode in which a plurality of still images are shot per second. is doing. For this reason, when a subject is photographed in the continuous shooting mode, the capacity of the storage medium is limited, and the photographing opportunity may be limited to once, which is very inconvenient for the user.

The present invention solves the above-mentioned drawbacks of the prior art, and provides an image pickup recording apparatus capable of switching the compression contents depending on the shooting mode for continuous shooting and single shooting and performing image compression processing suitable for each. To aim.

[0014]

In order to solve the above problems, the present invention is directed to a still image capturing / recording apparatus that compresses and stores a still image of a subject captured in either a single shooting mode or a continuous shooting mode. A first compression means for compressing an image signal of an image in a frame or a field; and a second compression means for compressing an image signal of a still image by at least a difference between frames or fields and motion application compensation; In the shooting mode, the first compression means compresses the image signal, and in the continuous shooting mode, the switching means performs a switching process of compressing the image signal by the second compression means. The compression means detects the motion vector between the first image signal and the second image signal input thereafter, and if the motion vector is within a predetermined range, Or by the inter-field difference and motion applying compensation for compressing said first image signal.

[0015]

According to the present invention, the compression in the frame or field is performed in the single-shot mode. In the continuous shooting mode, the motion vector is searched by the correlation between the frames of the image signal, and the frame or the inter-field difference or the frame using the motion compensation is selectively used according to the magnitude of the motion vector. Alternatively, compression is performed by the difference between fields.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT An embodiment of an image recording apparatus according to the present invention will be described in detail with reference to the accompanying drawings.

A digital still camera, for example, is known as an image recording device for recording a still image of a single-shot or continuous-shot subject as a digital signal in a memory card such as an IC card. FIG. 5 is a functional block diagram showing an embodiment of such an image recording apparatus according to the present invention.

In FIG. 5, a lens 50 is a so-called optical lens, and a subject (not shown) formed by the lens 50 is photoelectrically converted by an image pickup device 51 such as CCD and taken in as an electric signal. The analog image information captured as an electric signal is converted into a digital signal by the A / D converter 52. The digital image signal thus captured is converted by the signal processing circuit 42 into a luminance signal and a color difference signal.

Such an image-capturing recording device capable of recording a still image on the memory card 56 usually has an output terminal 58 for outputting a video signal of a subject, and the video signal from the terminal 58 is incorporated, for example, in an LCD display or. It is designed to be displayed on an external display device. For this reason,
One of the image signals sent to the output terminal 58 is a color difference signal modulated into a chroma signal, added to a luminance signal, output to the D / A converter 57, converted to an NTSC analog signal, and output from the output terminal 58. It The other image signal recorded in the IC memory card 56 is input to the compression encoding processing block 54 and image compression is performed. The compression-coded image signal is transferred to the interface circuit 55.
Through the IC memory card 56.

FIG. 1 is a block diagram showing an example of the compression encoding circuit 54 of the image pickup recording apparatus in this embodiment. The compression encoding circuit 54 uses the luminance signal (Y) and the color difference signal (R
-Y, BY) is input from the digital input terminal 1, subjected to the predetermined signal processing described below, and output from the output terminal 11. The compression encoding circuit 54 will be described below. Here, the luminance signal will be described as an example.

First, in the single shooting mode, a mode identification signal indicating the single shooting mode is input from the continuous shooting mode determination signal input terminal 18 for inputting the mode identification signal. Based on this single-shot mode identification signal, the motion compensation discriminator 1
The switch 7 selects the single-shot mode and selects the switch 22 to the contact a side. In the single shooting mode, the motion compensation circuit 17 does not discriminate the original motion in this way. In this way, the image signal first obtained from the image pickup device 51 (FIG. 1) is input from the input terminal 1 and temporarily stored in the frame memory 2.

The image signal from the frame memory 2 is processed in block units as block data of a total of 64 pixels of horizontal 8 × vertical 8 pixels. In the image signal input in the single shooting mode, the switch 22 is selected on the contact a side. Therefore, the subtractor 3 becomes substantially through, and the image signal from the frame memory 2 remains as it is in the DCT converter 4.
Entered in.

Two-dimensional discrete cosine transformer (hereinafter DCT)
Reference numeral 4 is a converter for performing cosine transformation in the horizontal direction and the vertical direction, respectively, and as a result, it is transformed into 8 × 8 DCT coefficient matrices, that is, two-dimensional spatial frequencies. In the case of natural images, the fact that the converted horizontal and vertical coefficients are large in the low frequency range and small in the high frequency range is used, and
It is processed to have a higher compression rate by utilizing the human visual characteristics.

Specifically, each coefficient is weighted so that the number of coefficients becomes "0".
For this reason, the vector quantizer 5 performs vector calculation of the 8 × 8 weighted data extracted from the set quantization coefficient table 24 and the spatial frequency coefficient output from the DCT 4.
, And the coefficient is scaled and output. The quantized coefficients obtained in this way are AC components (hereinafter referred to as AC components) and DC components (hereinafter referred to as DC components).
In addition, the encoding processing is changed due to the nature of each signal.
At 4, the encoding process is performed.

The AC components are zigzag-scanned by the zigzag scan converter 6 in the order of increasing spatial frequency components. The quantized coefficient of the spatial frequency by this zigzag scan is converted by the coefficient quantizer 5 so that "0" s are arranged more continuously because the data of the higher frequency has "0" s. In this way, the converted bit stream is packed by "0" in the zero run length circuit 7 and the compression rate is further increased. The data bit stream thus "0" -compressed is further compressed by the Huffman encoder 8 by allocating the one having the highest appearance probability to the one having the shortest code length based on the data appearance probability. You can increase the rate.

The DC component differs from the AC component in the nature of its signal. That is, the DC component is large as a data value due to the nature of the image signal, but the image signal is 8
When divided in block units of × 8 pixels, the correlation between adjacent blocks becomes very high. Therefore, by transmitting the data difference between the blocks, it is possible to further increase the compression rate, and thus the DPCM circuit 12 transmits the difference. The Huffman encoder 13 assigns a code length to the DC component thus converted into the DPCM in accordance with the appearance probability in the same manner as the AC component.

In this way, the respective outputs of the Huffman encoders 8 and 13 are input to the multiplexer 9 and multiplexed as serial data and input to the buffer 10. Then, the bit stream data obtained by this is variable-length coded, so that the data amount thereof changes depending on the image signal. Therefore, the buffer 10
Depending on the amount of data accumulated in the above, if the amount is large, feedback is performed such as increasing the quantization step of the coefficient quantizer 5, so that the data amount is controlled to be constant. The image signal thus obtained from the image pickup device 51 is compressed and output from the output terminal 11, and is stored in the IC memory card 56 which is a storage medium via the interface circuit 55. In this way, the single-shot mode is the same as the conventional compression process.

Next, the continuous shooting mode will be described. In the continuous shooting mode, the coefficient inverse quantizer 15, 8 * 8 IDCT 16, the motion compensation discriminator 17, the frame memory 19, the loop filter 20, the switches 21, 22, 59 and addition are performed so that the capacity of the storage medium 45 is effectively used. The compensation block 23 constituted by the device 60 functions.

During continuous shooting, multiple images per second (10 in and out)
In this mode, the captured image signals often have a very strong correlation between the captured image signals.
At the time of continuous shooting, first, the continuous shooting mode is input to the motion compensation discriminator 17 from the mode determination signal terminal 18 and is identified as the continuous shooting mode.

The first first image from the image pickup device in the continuous shooting mode is taken in in the same manner as in the single shooting mode, compressed as described above, and output from the output terminal 11. Along with this, the spatial frequency coefficient sequence output from the coefficient quantizer 5 is the coefficient inverse quantizer 1 of the compensation block 23.
5 is input, inversely quantized, and inversely transformed into an 8 × 8 spatial frequency sequence. In this way, the discrete cosine transform coefficient which is the irreversibly restored spatial frequency is irreversibly converted into an image signal by the inverse cosine converter 16, and the switch 59 is moved to the contact f side by the control signal from the motion compensation discriminator 17. Is input to the frame memory 19.

Next, when the second and subsequent image signals which have been continuously shot are input from the input terminal 1, they are temporarily stored in the frame memory 2. The difference from the processing of the first image signal is that a macroblock that is made into a large number of blocks is further defined as a set of several 8 × 8 pixel block units in order to detect a motion, and a frame is set in each macroblock unit. The presence or absence of motion is determined by calculating the difference between the image signal of the previous frame in time recorded in the memory 19 and the macro block at the same position and comparing the difference with the reference level.

When Motion is Small When the motion compensation discriminator 17 determines that the motion is small in macroblock units, the switch 22 is selected on the contact b side and the switch 21 is selected on the contact d side. In this way, the switch is selected and the frame memories 2 to 8 are selected.
The data is input to the comparison and subtractor 3 in block units of × 8 pixels.
A signal image of a block of 8 × 8 pixels at the same position is input from the other frame memory 19, and the difference between blocks at the same position is output from the comparison / subtractor 3.

In this way, the image data composed of the difference components are similarly subjected to the two-dimensional cosine transformation, and the following processing is performed in the same manner as the single shooting mode. As a result, most of the two-dimensional cosine-transformed coefficient sequence is “0”, so that the image signal data is compressed and recorded on the storage medium (for example, the IC card 56 here).

On the other hand, the coefficient-quantized data is input to the coefficient quantizer 15, inversely quantized, and input to the cosine inverse transformer 16 to be a differential image signal. At this time, the switch 59 is selected on the contact e side and the switch 21 is selected on the contact d side by the determination signal. Therefore, the image signal output from the frame memory 19 is input to the adder 60 without change from the previous image signal, is added to the difference component from the DCT 16, is input to the frame memory 19, and is secondly recorded. The resulting image signal is restored as a result and held in the frame memory 19. In this way, the image signals are successively compared with the previous image signal, and the difference component is compressed and recorded.

When the motion is large In macroblock units, when the motion is determined to be large by the motion compensation discriminator 17, the block of the frame to be processed was recorded earlier in time in order to detect the motion vector. With respect to the block 19 of the frame memory in which the image signal is held, all the blocks having the smallest difference component between blocks within a certain pixel distance are searched to obtain a motion vector. In this way, the motion vector is obtained to determine which block of the previous frame has the smallest difference component with respect to the block to be processed of the frame to be processed.

At this time, the switch 22 is selected on the contact b side, and the switch 21 is selected on the contact c side. In this way, the switch is selected, and the processed block of 8 × 8 pixels from the frame memory 2 is input to the comparison / subtractor 3 in block units. From the other frame memory 19, the block of the previous image signal determined by the motion vector is subjected to filtering processing by the loop filter 20 in order to reduce the distortion of the edge portion of the block, and is output. The component is output from the subtractor 3.

In this way, the image data composed of the difference components between the blocks are similarly subjected to the two-dimensional cosine transformation,
The following processing is performed in the same way as when the movement is small. As a result, the coefficient sequence subjected to the two-dimensional cosine transform by the DTC converter 4 is mostly "0", so that the image signal data is compressed and the storage medium (for example, I
It will be recorded on the C card 56).

On the other hand, the coefficient-quantized data is input to the coefficient quantizer 15 and inversely quantized to obtain the cosine inverse transformer 1.
It is input to 6 and becomes a differential image signal. Since the contact point of the switch 59 is selected to the e side and the contact point of the switch 21 is selected to the d side by the determination signal, the image signal output from the frame memory 19 is the same as the previous image signal input to the adder 60 as it is. , And the difference component from the IDCT 16 is added to the frame memory 19 and the image signal to be recorded second is restored as a result and held in the frame memory. In this way, the image signals are successively compared with the previous image signal, and the difference component is compressed and recorded.

When there is no correlation between images When the motion is very large, or when panning or a background appears newly, there is no correlation between the previous image and the image to be processed. A compression method using the correlation between frames as described above cannot be used. Therefore, in such a case, the image signal is recorded in the storage medium by the compression by the DCT conversion within the frame as in the single shooting mode.

FIG. 4 is a flow chart showing the operation of the compression encoding process in this embodiment. As shown in the figure, in the present embodiment, whether the continuous shooting mode or the single shooting mode is determined by the continuous shooting determination signal (S100), and in the single shooting mode, the intra-frame DCT compression is executed (S102). If it is the continuous shooting mode, the intra-frame DCT compression is executed in the case of the first frame (S102), and if not, the presence or absence of motion compensation is determined by comparing with the previous image signal.
If there is no motion compensation, the inter-frame difference DCT
Compression is performed (S108), and it is determined whether the motion between macroblocks is large (110). If the motion between macroblocks is large, it is determined that there is no correlation (S116), and intra-frame DCT compression is executed (S1).
02). Also, if the movement between macroblocks is small,
The motion vector is searched (S112), and interframe DCT compression by motion compensation is executed (S114).

As described above, according to this embodiment, in the single shooting mode, the compensation block 23 representing the entire motion application block used in the continuous shooting mode is not used, so that the block 23 is synchronized with the continuous shooting determination signal 18. By stopping the operation of, the function for continuous shooting can be added without increasing the power consumption in the single shooting mode.

[0042]

As described above, even in the limited storage medium capacity, by performing motion compensation between frames or fields and compression by frame or field differential transmission or frame or field differential transmission in the continuous shooting mode. Recording can be performed without missing a photo opportunity, and great convenience can be given to the user without increasing power consumption during single shooting in normal use.

[Brief description of drawings]

FIG. 1 is a functional block diagram showing an embodiment of a compression encoder in an image recording apparatus according to the present invention.

FIG. 2 is a functional block diagram of a compression encoder in a conventional image recording device.

FIG. 3 is a functional block diagram of a conventional image recording device.

FIG. 4 is a flowchart showing processing of the compression encoder shown in FIG.

FIG. 5 is a functional block diagram showing an embodiment of an image pickup recording apparatus according to the present invention.

[Explanation of symbols]

 1 Digital Input Terminal 2, 19 Frame Memory 3 Subtractor 4 DCT Converter 5 Coefficient Quantizer 6 Zigzag Scanner 7 Zero Run Length Converter 8 Huffman Encoder 9 Multiplexer 10 Buffer 11 Output Terminal 12 DPCM Unit 13 Huffman Encoder 14 Encoder 15 Coefficient inverse quantizer 16 Inverse DCT converter 17 Motion compensation discriminator 18 Continuous shooting mode discrimination signal input terminal 20 Loop filter 21, 22, 59 switch 23 Motion compensation block during continuous shooting 24 Quantum Conversion coefficient cable 50 Lens 51 Image sensor 52 A / D converter 53 Signal processing circuit 54 Compression encoder 55 Interface 56 IC memory card 57 D / A converter 58 Output terminal 59 Switch

Claims (1)

[Claims] 1. An imaging recording apparatus for compressing and storing a still image of a subject imaged in one of a single shooting mode and a continuous shooting mode, the first compression for compressing an image signal of the still image within a frame or a field. Means, second frame compression means for compressing the image signal of the still image by at least frame or field difference, and in the case of the single shooting mode, the first compression means A switching means for compressing the image signal and for switching the image signal by the second compression means in the case of the continuous shooting mode, wherein the second compression means includes the first image signal and the first image signal. The motion vector with respect to the second image signal input thereafter is detected, and when this motion vector is within a predetermined range, the difference between the frames or fields is detected. And the motion applying compensation, imaging and recording apparatus which is characterized in that the compression of the first image signal.