JPH04220875A - Encoding device for picture data - Google Patents

Abstract

PURPOSE:To encode a data so as to be within a prescribed code quantity in a shorter processing time with a simple circuit by decreasing a loss of picture quality as small as possible. CONSTITUTION:A total code quantity from a code quantity calculation circuit 32 and a setting code quantity are compared and when the total code quantity is larger than the setting code quantity, a code allocation circuit 38 uses a code generated in each block to decide the allocated code quantity for each block and a code limit circuit 34 is featured to limit a variable length code so that the encoded data stored in advance in a memory 12 does not exceed the allocated code quantity.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a coding apparatus for highly compressing and coding image data to a predetermined capacity.

0002

2. Description of the Related Art When image signals are stored as digital data in storage devices such as memory cards, magnetic disks, or magnetic tapes, the amount of data is enormous, so many frames are used. In order to record a system image within a limited storage capacity, it is necessary to perform some kind of highly efficient compression on the obtained image signal data. Furthermore, in digital electronic still cameras, captured images are stored as digital data on storage media such as memory cards, magnetic disks, or magnetic tape devices instead of silver halide film. memory card,
The number of images that can be recorded on a magnetic disk or one roll of magnetic tape device is specified, and recording of this specified number of images must be guaranteed, and moreover, the time required for data recording and reproduction processing is limited. Must be short and constant.

Similarly, when recording moving images on a digital VTR (video tape recorder), digital moving image file, etc., a predetermined amount of data is recorded without being affected by the amount of image data per frame. Must be able to record frames. In other words, whether it is a still image or a moving image, it is necessary to be able to reliably record the required number of frames, and the time required for data recording and playback processing must be short and constant. be.

By the way, as a highly efficient image data compression method, a coding method that combines orthogonal transform coding and variable length coding is widely known. A typical example is a method being considered in the international standardization of still image coding.

[0005] This method will be briefly explained below. First, image data is divided into blocks of a predetermined size, and each divided block is orthogonally transformed to create a two-dimensional D
Perform CT (discrete cosine transform). Next, linear quantization is performed according to each frequency component, and Huffman encoding is performed as variable length encoding on the quantized value. At this time,
As for the DC component, the difference value between it and the DC component of neighboring blocks is Huffman encoded. The AC component is scanned from a low frequency component to a high frequency component, which is called a zigzag scan, and two-dimensional Huffman encoding is performed from the consecutive number of invalid (value 0) components and the value of the subsequent valid component. .

The above operation will be explained with reference to FIG. First, in (a), one frame of image data (one frame of image exemplified in the international standardization proposal is 720×
576 pixels) into a block of a predetermined size (for example, 8×
Divide into blocks A, B, C, ...) each consisting of 8 pixels,
As shown in (b), a two-dimensional DCT (discrete cosine transform) is performed as an orthogonal transform for each divided block, and the resultant blocks are sequentially stored on an 8×8 matrix. When image data is viewed on a two-dimensional plane, it has a spatial frequency that is frequency information based on the distribution of gradation information.

Therefore, by performing the above DCT, the image data is converted into a direct current component DC and an alternating current component DC, and the origin position, that is, (0
, 0), data indicating the value of the DC component DC, and position (0, 7), data indicating the maximum frequency value of the AC component AC in the horizontal axis direction, (7, 0). At the position, there is data indicating the maximum frequency value of the alternating current component AC in the vertical axis direction.
(7, 7) positions respectively store data indicating the maximum frequency value of the alternating current component AC in the diagonal direction, and at the intermediate position, frequency data in the direction related by each coordinate position is stored from the origin side. The data will be stored in such a way that those with higher frequencies appear in sequence.

Next, by dividing the data stored at each coordinate position in this matrix by the quantization width for each frequency component, linear quantization is performed according to each frequency component (c), and this quantization Huffman encoding is performed on the obtained value as variable length encoding. At this time, the direct current component DC
Regarding, the difference value with the DC component of the neighboring block is expressed by a group number (number of additional bits) and additional bits, the group number is Huffman encoded, and the obtained code word and additional bits are combined and Let it be encoded data (d1, d2,
e1, e2).

Valid for the alternating current component AC, ie:
A coefficient whose value is not 0 is expressed by a group number and additional bits. Therefore, the AC component AC performs a scan called zigzag scan from low frequency components to high frequency components, and detects the number of consecutive invalid components (run number of 0) and the following valid components. Two-dimensional Huffman encoding is performed using the value group number, and the resulting code word and additional bits are combined to form encoded data.

[0010] Huffman encoding is based on the peak frequency of occurrence in the data distribution of each of the DC component DC and AC component AC per frame image. This is done by encoding data and obtaining a code word by assigning bits such that the number of bits decreases and the number of bits increases toward the periphery. The above is the basic part of this method.

[0011] However, in this basic part alone, the amount of code is not constant for each image because Huffman coding, which is variable length coding, is used. Therefore, as a method for controlling the amount of code, the present inventors proposed the following method in Japanese Patent Application No. 2-137222. That is, in a compression method that combines orthogonal transformation and variable length coding, in order to control the amount of generated code, the image signal stored in memory is divided into blocks, and orthogonal transformation is performed for each divided block. Then, this converted output is quantized using a provisional quantization width, and this quantized output is variable-length coded, and the amount of generated codes for each block and the total amount of generated codes for the entire image are calculated. , Next, a new quantization width is predicted from the provisional quantization width, the total generated code amount, and the target total code amount. Further, the amount of generated codes for each block, the total amount of generated codes,
The allocated code amount for each block is calculated from the target total code amount. Then, the image signal in the image memory is divided into blocks, orthogonally transformed, quantized, and variable length encoded again using the new quantization width, and if the generated code amount of each block exceeds the allocated code amount of each block, stops variable-length encoding midway through and moves on to processing the next block. Thereby, the code amount is controlled so that the total generated code amount for the entire image does not exceed the target set code amount.

0012

As described above, in digital electronic still cameras and other applications, it is desirable to be able to compress image data with high efficiency. The above-mentioned international standard method is a compression method that satisfies these requirements.In this method, the method that combines block-by-block orthogonal transformation and variable-length coding, as shown in the example, cannot achieve high compression of image data. Although this can be done efficiently, since variable length encoding is used, the amount of code is not constant, and the number of images that can be recorded on a recording medium such as a single memory card, magnetic disk, or single magnetic tape is limited. There was a drawback that the number of sheets was indefinite.

In recording image data, it is necessary to guarantee the number of images that can be recorded on one memory card, magnetic disk device, or magnetic tape. Furthermore, from the viewpoint of operability and power consumption, the processing time for data compression must be as short as possible.

[0014] In contrast, the above-mentioned patent application No. 2-13722
The method proposed in No. 2 can obtain very good results by keeping the amount of code constant, but it requires memory to store the image signal, which increases the circuit size, cost, and power consumption. Become. For this reason, it has been difficult to apply the above method to products for which miniaturization is of great importance or to low-priced products. In recent years, there have been various needs of users, such as those who place importance on image quality even if the product is expensive, and those who want an inexpensive product even if they sacrifice some image quality.To meet these needs, the above-mentioned We need not only a method that can provide high image quality even though it requires a somewhat complicated circuit, such as the method shown in , but also a method that can be realized with a simple circuit even if the image quality is sacrificed to some extent. Furthermore, these two systems must be compatible with each other when reproducing compressed data.

The image data encoding device of the present invention has been developed with attention to such problems, and its purpose is to minimize the loss of image quality and to perform short processing using a simple circuit. An object of the present invention is to provide an image data encoding device capable of encoding data within a certain amount of code within a certain amount of time.

[0016]

[Means for Solving the Problems] In order to solve the above problems, the image data encoding device of the present invention includes a blocking means for dividing image data into blocks, and a blocking means for dividing image data into blocks. an orthogonal transform means that performs orthogonal transform on the transform, a quantizer that quantizes the transform output, a variable length encoder that encodes the quantized output with variable length, and a means for storing code amount; code amount allocation means for determining the allocated code amount for each block using the generated code amount for each block; means for comparing with a set code amount, and if the total code amount generated in the variable length encoding is larger than the set code amount, the allocated code amount for each block is determined using the generated code amount for each block. The apparatus further comprises means for determining the variable length code and terminating the variable length code so that the encoded data stored in the memory does not exceed the allocated allocated code amount.

[0017]

[Operation] That is, in the present invention, data is stored in the recording medium.
When recording data, set the desired amount of code, first perform orthogonal transform for each block, then quantize the transform output, and then variable-length encode the quantized output to create the encoded data. - record the data in memory. The total code amount generated at this point is compared with the set code amount, and if the total code amount is within the set code amount, the encoding is terminated. On the other hand, if the total code amount exceeds the set code amount, the code amount allocated for each block is determined using the code amount for each block generated in the previous encoding, and the amount of code allocated to each block is determined and recorded in the memory. This method controls the total code amount to be within the set code amount by adding processing to terminate the variable length code in each block so that the coded data does not exceed the allocated code amount. be.

[0018]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An image encoding apparatus according to the present invention will be described below with reference to the drawings.

FIG. 2 shows the encoding circuit 10 of the present invention shown in FIG.
1 is a block diagram showing the configuration of a digital electronic still camera body (hereinafter abbreviated as electronic camera body) to which this is applied. Here, illustrations and explanations of mechanisms of the electronic camera body that are not directly related to the present invention will be omitted. In FIG. 2, an electronic camera body 1 includes a lens 2, an image sensor 4 such as a CCD, a process circuit 6, an A/D converter 8, an encoding circuit 10, and a memory 1.
2, a recording medium 14 and a system controller 16.

[0020] The lens 2 is arranged in front of the image sensor 4,
A subject image is projected onto an image sensor. The video signal output by the image sensor 4 is subjected to processing such as color separation, white balance, and gamma correction by the process circuit 6, and is converted into a digital signal by the A/D converter 8. Encoding circuit 1
0 is compressed data that performs compression processing on the input image data.
output data to memory 12. The memory 12 is, for example, a FIF.
It can be configured with O (first-in first-out memory). The data that has been compressed is stored on recording medium 1.
Recorded in 4. The system controller 16 controls each of the above-mentioned components.

As shown in FIG. 1, the encoding circuit 10 includes a blocking circuit 20, an orthogonal transform circuit 22, a quantization circuit 24,
It includes an entropy encoding circuit 26, a code output circuit 28, a quantization width setting circuit 30, a code amount calculation circuit 32, a code truncation circuit 34, a control circuit 36, and a code allocation circuit 38. The blocking circuit 20 performs blocking processing to divide input image data into blocks of a predetermined size. Here, as an example, the block size is 8×
8, but is not limited to this.

[0022] Such a blocking circuit is constructed by, for example, having a RAM (random access memory) for eight horizontal lines with a capacity corresponding to the luminance and color difference components. The orthogonal transform circuit 22 orthogonally transforms the block-processed image data for each block.
This is a CT scan.

The quantization circuit 24 linearly quantizes the orthogonal transform coefficients output from the orthogonal transform circuit 22 using a quantization width for each frequency component that is preset for each frequency component. The entropy encoding circuit 26 entropy encodes the quantized transform coefficients. Here, an example using Huffman encoding is shown. The code output circuit 28 performs control to output the encoded image data to the memory 12 with a constant bus width, for example, 8 bits.

The quantization width setting circuit 30 outputs the quantization width for the set target code amount to the quantization circuit 24. The code amount calculation circuit 32 calculates the code amount of the encoded image data and outputs it to the code allocation circuit 38 and the control circuit 36.

The code terminating circuit 34 performs the following on the encoded data:
The variable length code is truncated in each block so that the allocated code amount is not exceeded. The code allocation circuit 38 uses the code amount given from the code amount calculation circuit 32 to determine the allocated code amount for each block.

The control circuit 36 is the system controller 16
The target code amount information given from the code amount calculation circuit 32
The code amount input from the encoder 10 is used to control the progress of the encoding process and to control each component of the encoding circuit 10.

The operation of the electronic camera configured as above will be explained. Prior to photographing an image, target code amount information is given from the system controller 16 to the control circuit 36. Here, the target code amount may be fixed to a value specific to the camera, or a desired value may be selected each time using a changeover switch.

Next, when a shutter (not shown) is pressed to start photographing, the image of the subject projected by the lens 2 is converted into a video signal by the image sensor 4 and output to the process circuit 6. The process circuit 6 performs processing such as color separation, white balance, and gamma correction on this video signal. The processed video signal is converted into a digital signal by the A/D converter 8, and is input to the encoding circuit 10 as a luminance and color difference line sequential signal.

In the encoding circuit 10, the input image data is first subjected to blocking processing in which the input image data is divided into 8×8 blocks by the blocking circuit 20. Here, the processing order is such that first, two blocks of luminance signals are output, then one block of color difference signals is output, and so on, using a so-called interleaving method. Luminance and color difference signals are processed alternately for each block.

The quantization circuit 24 linearly quantizes the DCT coefficients obtained as a result of the blocking process for each frequency component using a quantization width for each frequency component. The quantization width here is determined by the system control first.
The quantization width set by the quantization width setting circuit 30 is used in accordance with the target code amount information given from the controller 16 to the control circuit 36.

The quantized transform coefficients are Huffman encoded by the entropy encoding circuit 26. At this time,
Regarding the DC component DC, the difference value with the DC component of neighboring blocks is expressed by a group number (number of additional bits) and additional bits, the group number is Huffman encoded, and the obtained code word and additional bits are Together, they are used as encoded data. The alternating current component AC performs a scan called zigzag scan from low frequency components to high frequency components, and is invalidated (
Two-dimensional Huffman encoding is performed from the number of consecutive components (value is 0) (run number of 0) and the group number of the subsequent valid component values (value is not 0). The code word and additional bits are combined to form encoded data.

The encoded image data is sent to the encoder output circuit 2.
The data is divided into 8 bits by 8 and output to the memory 12. At the same time, the code amount calculation circuit 32 calculates the code amount of the encoded image data and outputs it to the code allocation circuit 38. The code allocation circuit 38 stores the amount of codes generated for each block.

The above processing is performed sequentially for each block. When the processing of blocks for the entire screen is completed, the total code amount generated is output from the code amount calculation circuit 32 to the control circuit 36 and compared with the set code amount. If the total code amount is within the set code amount, the control circuit 36 will terminate the encoding and the encoded data recorded in the memory 12 will be deleted.
The data is transferred and recorded on the recording medium 14, and the photographing ends. On the other hand, if the total code amount exceeds the set code amount, the following process is performed.

First, the total code amount generated in the previous encoding and the code amount for each block are given from the code amount calculation circuit 32 to the code allocation circuit 38, and the code amount for each block is divided into the target code amount and the total code amount. The allocated code amount for each block is determined by multiplying by the ratio. Next, the encoded data is sequentially read out from the memory 12 to the code abort circuit 34, and processing is performed to abort the variable length codes for the amount of code that exceeds the amount of code allocated to each block. That is, among the two-dimensional Huffman codes expressing the AC component AC, components within the allocated code amount are left,
High frequency component code words and additional bit data exceeding the allocated code amount are truncated and an EOB (end of block) code is added. The processed image data is output from the abort circuit 34 to the code output circuit 28, where it is
The data is divided into 8 bits and output to the memory 12. When processing of blocks for the entire screen is completed, memory 1
The encoded data stored in 2 is transferred to and recorded on the recording medium 14, and the photographing is completed.

As explained above, in the present invention,
The code amount can be controlled to stay within a certain code amount, and no memory is required to store the image signal.
It is sufficient to have a memory with a capacity sufficient to store encoded data after being compressed. Therefore, there is an advantage that the scale and cost of the circuit can be reduced.

In the above embodiment, the quantization width when linear quantization is performed may be set to exactly the target code amount for an image with many high frequency components among general images. That is, for example, images with many high-frequency components, such as landscape photographs, tend to have a large amount of code, so by adjusting the amount of code exactly to the target amount for such images, the first encoding process can be It is possible to reduce the occurrence of the generated code amount exceeding the target code amount. In this case, for example, in images with few high-frequency components, such as portraits, the amount of code is often much lower than the target amount of code, but even if such images are encoded with a small amount of code, the image quality may deteriorate. Since it is small, there is no problem in practical use.

Further, the capacity of the memory 12 may be set to allow for a slight margin for the target code amount, for example, from 1.5 times to twice the target code amount. If this value is set, encoding can be performed without problems in most cases, but in rare cases, the amount of code generated in the first encoding process may exceed the capacity of the memory 12. . To solve this problem, set the upper limit of the amount of generated code in advance during encoding, for example, from 1/3 to 1/2 of the total number of blocks, and then perform the encoding up to this point. If the amount of generated codes exceeds the upper limit at that point, the variable-length codes may be truncated for each block for the code data generated up to that point. That is, when the amount of generated code exceeds the upper limit, the control circuit 36 temporarily stops the encoding, and the encoded data is sequentially read out from the memory 12 to the code abort circuit 34, and the encoded data is read out for each block. Next, processing is performed to terminate the variable length code for the amount exceeding the predetermined upper limit code amount.

At the same time, a temporary stop signal is output from the control circuit 36 to the system controller 16, and the image sensor 4
reading is suspended. When the variable length code abort process is completed, reading and encoding from the image sensor 4 are restarted, but the above process has created enough free space in the memory 12 for subsequent encoding. , it is possible to continue encoding and complete data processing for one screen.

Furthermore, in the above-mentioned embodiment, it was assumed that all encoded data was once recorded in the memory 12, but in this case, the recording medium 14 itself is used as a memory, and there is a margin for the target code amount. It is also possible to provide additional memory outside the recording medium with only the capacity that allows for this. In this case, the overall memory capacity can be further reduced.

[0040]

[Effects of the Invention] As detailed above, according to the present invention, data can be encoded within a certain amount of code within a short processing time using a simple circuit, with the loss of image quality as small as possible. A possible image data encoding device can be provided.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing a circuit configuration of an image data encoding device according to the present invention.

FIG. 2 is a block diagram showing the configuration of an electronic camera including the encoding device of FIG. 1.

FIG. 3 is an operation transition diagram for explaining the principle of a conventional encoding method.

[Explanation of symbols]

20... Blocking circuit, 22... Orthogonal transform circuit, 24... Quantization circuit, 26... Entropy encoding circuit, 28... Code output circuit, 30... Quantization width setting circuit, 32... Code amount calculation circuit, 34... Code Termination circuit, 36...control circuit, 38...code assignment circuit.

Claims (1)

[Claims] Claim 1: Blocking means for dividing image data into blocks; orthogonal transformation means for performing orthogonal transformation for each divided block; quantization means for quantizing the output of this transformation; variable-length encoding means for variable-length encoding an output; means for storing the amount of code for each block generated in the variable-length encoding; and an allocation code for each block using the amount of codes generated for each block. and a means for calculating a total code amount generated in variable length encoding and comparing it with a set code amount, the total code amount generated in the variable length encoding being set. If it is larger than the code amount, determine the allocated code amount for each block using the generated code amount for each block, and ensure that the coded data stored in the memory does not exceed the allocated allocated code amount. An image data encoding device further comprising means for terminating the variable length code in such a manner.