JP4092830B2 - Image data compression method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image data compression method, and is particularly useful when applied to compression / decompression of multilevel halftone image data having a large number of gradations. Image data compression method About.
[0002]
[Prior art]
With the development of image input devices (for example, CCDs used in digital cameras), image data has been captured with high definition. The original data (RAW data) obtained in such a high definition is subjected to image processing by the user with a personal computer or the like, and then desired image data is obtained with an output device (for example, CRT, printer). It is supposed to be.
[0003]
Image data represented by a signal (signal charge) from the CCD is converted into a digital signal of 10 to 12 bits per pixel by, for example, an A / D converter. When converted into a 12-bit digital signal, the maximum gradation is “4096”.
Thus, the image data captured in high definition with 10 to 12 bits is changed to image data of about 8 bits per pixel, and the image is reproduced by the output device.
[0004]
As described above, the number of bits of image data per pixel at the time of capture is represented by 10 to 12 bits. At the time of capture, the shooting environment varies widely from when the shooting environment is extremely bright to when it is extremely dark. This is because it is necessary to obtain image data. On the other hand, at the time of reproduction, the brightness difference actually represented in one frame (screen) is smaller than that in the shooting environment, so that the brightness of the image in one frame can be sufficiently expressed with about 8-bit image data.
[0005]
By the way, the image data captured by the CCD is stored in a storage medium, and the image data is appropriately read out from the storage medium. Conventionally, the image data is changed to 8 bits per pixel in the storage medium. Was remembered.
Here, if 8-bit image data per pixel is stored as it is, the amount of data per frame (one screen) becomes enormous, so that the image data is compressed and stored in a storage medium. Yes.
[0006]
As general image data compression processing, DPCM processing for original data (RAW data), Huffman coding, arithmetic coding, JPEG lossless coding using these as appropriate, universal codes represented by the Ziv-Lempel method Is known.
[0007]
[Problems to be solved by the invention]
However, if image data of 8 bits per pixel is stored as original data in the storage medium, when the user performs desired processing / correction on the image data of 8 bits per pixel, The image data becomes 8 bits or less per pixel, and the image quality deteriorates.
[0008]
Therefore, the original data (10 to 12 bits) of the high-definition image obtained by the CCD is stored in the storage medium as it is, and the original data is read out and processed / modified. By doing so, the image quality of the processed / modified image data can be improved.
When the high-definition original data is stored in the storage medium, it is necessary to perform the above-described compression processing (particularly, lossless compression processing). However, 10-bit to 12-bit image data is converted to 8-bit image data. Even if similar compression processing is performed, a high compression ratio cannot be achieved. This is because when image data is compressed (especially in the case of DPCM coding), the higher the compression rate is, the higher the correlation between pixel data between adjacent pixels is. For example, an image of about 12 bits In the case of data, the correlation of image data between pixels is remarkable in the upper 6 bits to 8 bits, but there is almost no correlation in the lower 3 to 4 bits. Therefore, even if lossless compression is applied to 10-bit to 12-bit image data as it is, a high compression rate cannot be obtained, and even after compression, a storage medium having a large image data and a large capacity is still required. become.
[0009]
Furthermore, the image data represented by 10 to 12 bits has a problem that the processing time is long to read and reproduce the image data from the storage medium because the image data amount is large.
The present invention has been made in view of such circumstances, and a first object thereof is to provide an image data compression method capable of compressing image data acquired with high definition at a high compression rate.
[0010]
The second purpose is image data stored in a storage medium. The When reproducing, a coarse image with a small amount of data can be selectively reproduced at high speed. Image data compression method Is to provide.
[0011]
[Means for Solving the Problems]
In order to achieve the above object, according to the first aspect of the present invention, the image data acquired by the image input means (1) at a constant number of bits per pixel, the data for each pixel of the constant number of bits is transferred to the upper side. Separating the higher-order bit data represented by a predetermined number of bits and the lower-order bit data represented by a lower-order predetermined number of bits; performing variable length encoding on the higher-order bit data; and Packing side bit data by a bit shift operation addressed to a plurality of pixels, and generating management data for individually managing the variable-length-encoded higher-order bit data and the packed lower-order bit data A step, a variable length encoded upper bit data, the packed lower bit data, and the generated management data And storing (2) The image input means (1) is an input means in which pixels are arranged in a matrix, and the variable-length-encoded higher-order bit data is at least for each pixel included in the same row of one row. Blocked and packed lower side bit data is blocked for each pixel included in the same row of at least one row, and as the management data, the upper side bit data and the blocked upper side bit data are blocked. Position information indicating each storage position of the lower-order bit data is generated, and the compressed image data obtained by the variable-length encoding is stored in the management data in the block-order higher-order bit data. Position information indicating the position is provided at a position ahead of the variable-length encoded higher-order bit data. It is what I did.
[0012]
According to a second aspect of the present invention, in the image data compression method according to the first aspect, The variable-length encoded high-order bit data is blocked for each pixel included in all the rows of the image data for one frame, and the packed low-order bit data is the image data for one frame. Blocked for every pixel in every row It is what I did.
[0013]
According to a third aspect of the present invention, in the image data compression method according to the first or second aspect, the variable length coding is performed by Huffman coding and DPCM coding. .
According to a fourth aspect of the present invention, there is provided the image data compression method according to any one of the first to third aspects, wherein a certain number of bits of the data for each pixel is detected by the detection accuracy of the image input means (1). The predetermined number of bits on the upper side is set to the number of upper bits on which there is a strong tendency to generate a correlation with neighboring pixels, and the value is obtained based on empirical rules.
[0014]
The invention according to claim 5 is: 5. The image data compression method according to claim 1, wherein the fixed number of bits per pixel acquired by the image input means (1) is 12 bits or more. It is what I did.
[0015]
(Function)
Book According to the invention, for high-order image data obtained with high definition, high-order bit data having high correlation with neighboring pixels can be variable-length encoded with high compression, and low-order bits that have low correlation. For data, bit shift and packing processing can be performed at high speed. Further, by storing the management information, desired image data can be accessed at high speed, and the restoration process can be speeded up.
[0016]
or, Book According to the invention, the higher-order bit data and the lower-order bit data are individually blocked for each row, so that the processing speed when selectively reproducing only the image data of a specific row can be increased. In this case, the image can be reproduced at high speed by the position information indicating the storage position of each block.
or, Book According to the invention, it is possible to encode image data at a high compression rate with respect to high-order bit data having strong correlation.
[0017]
or, Book According to the invention, high-definition image data corresponding to the capture accuracy of the image input means (1) can be compressed / restored.
or, Book According to the invention, the high-definition image data obtained by the image input means (1) is separated into upper bit data and lower bit data and stored in the storage medium (2). By selectively restoring only the data, a rough image can be reproduced at high speed.
[0018]
DETAILED DESCRIPTION OF THE INVENTION
(First embodiment)
DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, a first embodiment of the invention will be described with reference to the accompanying drawings. The first embodiment corresponds to claims 1 to 5.
200801181359070460______________________11999328380_____APH_0
Of the present invention Image data compression method 1 is a block diagram illustrating a configuration of an encoding processing device 10 to which is applied.
[0019]
The encoding processing apparatus 10 digitizes the image signal input from the image input device (CCD) 1 and then compresses the image data by JPEG lossless encoding in which DPCM encoding and Huffman encoding are used in combination. And stored in the storage medium 2. Here, the encoding processing device 10 is provided integrally with the CCD 1 (for example, in a digital camera). Further, the compressed image data stored in the storage medium 2 is decoded by a decoding processing device 20 (FIG. 6) such as a personal computer, and then processed / modified by the user, as will be described later. The image is restored.
[0020]
Here, the CCD 1 connected to the input side of the encoding processing device 10 has pixels (cells) arranged in a matrix of m rows × n columns, and as shown in FIG. In the illustrated example, color filters “R”, “G”, and “B”) are arranged.
The encoding processing apparatus 10 includes an A / D converter 11 that converts a signal (signal charge) from the CCD 1 into a digital signal (12 bits) of 4096 gradations, and temporarily converts the digitized signal (original data). The input data buffer 12 to be stored, the CPU 13 that encodes the signal (original data) temporarily stored in the input data buffer 12, the main memory 14 that stores the program executed by the CPU 13, and the signal encoded by the CPU 13 (Compressed image data) is temporarily stored, output to the storage medium 2 side at a predetermined timing, and constituted by an output data buffer 15 and the like.
[0021]
Among them, the input data buffer 12 stores pixel data (m × n) for each pixel obtained by the CCD 1 for each row (line), and data for one row (n The target line buffer 12A for storing a plurality of pixel data), the one line previous buffer 12B for storing the previous line data shifted from the target line buffer 12A, and the one line previous buffer 12B. A two-line-before buffer 12C for storing the data two lines before is configured.
[0022]
The output data buffer 15 includes an address buffer 15A, a high buffer 15B, and a row buffer 15C. Among these, the address buffer 15A includes “higher-order compressed data block position information” described later, and the high buffer 15B includes “higher-order” “Side compressed data” and “lower side data block position information”, and “lower side data” are stored in the row buffer 15C, respectively.
[0023]
Next, an image data encoding process for one frame (screen) executed by the CPU 13 of the encoding processing apparatus 10 will be described with reference to a program flowchart of FIG.
When this encoding process is started (start), first, in step S1, from the pixel data (12-bit original data) stored in the target line buffer 12A of the input data buffer 12 for every n rows. , The data of the pixel of interest is read.
[0024]
In step S2, the pixel data (12 bits) of the read target pixel is separated into upper 8 bits (upper bit data) and lower 4 bits (lower bit data). In this way, the 12-bit pixel data (original data) is separated into the upper 8 bits and the lower 4 bits. Generally, in the image data obtained by 12 bits, the upper 6 bits to the upper 8 bits are related. Is highly correlated with other neighboring image data and can be encoded at a high compression rate, whereas when it is in the lower 4 bits, it is between the lower 4 bits of other neighboring pixels. This is because the correlation at the time becomes low.
[0025]
In step S3, 8-bit higher-order bit data is encoded by JPEG lossless encoding, and the encoded pixel data is stored in the high buffer 15B as “upper-side compressed data”.
In this first embodiment, the JPEG lossless encoding in step S3 is generally performed according to the following procedure by DPCM encoding and Huffman encoding.
[0026]
First, a predicted value of a target pixel is determined based on a pixel value (high-order bit data value) of a neighboring pixel (same color pixel) of the same color filter or a neighboring pixel (high-order bit data value). Calculated according to the prediction formula.
Here, as the neighboring pixels, the same color pixels (R42 two pixels before), the same color pixels two lines before (R24, R22), or adjacent pixels with respect to the target pixel (for example, R44 in FIG. 2) Among (G43, B33, G34), a pixel that can obtain a pixel value with the smallest prediction error is used.
[0027]
Which pixel value is used to calculate the predicted value (which prediction value is used as a variable) is specifically the same color pixel in the vicinity of the pixel two pixels before or the adjacent pixel two pixels before A plurality of provisional prediction values in a plurality of prediction formulas using the respective pixel values of the higher-order bit data of each of the two, and comparing the pixel values two pixels before and the plurality of provisional prediction values, respectively, A prediction formula (optimum prediction formula) that minimizes the prediction error is stored. Then, the prediction value of the target pixel in the current loop is calculated using the optimum prediction formula stored two pixels before.
[0028]
The prediction value of the higher-order bit data calculated in this way is compared with the pixel value of the higher-order bit data (8 bits) separated in step S2 to obtain the prediction error Δ (DPCM coding). . Then, the prediction error Δ is subjected to Huffman coding according to the occurrence distribution, and the upper bit data is coded (variable length coding).
[0029]
Here, in the calculation of the predicted value, a prediction formula using the same color pixel of the same row and the pixel value of the same color pixel of the previous row as the same color pixel is used. As shown in FIG. In the case of a primary color CCD, for R and B, pixels in which color filters of the same color component are arranged do not exist in one row (one line before), and neighboring pixels of the same color exist in two rows before. It is.
[0030]
When the upper bit data is encoded in step S3 as described above, in the next step S4, the lower 4 bits of the pixel of interest are adjacent pixels (pixels having different color components in the example shown in FIG. 2). It is combined with the lower bit data and processed as 8-bit data as a whole.
That is, of the pixels arranged in a matrix (FIG. 2), the pixels (n) included in the same row are odd-numbered or even-numbered from the extreme end position (B11 in the first row of FIG. 2). If it is an odd number (B11, B13, B15), the lower 4 bit data is shifted to the upper 4 bits of the row buffer 15C (8 bits) and stored, and the even number (G12, G14) is stored. ), The lower bit data is stored as it is in the lower 4 bits of the row buffer 15C. As a result, in the row buffer 15C, odd-numbered (B11, B13, B15) lower-order bit data (4 bits) is stored in the upper 4 bits of 1 byte (8 bits), and the lower 4 bits of the row buffer 15C. Even-numbered lower-order bit data (4 bits) is stored (bit shift processing and packing processing).
[0031]
In step S5, it is determined whether or not the processing in steps S1 to S4 described above has been performed for one row, that is, whether all n pixels included in the current row are processed. When the above process is completed (the determination result is “Yes”), the process proceeds to the next step S6.
In step S6, the high-order compressed data obtained for one row (n pixels) is combined as a high-order compressed data block, and similarly, the low-order data is combined as a low-order data block (blocking). .
[0032]
In step S7, the position information of the lower data block in the same row is added to the head position of the upper compressed data block.
In step S8, position information indicating the head position of the upper compressed data block in each row is added to the address buffer 15A.
In this way, the upper side compressed data block position information stored in the address buffer 15A, the lower side data block position information to which the upper side compressed data and the lower side data are added are stored in the storage medium 2. It is stored at a predetermined timing.
[0033]
When the encoding of the image data for one line, which is the subject of this time, is completed, in the next step S9, the above-described processing in steps S1 to S8 is performed for all m lines constituting one frame (one screen). It is determined whether or not it has been received. If the processing for one frame (m-line processing) has not been completed yet, the determination result in step S9 is “No”, and the above-described steps S1 to S8 are repeatedly executed.
[0034]
On the other hand, when the determination result of step S9 is “Yes”, that is, when the processing for all the rows (m rows) of one frame is finished, the program is finished as it is.
FIG. 4 shows the data format of the encoded image data (one frame) stored in the storage medium 2.
[0035]
As shown in this figure, the image data for one frame is stored with the upper compressed data block position information for one frame (for m rows) at the earliest position, and thereafter the image data block for each row is stored. It is stored in a repeated pattern from the first line to the m-th line. Each row of image data blocks is appended with one row of lower data block position information at the beginning, followed by one row of upper compressed data blocks and one row of lower data blocks. Are stored sequentially.
[0036]
Here, the higher-order compressed data block position information provided at the earliest position of the image data for one frame is the position information of the image data block in each row of the image data for one frame (for m rows) (in FIG. 4). The information on the upper side compressed data position indicated by ↓). The lower data block position information provided at the head position of the image data block in each row indicates the position information (position information indicated by ↑ in FIG. 4) of the lower data block in the image data block in each row. Is.
[0037]
Specifically, as shown in an enlarged view in FIG. 4 (the image data block in the second row is enlarged in the figure), the upper compressed data following the lower data block position information in the second row In the block, the high-order compressed data of all the pixels included in the second row (for n pixels) is blocked. Further, the lower side data (8 bits) that is packed for two pixels and represented by one byte is divided into one row (n pixels) as a lower side data block.
[0038]
As described above, the high-order compressed data block position information is stored at the earliest position of the image data for one frame, whereby the high-order compressed data of a specific row in one frame image is randomly read out. Is possible. Here, the higher-order compressed data block position indicating the position of the higher-order compressed data block simultaneously indicates the position of the image data block in each row. Therefore, it is also possible to read out both the high-order compressed data and the low-order data in a specific row at random.
[0039]
Also, as described above, the lower data block position information is added to the head position of the image data block in each row because the length of the upper compressed data block is variable by JPEG lossless encoding as described above. This is because the data length is different for each row after encoding.
FIG. 5 shows another data format example when the upper compressed data block and the lower data block are stored in the storage medium 2.
[0040]
In the data format shown in this figure, when compressing and storing image data for one frame, the upper side compressed data block position information and the lower side data block position information are stored for each row at the earliest position. The data format is different from the data format shown in FIG. Therefore, in the second data format example, random reading of higher-order compressed data of a specific row based on higher-order compressed data block position information, and higher-order side of a specific row based on higher-order compressed data block position information In addition to being able to read out both the compressed data and the lower-order data at random, it is possible to read out the lower-order data in a specific row based on the lower-order position data at random.
[0041]
Next, decoding of compressed image data encoded according to the above procedure will be described.
FIG. 6 is a block diagram showing a decoding processing device 20 for decoding the compressed image data stored in the storage medium 2.
The decoding processing device 20 decodes the compressed image data from the storage medium 2 and outputs the decoded image data (12-bit original data) to an image processing / correction device (personal computer) 30. An input data buffer 21 that reads compressed image data from the storage medium 2, a CPU 22 that restores compressed image data from the storage medium 2, and restored data that temporarily stores and outputs the restored image data (original data) for each row. The buffer 23 and the main memory 24 in which programs executed by the CPU 22 are stored.
[0042]
In addition, an image processing / correcting device 30 is connected to the restored data buffer 23 of the decoding processing device 20. This image processing / correction device 30 processes / corrects image data (12-bit original data) restored based on externally input processing / correction information (information input by a user's keyboard operation or the like). Is to be applied. The processed / corrected image data (for example, 8-bit image data) is output to an image output device (for example, CRT, printer) 40 so that a desired image can be obtained. When the above-described decoding processing apparatus 10 is used in a digital camera, the decoding processing apparatus 20 and the image processing / correction apparatus 30 are mainly configured by a personal computer.
[0043]
Here, the input data buffer 21 of the decoding processing device 20 constitutes an address buffer 21A, a high buffer 21B, and a row buffer 21C. Then, “upper compressed data block position information” from the storage medium 2 is temporarily stored in the address buffer 21A, and “upper compressed data” and “lower data block position information” from the storage medium 2 are stored in the high buffer. 21B is temporarily stored. In addition, the “lower data” from the storage medium 2 is temporarily stored in the row buffer 21C.
[0044]
Further, the restoration data buffer 23 of the decoding processing device 20 constitutes a target line buffer 23A, a one-line front buffer 23B, and a two-line front buffer 23C. Then, the restored image data (12-bit original data) of the current restoration target line (restoration target line) is temporarily stored in the target line buffer 23C. The previous line image data (12-bit original data) shifted from the target line buffer 23A is temporarily stored in the previous line buffer 23B, and the previous line buffer 23C is stored in the previous line buffer 23C. Image data (12-bit original data) two rows before shifted from 23B is temporarily stored.
[0045]
FIG. 7 is a program flowchart showing an image decoding process executed by the CPU 22 of the decoding processing apparatus 20 described above.
When the image decoding program is started, first, in step S21, the position information (upper compressed data block position information, lower data block position information) taken into the input data buffer 21 from the storage medium 2 is stored. , Read by the CPU 22.
[0046]
In step S22, all of the image data blocks in the target row (for one row) in the current loop based on the read higher-order compressed data block position information (information indicating the position indicated by ↓ in FIG. 4). The data is read.
In step S23, based on the lower data block position information (information indicating the position of ↑ in FIG. 4) read in step S21, the data read in step S22 is converted into the upper compressed data block and the lower data block. The separation process is performed.
[0047]
In step S24, the well-known JPEG lossless decoding is performed on the high-order compressed data stored in the separated high-order compressed data block. This JPEG lossless decoding is performed in the reverse order of the encoding procedure performed in step S3 of the above-described image encoding process (FIG. 3).
That is, when decoding the higher-order bit data (8 bits) of the target pixel, first, the pixel value of two pixels before (same color pixel) decoded two loops before is obtained. Next, a plurality of provisional prediction values of the same color pixel two pixels before are obtained based on a plurality of prediction expressions using the pixel of the same color pixel in the vicinity or the pixel value of the adjacent pixel. Then, the pixel value two pixels before and the provisional prediction value are respectively compared to determine the prediction error Δ, and the prediction formula (optimal prediction formula) that minimizes the prediction error is determined. Finally, using this optimal prediction formula, the predicted value of the target pixel in the current loop is obtained, and the upper bit data (8 bits) of the target pixel is decoded from the stored prediction error Δ and this predicted value. .
[0048]
In step S25, the lower-order data (two pixels are represented by 8 bits) packed in step S4 of the above-described image encoding process (FIG. 3), and the higher-order 4 bits are odd-numbered image data. As (lower bit data), the lower 4 bits are restored as even-numbered image data (lower bit data).
[0049]
In step S26, 12-bit image data (original data) is obtained from the upper bit data (8 bits) restored in step S24 and the lower bit data (4 bits) restored in step S25. .
In step S27, it is determined whether or not the processing in steps S24 to S26 described above has been performed for one row, that is, all the n pixels included in the same row, and when the decoding processing for one row is completed. The process proceeds to the next step S28.
[0050]
In step S28, it is determined whether or not the processes in steps S22 to S27 described above have been performed for all the lines (m lines) constituting one frame (one screen), and the processing for one frame has been completed. If not, the process returns to step S22 and the process is repeated. When the processing for all the lines (m lines) of one frame is completed (the determination result of step S28 is “Yes”), the program is ended as it is (end).
[0051]
As described above, according to the image data encoding method and decoding method of the first embodiment, image data (12-bit image data) acquired with high definition by the CCD 1 is correlated. High-order bit data (8 bits) is compressed at a high compression rate by JPEG lossless encoding, and low-order bit data (4 bits) with low correlation is bit-shifted and packed to increase the processing speed. Thus, the speed of the encoding process can be increased while ensuring a high compression rate for the entire 12-bit image data (original data).
[0052]
Incidentally, with respect to the encoded image data, a coarse image can be reproduced at high speed by decoding only the higher-order compressed data. At this time, the higher-order compressed data block position information is stored at the earliest position of the compressed image data for one frame, and the higher-order compressed data block position (position indicated by ↓ in FIG. 4) can be recognized immediately. Therefore, a rough image can be easily reproduced at high speed. It is also possible to reproduce a partial image at a high speed by selecting a specific line from m lines.
In the first embodiment, in order to efficiently perform the process, the process returns to step S24 while the determination result of step S27 is “No”, and the determination result of step S28 is “No”. Return to step S22 to repeat the process, but as shown in FIG. 8, when the determination result of step S27 is "No" or the determination result of step S28 is "No" In addition, the processing may be repeated after the position information is read in step S21.
[0053]
(Second Embodiment)
Next, a second embodiment of the present invention will be described with reference to FIGS. This second embodiment also corresponds to claims 1 to 5.
In the second embodiment, when the upper compressed data block and the lower data block are stored in the storage medium 2, the upper compression of all the rows (m rows) is performed on the image data for one frame. The data block and the lower data blocks of all rows (m rows) are stored separately.
[0054]
The data format in this case is shown in FIG. The configuration of the encoding processing device and the decoding processing device for realizing the second embodiment is the same as the encoding processing device 10 (FIG. 1) and the decoding processing device of the first embodiment described above. 20 (FIG. 6) and the description thereof is omitted.
Further, with regard to the encoding of the higher-order bit data (JPEG lossless encoding), only the storage procedure in the storage medium 2 performed in steps S6 to S9 of the image encoding process shown in FIG. 3 is different. That is, in the second embodiment, as shown in FIG. 9, both the upper compressed data block position information and the lower data block position information are stored at the earliest position of the image data for one frame, and thereafter , M rows are stored together and upper data blocks are stored, and then m rows are stored together and lower data blocks are stored.
[0055]
In this way, the upper data blocks for m rows are collected, the lower data blocks for m rows are collected, and stored in the storage medium 2 in a separated state. When decoding and outputting only bit data (when reproducing only an 8-bit coarse image), only the upper compressed data needs to be decoded based on the upper compressed data block position information stored at the earliest position. , Processing speed is increased.
[0056]
In the compressed image data format in the second embodiment, each of the higher-order compressed data block position information and the lower-order data block position information is stored together. Each of the higher-order compressed data and the lower-order data is stored together. Therefore, it becomes possible to read out all the high-order compressed data or all the low-order data included in one screen at high speed based on each position information.
[0057]
Hereinafter, a procedure for decoding only the upper compressed data and reproducing a coarse image (upper 8 bits) at high speed will be described with reference to a program flowchart shown in FIG. This program is executed by the CPU 22 of the decryption processing apparatus 20 (FIG. 6).
First, in step S31, the upper data block position information is read.
[0058]
In step S32, data in the upper compressed data block for one row (upper compressed data) is read using the upper data block position information.
In step S33, decoding of one pixel of the read higher-order compressed data is performed by JPEG lossless decoding according to the reverse procedure of JPEG lossless encoding executed in the encoding process. The decoding in step S33 is performed in the same procedure as in step S24 in FIG.
[0059]
In step S34, it is determined whether or not the processing in step S33 described above has been performed for one row, that is, n pixels included in the same row, and when the processing for one row is completed, the next step Proceed to S35.
In step S35, it is determined whether or not the processing in steps S32 to S33 described above has been performed for all the rows (m rows) constituting one frame (one screen), and the processing for one frame has been completed. If not, the process returns to step S32 to repeat the process.
[0060]
On the other hand, when it is determined in step S35 that the decoding process for all the lines of one frame has been completed, the present program is terminated as it is (END).
According to the second embodiment described above, the image data is restored from the storage medium 2 in which high-definition original data (RAW data) represented by 12 bits per pixel is compressed and stored. By restoring only the upper 8 bits with strong correlation, a coarse image can be reproduced at high speed.
Even in the second embodiment, in order to efficiently perform the process, the process returns to step S33 while the determination result of step S34 is “No”, and the determination result of step S35 is “No”. Returns to step S32 and repeats the process, but as shown in FIG. 11, when the determination result of step S34 is "No", or when the determination result of step S35 is "No" In addition, both of the processes may be repeated from step S31.
[0061]
In the first and second embodiments described above, the case where the original data (RAW data) of the image is 12 bits has been described, but either image data larger than 12 bits or image data smaller than 12 bits is used. Even in this case, the present invention is applicable, and in this case as well, the processing speed can be increased while ensuring a high compression rate.
In the first and second embodiments described above, an example in which 12-bit original data (RAW data) is separated into upper 8 bits and lower 4 bits has been described. If it is 12-bit image data, it is empirically known that it is generally in the range of 6 bits to 9 bits.), It is possible to set the number of bits other than 8 bits. In this case, it may be determined for each frame (one screen) how many upper bits are used as upper bit data.
[0062]
In the first and second embodiments described above, both the encoding processing device and the decoding processing device are configured separately (for example, the encoding processing device is on the digital camera side, the decoding processing is Although the apparatus is described on the personal computer side), these two apparatuses may be incorporated in one apparatus (for example, both may be incorporated in a digital camera or the like).
[0063]
In the first and second embodiments described above, both the processing for separating the image data into the upper side and the lower side and the encoding processing by DPCM encoding of the upper side bit data are both performed in the encoding processing apparatus 10. However, the present invention is not limited to this, and the encoding by DPCM encoding is performed by a dedicated LSI integrated in an external chip so that the CPU 13 in the encoding processing apparatus 10 Only the lower side separation may be performed. In this case, an LSI that performs encoding by Ziv-Lempel or the like may be used as the external dedicated LSI instead of DPCM encoding. For decoding, a chip in which dedicated LSIs are similarly integrated is externally attached to the decoding processing device 20, and DPCM decoding (or Ziv-Lempel or the like) of higher-order bit data is performed by the dedicated LSI in the chip. (Decryption) may be performed. In this case, the CPU 22 of the decoding processing device 20 may perform the combination of the higher-order bit data and the lower-order bit data after decoding.
[0064]
In the first and second embodiments described above, the case where a still image is acquired by a digital camera or the like has been described. However, the present invention can also be applied to a case where a moving image is acquired.
[0065]
【The invention's effect】
Explained above Book According to the invention, high-definition image data is separated into higher-order bit data having high correlation with neighboring pixels and lower-order bit data having low correlation, and image data is encoded. Therefore, high-order bit data is encoded with a high compression rate, and low-order bit data is processed at a high processing speed, so that the image data as a whole is encoded with a high compression rate. However, the processing speed can be increased. Further, the management information can speed up the processing when selectively encoding / decoding image data of a specific row.
[0066]
or, Book According to the invention, the higher-order bit data and the lower-order bit data are individually blocked for each row, so that processing for selectively reproducing an image of a specific row by restoring the higher-order bit data Can be further increased in speed.
or, Book According to the invention, image data is compressed at a high compression rate with respect to high-order bit data having strong correlation.
[0067]
or, Book According to the present invention, image data captured in high definition can be restored as it is in high definition, and the user can process / modify the high-definition image data to output the processed / corrected image data as an output device. The image quality according to the specifications can be obtained.
or, Book According to the invention, it is possible to selectively restore only the high-order bit data of high-definition image data, and it is possible to reproduce a coarse image at a high speed by using the position information, and further to partially reproduce an image of a specific row. The processing speed can be increased.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a configuration of an encoding processing apparatus according to a first embodiment of the present invention.
FIG. 2 is a diagram illustrating a layout of color filters arranged on a light receiving surface of the image input device 2;
FIG. 3 is a flowchart illustrating image encoding processing according to the first embodiment.
FIG. 4 is a diagram illustrating a first example of a format of compressed image data obtained by the first embodiment.
FIG. 5 is a diagram illustrating a second example of a format of compressed image data obtained according to the first embodiment.
FIG. 6 is a block diagram showing a configuration of a decoding processing apparatus according to the first embodiment of this invention.
FIG. 7 is a flowchart illustrating image decoding processing according to the first embodiment;
FIG. 8 is a flowchart showing a different aspect of the image decoding process of the first embodiment.
FIG. 9 is a diagram showing a format of compressed image data obtained by the second embodiment.
FIG. 10 is a flowchart illustrating image decoding processing according to the second embodiment;
FIG. 11 is a flowchart illustrating a different aspect of the image decoding processing according to the second embodiment.
[Explanation of symbols]
1 Image input device (CCD)
2 storage media
10 Encoding processing device
12 Input data buffer
13 CPU
15 Output data buffer
20 Decoding processing device
21 Input data buffer
22 CPU
23 Restored data buffer

Claims (5)

In an image data compression method for compressing image data acquired at a constant number of bits per pixel by an image input means,
Separating the data for each pixel having a certain number of bits into upper bit data represented by an upper predetermined bit number and lower bit data represented by a lower predetermined bit number;
Performing variable length encoding on the upper bit data;
Packing the lower bit data with a bit shift operation addressed to a plurality of pixels;
Generating management data for individually managing the variable-length encoded high-order bit data and the packed low-order bit data;
Ri Do and a step of storing the variable length coded significant bits data and the packed lower bit data and the generated management data and the storage medium,
The image input means is an input means in which pixels are arranged in a matrix.
The variable-length encoded high-order bit data is blocked for each pixel included in the same row of at least one row,
The packed lower bit data is blocked for each pixel included in the same row of at least one row,
As the management data, position information indicating each storage position of the blocked upper bit data and the blocked lower bit data is generated,
The compressed image data obtained by the variable-length encoding is the upper bit data in which the position information indicating the storage position of the block upper bit data in the management data is the variable length encoded An image data compression method characterized by being provided at a position ahead of the other . The image data compression method according to claim 1,
The variable-length encoded high-order bit data is blocked for each pixel included in all the rows of image data for one frame,
The packed lower-order bit data is divided into blocks for each pixel included in all the rows of image data for one frame. The image data compression method according to claim 1 or 2,
The variable length coding is performed by Huffman coding and DPCM coding. The image data compression method according to any one of claims 1 to 3,
The fixed number of bits of data for each pixel is a value according to the detection accuracy of the image input means,
2. The image data compression method according to claim 1, wherein the upper-order predetermined number of bits is the number of upper-order bits having a strong tendency to correlate with neighboring pixels, and the value is obtained by an empirical rule. The image data compression method according to any one of claims 1 to 4,
The image data compression method, wherein the fixed number of bits per pixel acquired by the image input means is 12 bits or more.

Images