JP2004040300A - Image processing apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing apparatus the compression rate of which can be increased more than that of a conventional image processing apparatus in the case of reversibly compressing three-primary-color image data by one frame obtained from an imaging element. <P>SOLUTION: In the case of reversely compressing R, G, B CCDRAW data by one frame obtained from a CCD 10 by using a Huffman table, a Huffman table different from each color channel is used. That is, a Huffman table optimizing circuit 20 selects an optimum Huffman table to each color channel among a plurality of Huffman tables prepared in advance on the basis of the R, G, B CCDRAW data temporarily stored in a memory 14 and gives the selected Huffman table to a Huffman encoding circuit 18. The Huffman encoding circuit 18 uses the optimum Huffman table different from each color channel to apply Huffman encoding to the R, G, B CCDRAW data. Thus, the high compression rate is obtained more than in the case of Huffman encoding by using a single Huffman table. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus, and more particularly to an image processing apparatus for reversibly compressing image data of three primary colors for one frame obtained from an image sensor.
[0002]
[Prior art]
Conventionally, when image data obtained from an image sensor such as a CCD is recorded on a recording medium, there is a recording method in which image data is compressed and recorded by an irreversible compression method such as JPEG, or CCD RAW data is recorded as it is. is there.
[0003]
There has also been proposed a data processing device that prepares a plurality of Huffman tables, reversibly compresses input data obtained from an image sensor such as a CCD, and sequentially selects the data having the smallest data size. (Japanese Patent Laid-Open No. 2001-61067). The input data is a difference value of CCD RAW (raw) data of the same color channel output from a DPCM conversion circuit (differential pulse-code modulation).
[0004]
[Problems to be solved by the invention]
By the way, when an irreversible compression method such as JPEG is used, there is a problem that sufficient bit accuracy cannot be obtained at the time of reproduction or processing of image data, and image quality deteriorates. On the other hand, if CCD RAW data is recorded as it is, There is a problem that the amount of data greatly increases as the number of pixels of the image sensor increases and the number of bits for A / D conversion increases.
[0005]
Further, the data processing apparatus described in Japanese Patent Application Laid-Open No. 2001-61067 does not disclose a method of compressing with a Huffman table that differs for each color channel.
[0006]
The present invention has been made in view of such circumstances, and when reversibly compressing image data of three primary colors for one frame obtained from an image sensor, an image that can have a higher compression ratio than a conventional apparatus. An object is to provide a processing apparatus.
[0007]
[Means for Solving the Problems]
In order to achieve the above object, an invention according to claim 1 is an image processing device that reversibly compresses image data of three primary colors for one frame obtained from an image sensor, Huffman coding means for reversibly compressing each color channel by Huffman coding using a different Huffman table for each color channel, and Huffman table optimization means for optimizing the Huffman table of the Huffman coding means for each color channel And recording means for recording the compressed data of the three primary colors compressed by the Huffman encoding means on a recording medium.
[0008]
According to the first aspect of the present invention, when the image data of the three primary colors for one frame is reversibly compressed using the Huffman table, a different Huffman table is used for each color channel. That is, since the image data of the three primary colors has different histograms for each gradation for each color channel, the image data for each color channel is used instead of using the optimum (one) Huffman table for the image data of the three primary colors. The compression rate is higher when the Huffman table that is optimal for the above is used. Therefore, in the present invention, image data of three primary colors for one frame is reversibly compressed by Huffman coding for each color channel using a different Huffman table for each color channel. Further, the Huffman table that is different for each color channel is optimized for each color channel by the Huffman table optimizing means and is given to the Huffman encoding means.
[0009]
3. The Huffman table optimizing means includes means for obtaining, for each color channel, a histogram for each gradation for the image data of the three primary colors for one frame, and the obtained histogram for each color channel. Based on the above, the optimum Huffman table for each color channel is selected from a plurality of Huffman tables prepared in advance, or is generated by calculation.
[0010]
That is, the optimum Huffman table for each color channel may be selected from a plurality of Huffman tables prepared in advance, or may be generated by calculation based on the histogram obtained for each color channel. Also good.
[0011]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, preferred embodiments of an image processing apparatus according to the present invention will be described in detail with reference to the accompanying drawings.
[0012]
FIG. 1 is a block diagram of a main part of a digital camera including an image processing apparatus according to the present invention.
[0013]
This digital camera records a still image picked up when a shutter button (not shown) is pressed on a recording medium such as a memory card. When the shutter button is turned on, image light indicating a subject passes through a lens optical system (not shown). Then, an image is formed on the light receiving surface of the imaging device (CCD) 10. The signal charges accumulated in the CCD 10 are sequentially read out as voltage signals corresponding to the signal charges, converted into digital signals (hereinafter referred to as “CCD RAW data”) by the A / D converter 12, and then the SDRAM or the like. Temporarily stored in the memory 14. The CCD 10 has an R, G, B color filter array, for example, a Bayer array, and the CCD RAW data output from the A / D converter 12 is a color filter as shown in FIG. Corresponding to the array, it is output in the order of R, G, B, G, R, G.
[0014]
The signal processing circuit 15 performs color interpolation processing corresponding to a white balance correction unit that adjusts R, G, and B gains of CCD RAW data to adjust white balance, a gamma correction unit, and a color filter array of the CCD 10. A color interpolation processing unit, a YC processing unit that performs generation processing (YC processing) of luminance data Y and color difference data C _r and C _b , a contour enhancement unit that adds an aperture signal to luminance data Y and performs contour enhancement. A noise reduction processing unit that performs noise reduction processing such as smoothing processing and median filter processing, and a saturation enhancement unit that increases and decreases the gains of the color difference data C _r and C _b are included.
[0015]
The image data processed by the signal processing circuit 15 is encoded by a video encoder (not shown) and output to a liquid crystal monitor provided on the back of the camera, whereby a subject image is displayed on the display screen of the liquid crystal monitor. .
[0016]
The digital camera also includes an image processing apparatus according to the present invention that reversibly compresses CCD RAW data and records the data in a recording medium 16 such as a memory card.
[0017]
The image processing apparatus includes a Huffman encoding circuit 18 that reversibly compresses R, G, and B CCD RAW data for one frame by Huffman encoding for each color channel using a different Huffman table for each color channel; A Huffman table optimization circuit 20 that optimizes the Huffman table in the Huffman encoding circuit 18 for each color channel, and compressed data for each of R, G, and B compressed by the Huffman encoding circuit 18 are recorded on the recording medium 16. And recording means.
[0018]
Next, the operation of the Huffman coding circuit 18 and the Huffman table optimization circuit 20 will be described.
[0019]
First, Huffman coding using the Huffman table will be briefly described.
[0020]
A typical lossless coding is Huffman coding. Huffman code is a statistical data compression technique that assigns long codes to those with low occurrence probability and short codes to those with high occurrence probability. This is a technique for shortening the average code length by adding the value.
[0021]
The CCD RAW data has, for example, a code length of 12 bits (2 ¹² = 4098 gradation), but the occurrence probability within one frame of the CCD RAW data of each gradation is not uniform. Therefore, the average code length can be shortened by assigning a short code length to a gradation having a high occurrence probability and assigning a long code length to a gradation having a low occurrence probability. The Huffman table is a conversion table for newly encoding by assigning a code length to input data.
[0022]
Incidentally, the CCD RAW data has a different frequency distribution for each of the R, G, and B color channels. Therefore, the average code length of the R, G, and B CCD RAW data is shorter when the Huffman coding is performed with the Huffman table suitable for each color channel than when the Huffman coding is performed with a single Huffman table (the compression rate is reduced). Can be high).
[0023]
Therefore, in the present invention, R, G, and B CCD RAW data are compressed by Huffman coding for each color channel using a different Huffman table for each color channel.
[0024]
That is, the CCD RAW data (FIG. 2A) temporarily stored in the memory 14 in the order of the color filter arrangement of the CCD 10 is decomposed for each color channel, R CCD RAW data, G CCD RAW data, and B CCD RAW data. (FIG. 2B), and the CCD RAW data in the order of the color filter temporarily stored in the memory 14 is rewritten to CCD RAW data for each color channel.
[0025]
The Huffman table optimizing circuit 20 has, for example, a plurality of types of Huffman tables. From these Huffman tables, the Huffman tables (Table R, Table G, optimal for CCD RAW data of R, G, and B color channels) are provided. Select Table B).
[0026]
As a method for selecting an optimum Huffman table, a histogram (number of appearances for each gradation) is obtained for each CCD RAW data of each color channel, and an optimum Huffman table that can minimize the average code length is selected from the histogram. .
[0027]
The Huffman table optimization circuit 20 outputs the optimum Huffman table (Table R, Table G, Table B) selected as described above to the Huffman coding circuit 18 (FIGS. 2C and 2D).
[0028]
The Huffman encoding circuit 18 uses the optimum Huffman table (Table R, Table G, Table B) that is different for each color channel provided from the Huffman table optimization circuit 20 to provide CCD RAW data for each color channel of R, G, and B. Is compressed by Huffman coding.
[0029]
The R, G, and B compressed data that are Huffman encoded in this way are recorded on the recording medium 16.
[0030]
FIG. 3 is a diagram showing an example of image files of R, G, and B compressed data recorded on the recording medium 16 as described above.
[0031]
As shown in the drawing, Huffman-coded R, G, and B compressed data is recorded in the main area of the image file, and the compressed data is decoded into uncompressed CCD RAW data in the header area. Information (digital camera model data, recording size, Huffman table information for each of R, G, and B) and the like are recorded.
[0032]
The color filter array of the CCD can be known from the model data of the digital camera. Also, it is possible to know which Huffman table the R, G, and B compressed data is encoded by using the Huffman table information (for example, information specifying the Huffman table such as a table number).
[0033]
In this embodiment, the Huffman table optimization circuit 20 selects an optimum Huffman table that maximizes the compression rate for each color channel from a plurality of Huffman tables prepared in advance. An optimal Huffman table may be generated by calculation based on a histogram indicating the distribution. Further, instead of directly encoding the CCD RAW data, the difference data of adjacent pixels of the CCD RAW data decomposed for each of R, G, and B is obtained, and the difference data may be encoded as described above. Good.
[0034]
【The invention's effect】
As described above, according to the present invention, when image data of three primary colors for one frame is reversibly compressed using a Huffman table, Huffman encoding is performed using a different Huffman table for each color channel. Therefore, the compression rate can be made higher than when Huffman encoding is performed using a single Huffman table.
[Brief description of the drawings]
FIG. 1 is a block diagram of a main part of a digital camera including an image processing apparatus according to the present invention. FIG. 2 is a diagram illustrating a compression process using Huffman coding according to the present invention. Figure showing an example of an image file [Explanation of symbols]
DESCRIPTION OF SYMBOLS 10 ... Imaging device (CCD), 12 ... A / D converter, 14 ... Memory, 15 ... Signal processing circuit, 16 ... Recording medium, 18 ... Huffman encoding circuit, 20 ... Huffman table optimization circuit

Claims (2)

In an image processing apparatus that reversibly compresses image data of three primary colors for one frame obtained from an image sensor,
Huffman encoding means for reversibly compressing the image data of the three primary colors for one frame by Huffman encoding for each color channel using a different Huffman table for each color channel;
Huffman table optimizing means for optimizing the Huffman table of the Huffman encoding means for each color channel;
Recording means for recording the compressed data of the three primary colors compressed by the Huffman encoding means on a recording medium;
An image processing apparatus comprising: The Huffman table optimizing means has means for obtaining, for each color channel, a histogram for each gradation for the image data of the three primary colors for one frame, and for each color channel based on the obtained histogram for each color channel. The image processing apparatus according to claim 1, wherein an optimum Huffman table is selected from a plurality of Huffman tables prepared in advance or is generated by calculation.

Images