JP2010050572A - Image processing apparatus and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To carry out correction of axial chromatic aberration generated in an image focusing unit, and at the same time, and also to carry out the axial chromatic aberration generated in a non-focusing unit. <P>SOLUTION: An image processing apparatus corrects the axial chromatic aberration, and is provided with: the same time processing circuit (600) which separates an input image signal into a plurality of image signals of different optical frequency bands which at least includes a first image signal of a standard first optical frequency band, and a second image signal of a second optical frequency signal except the first optical frequency band; independent components analysis circuits (614, 618) which perform independent components analysis for the second image signal; and a YCMTX circuit (619) which composes a result of the independent components analysis of the second image signal by the independent components analysis circuits, and the first image signal. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an image processing apparatus and method, and more particularly to an image processing apparatus and method for correcting axial chromatic aberration using independent component analysis.

  In the lens of the image pickup apparatus, chromatic aberration caused by the difference in refractive index depending on the wavelength of light is known. Above all, axial chromatic aberration is a phenomenon in which the image formation position of light of each wavelength shifts back and forth on the axis. When focusing on an image of a certain wavelength, the image of light of another wavelength is not in focus. It was the cause of bleeding.

  Therefore, a conventional imaging apparatus has a technique in which a focal length correction value corresponding to the axial chromatic aberration inherent to the lens is stored in advance, and correction is performed using the focal length correction value for an image signal of light of each wavelength. (For example, refer to Patent Document 1).

JP-A-6-113309

  However, the above-described conventional imaging apparatus can correct axial chromatic aberration that occurs in an in-focus portion of an image, but cannot correct axial chromatic aberration that occurs in an out-of-focus portion of an image. There was a problem.

  The present invention has been made in view of the above problems, and can correct axial chromatic aberration generated in an in-focus portion of an image and also correct axial chromatic aberration generated in an out-of-focus portion of an image. The purpose is to be able to.

  In order to achieve the above object, an image processing apparatus according to the present invention for correcting axial chromatic aberration uses an input image signal as a first image signal in a first optical frequency band serving as a reference, and the first light. Separating means for separating a plurality of image signals in different optical frequency bands including at least a second image signal in a second optical frequency band other than the frequency band; and independent component analysis for the second image signal Analyzing means for performing, and combining means for combining the result of the independent component analysis of the second image signal by the analyzing means and the first image signal.

  Further, the image processing method of the present invention for correcting axial chromatic aberration uses an input image signal as a first image signal in a first optical frequency band serving as a reference and a second image signal other than the first optical frequency band. A separation step of separating into a plurality of image signals of different optical frequency bands including at least a second image signal of the optical frequency band, an analysis step of performing independent component analysis on the second image signal, A synthesis step of synthesizing the result of the independent component analysis of the second image signal by the analysis step and the first image signal.

  According to the present invention, it is possible to correct axial chromatic aberration that occurs in an in-focus portion of an image and to correct axial chromatic aberration that occurs in an out-of-focus portion of an image.

  The best mode for carrying out the present invention will be described below in detail with reference to the accompanying drawings.

<First Embodiment>
FIG. 1 is a block diagram illustrating a functional configuration example of an imaging apparatus according to the first embodiment of the present invention.

  In FIG. 1, 1 is a lens, 2 is an aperture, 3 is an image sensor, 4 is an analog front end (AFE) including a double correlation sampling (CDS) circuit and an A / D converter, and 5 is a lens 1 and an aperture. 2 and a drive circuit for the image sensor 3.

  A camera signal processing circuit 6 generates a luminance signal and a color difference signal from the captured image data. Reference numeral 7 denotes an image memory for storing image data. Reference numeral 8 denotes a compression encoding processing circuit for compressing and encoding the image data processed by the camera signal processing circuit 6. Reference numeral 10 denotes a recording medium that can be removed from the image pickup apparatus, and reference numeral 9 denotes a recording processing circuit that records the compression-coded image data on the recording medium 10.

  Reference numeral 11 denotes a system control unit that controls the entire imaging apparatus, and reference numeral 14 denotes a bus that performs communication between the system control unit 11 and each block in the imaging apparatus. A non-volatile memory (ROM) 12 stores a program describing a control method executed by the system control unit 11 and control data such as parameters and tables used when the program is executed. Reference numeral 13 denotes a volatile memory (RAM) that transfers and stores the program and control data stored in the nonvolatile memory 12 and is used when the system control unit 11 controls the imaging apparatus.

  The system control unit 11 controls the drive circuit 5, the camera signal processing circuit 6, the compression encoding processing circuit 8, and the recording processing circuit 9 via the bus 14. Information such as a control parameter necessary for each signal processing circuit and a program for controlling the circuit is read from the nonvolatile memory 12.

  Hereinafter, a photographing operation in the imaging apparatus having the above configuration will be described. Prior to the shooting operation, necessary programs, control data, and correction data are transferred from the nonvolatile memory 12 to the volatile memory 13 and stored at the start of the operation of the system control unit 11 such as when the imaging apparatus is turned on. Shall be kept. These programs and data are used when the system control unit 11 controls the imaging apparatus. Furthermore, if necessary, additional programs and data are transferred from the nonvolatile memory 12 to the volatile memory 13, or the system control unit 11 directly reads and uses the data in the nonvolatile memory 12.

  First, the drive circuit 5 drives the diaphragm 2 and the lens 1 according to a control signal sent from the system control unit 11 to form a subject image adjusted to an appropriate brightness on the image sensor 3. The image pickup device 3 is driven by a drive pulse from the drive circuit 5 controlled by the system control unit 11, converts the subject image into an electrical signal by photoelectric conversion, and outputs it as an analog image signal.

  The analog image signal output from the image sensor 3 is removed from the clock synchronization noise by the CDS circuit in the AFE 4 according to the operation pulse from the drive circuit 5 controlled by the system control unit 11, and the A / It is converted into a digital image signal by a D converter. The camera signal processing circuit 6 controlled by the system control unit 11 generates a luminance signal and color difference signals (Y, Cb, Cr) from the digital image signal output from the AFE 4.

  The image memory 7 is used for temporarily storing a digital image signal during signal processing or for storing image data which is a digital image signal subjected to signal processing. The compression encoding processing circuit 8 compresses and encodes the signal-processed digital image signal. The image data encoded by the compression encoding processing circuit 8 is converted into data suitable for the recording medium 10 (for example, file system data having a hierarchical structure) by the recording processing circuit 9 and recorded on the recording medium 10.

  FIG. 2 is a block diagram illustrating a functional configuration example of the camera signal processing circuit 6 according to the first embodiment.

  In the first embodiment, the case where the image sensor 3 is a single-plate image sensor in which primary color filters are arranged in a Bayer manner will be described.

  As an output of the AFE 4, digital image data of any color (hereinafter referred to as “RAW data”) corresponding to the color filter of the image sensor 3 is input to the camera signal processing circuit 6 for each pixel. In this RAW data state, each pixel position has only a color value corresponding to one of RGB color filters. For this reason, the synchronization processing circuit 600 separates RGB pixels from input RAW data, and then performs an interpolation operation so as to have RGB values at all pixel positions.

  The RGB image signals generated by the synchronization processing circuit 600 are respectively converted into orthogonal transform circuits 601 (second frequency analysis means), 602 (first frequency analysis means), and 603 (second frequency analysis means). Is orthogonally transformed.

  Next, orthogonal transform coefficients (second frequency analysis values) of R and B (second image signals in a second optical frequency band other than the first optical frequency band) subjected to orthogonal transform in the orthogonal transform circuits 601 and 603 are used. ) Is processed as follows. That is, the effective bands are set in the effective band determination circuits 604 and 605 excluding high frequency components in which the ratio of the noise component is increased. The effective band determination circuits 604 and 605 set all the points where the absolute values of the orthogonal transform coefficients of R and B are smaller than a certain threshold value Th to 0, and set the portions that are not 0 as effective bands. What value should be set for the threshold Th may be obtained experimentally by detecting the amount of the noise component. The determined effective band is output to the band limit frequency calculation circuits 606 and 609. The band limit frequency calculation circuits 606 and 609 extract the highest frequency (band limit frequency) within the effective band from the orthogonal transform coefficient, and output the value of the highest frequency to the band limit processing circuits 607 and 608. Further, the band limited frequency calculation circuits 606 and 609 output the orthogonal transform coefficient representing the frequency component higher than the highest frequency as 0 to the target frequency calculation circuits 610 and 611, respectively.

  The band limit processing circuits 607 and 608 perform the same process on the orthogonal transform coefficient of G that has been orthogonally transformed by the orthogonal transform circuit 602 using the highest frequency extracted by the band limit frequency calculation circuits 606 and 609. . That is, the orthogonal transform coefficient representing the frequency component higher than the highest frequency among the orthogonal transform coefficients (first frequency analysis values) of G (first image signal in the first optical frequency band) is set to 0, and the target Output to the frequency calculation circuits 610 and 611, respectively.

  In the target frequency calculation circuits 610 and 611, for example, the target value of the orthogonal transformation coefficient of R and B is calculated by calculation using the following equation (1).

Here, i and j are the positions of the orthogonal transform coefficients, Gx _ij is the value of the orthogonal transform coefficient at the positions i and j of G, and Rexp _ij and Bexp _ij are the target values at the positions of R and B, respectively. R _DC , G _DC , and B _DC indicate the DC component values of the orthogonal transform coefficients, and Rexp _ij and Bexp _ij are output to the ratio extraction circuits 612 and 613, respectively.

  The ratio extraction circuits 612 and 613 obtain the ratio of the output of the band limited frequency calculation circuits 606 and 609 and the output of the target frequency calculation circuits 610 and 611, for example, by calculation using the following equation (2), and select filter coefficients The data is output to the processing circuits 615 and 617.

Here, i and j are orthogonal transformation coefficient positions, Rx _ij and Bx _ij are orthogonal transformation coefficient values at the respective positions, and Rexp _ij and Bexp _ij are target values at the respective positions. Rdiff _ij and Bdiff _ij indicate the ratio between the target value of the orthogonal transform coefficient at each position and the actual coefficient value.

  The filter coefficient selection processing circuits 615 and 617 read the filter coefficient from the coefficient table 616 according to the ratio from the ratio extraction circuits 612 and 613, and output it to the independent component analysis circuits 614 and 618, respectively.

  Independent component analysis circuits 614 and 618 perform independent component analysis on the R and B images from the synchronization processing circuit 600 using the outputs (filter coefficients) from the filter coefficient selection processing circuits 615 and 617 as initial values, Each is sharpened.

  Here, a generally blurred image is represented by a linear sum of a non-blurred image and its differential image.

  Therefore, in the independent component analysis, in order to obtain a non-blurred image, a separation filter that separates a blurred image signal into a non-blurred image and N differential images is obtained. Is done.

Here, X (x, y) is a blurred image signal, and (W ₁ , W ₂ ,... W _N ) is a separation filter. If y ₁ (x, y) is an image signal that is not blurred among (y ₁ (x, y), y ₂ (x, y), ... y _N (x, y)), y ₂ ( x, y) ... y _N (x, y) represents a differential image.

  Therefore, in the first embodiment, first, from an R or B image that is out of focus due to axial chromatic aberration, an unblurred R or B image and an R or B blurred component image (differential image). The separation filter that separates the two is obtained by independent component analysis. The axial chromatic aberration is corrected by acquiring a non-blurred R or B image.

By the way, in general independent component analysis, the initial values of the filter coefficients of the separation filters (W ₁ , W ₂ ,... W _N ) are set to random values, and the image separation accuracy is evaluated while improving the filter coefficients. Acquired by doing so-called learning. Therefore, if the number of times of learning increases to some extent, the image separation accuracy converges to some extent, but it does not converge or takes time to converge depending on the state of the input image and the learning process.

  On the other hand, since an unblurred image and a differential image can be acquired from independent component analysis, the separation filter includes a sharpening filter and a differential filter.

  Therefore, in the first embodiment, the sharpening filter and the filter coefficient of the differential filter from the filter coefficient selection processing circuits 615 and 617 are used as initial values of the separation filter, so that the separation accuracy is reliably converged, It was set as the structure which can shorten the time to convergence.

  In the YCMTX circuit 619, for example, the luminance signal Y and the color difference signals Cb and Cr are converted and output to the compression coding processing circuit 8 by the calculation of the following equation (4).

  Here, R ′ is an output from the independent component analysis circuit 614, and B ′ is an output from the independent component analysis circuit 618.

  Next, an example of an axial chromatic aberration correction procedure performed by the camera signal processing circuit 6 which is a feature of the first embodiment will be described with reference to a flowchart shown in FIG.

  First, in step S101, RGB image signals are held in a line memory for 7 lines (the image memory 7 can be used). Then, the image signal interpolated so as to have RGB values at all pixel positions by the synchronization processing circuit 600 is divided into 8 × 8 pixel blocks by the orthogonal transformation circuits 601, 602, 603, and orthogonal As the conversion process, 8 × 8 Hadamard transform is performed. Here, the orthogonal transform may be a discrete cosine transform, and the block size is not limited to 8 × 8 pixels. FIG. 4 shows a conceptual diagram of coefficient distribution when orthogonal transformation is performed. The region where the coefficient increases depends on the frequency component of the image.

  In step S102, the effective bandwidth determination circuits 604 and 605 set all the points where the absolute values of the orthogonal transform coefficients of R and B are smaller than a certain threshold Th to 0, and set the portions that are not 0 as effective bands. . Next, in step S103, the band-limited frequency calculation circuits 606 and 609 extract the boundary where the respective orthogonal transform coefficients are larger and smaller than the threshold Th as the highest frequency. Here, for example, as shown in FIG. 5 (a), when f is the highest frequency, the coefficient is 0 (the plain portion in FIG. 5) and the coefficient is not 0 (the shaded portion in FIG. 5). If it remains, it is set to 0 even if it is not set to 0 as shown in FIG.

  Next, in step S104, the band limitation processing circuits 607 and 608 set the orthogonal transform coefficient having a frequency higher than the highest frequency extracted in step S103 out of the G orthogonal transform coefficients to 0.

  Next, in step S105, the target frequency calculation circuits 610 and 611 calculate the target values of the R and B orthogonal transform coefficients using Expression (1). Further, in step S106, the ratio extraction circuits 612 and 613 obtain the ratio between the target value of each orthogonal transform coefficient and the actual coefficient value using Expression (2).

  Next, in step S107, the filter coefficient selection processing circuits 615 and 617 read the filter coefficient from the coefficient table 616 according to the ratio obtained in step S106. For example, as shown in FIG. 6, it is assumed that when the orthogonal transform coefficients are divided into the regions (A) to (K), the ratio is large in the region (A). In this case, since the R or B image signal lacks the horizontal low-to-mid frequency components, filter coefficients that correct the horizontal low-to-mid frequency components are used in the coefficient table. Of these, the initial value of the sharpening filter among the separation filters is used. On the other hand, the differential filter reads a filter coefficient for differentiation, such as a Laplacian filter, from the coefficient table and sets it as an initial value of the differential filter. In addition, the area division of the orthogonal transform coefficient is not limited to this, and not only the filter coefficient corresponding to the area but also the filter coefficient corresponding to the magnitude of the ratio is prepared, and the magnitude of the area and the ratio is prepared. The filter coefficient may be read out according to the above.

  Next, in step S108 and step S109, the independent component analysis circuits 614 and 618 apply the independent component analysis to the R and B images, respectively, using the filter coefficient obtained in step S107 as an initial value for learning, thereby sharpening processing. I do. Further, it is determined whether or not the sharpening accuracy of the image has converged to some extent. If it has not converged, the process proceeds to step S108, and if it has converged, the process proceeds to step S110.

  Next, in step S110, the YCMTX circuit 619 converts (synthesizes) the luminance signal Y and the color difference signals Cr and Cb using the above-described equation (4) using the result of the independent component analysis obtained in steps S108 and S109.

  As described above, according to the first embodiment, by performing independent component analysis on the R and B image signals, axial chromatic aberration is corrected regardless of the in-focus portion and the out-of-focus portion. Can do. Also, by performing independent component analysis using the initial values of the separation filters for independent component analysis obtained from the RGB image signals, the number of independent component analysis learnings can be reduced and the processing time can be shortened. it can.

<Second Embodiment>
Next, a second embodiment of the present invention will be described with reference to FIGS. Note that in the configuration of the imaging apparatus according to the second embodiment, the description of the same configuration as that of the imaging apparatus described with reference to FIGS. 1 and 2 in the first embodiment is omitted. The first embodiment is different from the first embodiment in that the camera signal processing circuit 6 includes an orthogonal transform circuit 620, 621, a ratio extraction circuit 622, 623, a filter 624, 627, a filter coefficient selection processing circuit 625, 626, The coefficient table 628 is added. Second, a procedure for performing correction when the sharpening processing of the independent component analysis circuits 614 and 618 is overcorrected is added.

  FIG. 7 is a block diagram illustrating a functional configuration example of the camera signal processing circuit 6 in the second embodiment.

  In FIG. 7, orthogonal transform circuits 620 and 621 (third frequency analysis means) orthogonally transform the outputs from the independent component analysis circuits 614 and 618. Further, the ratio extraction circuits 622 and 623 (second ratio extraction means) output the orthogonal transform coefficients (third frequency analysis values) from the orthogonal transform circuits 620 and 621 and the outputs from the target frequency calculation circuits 610 and 611, respectively. This is a circuit for obtaining a ratio. Filter coefficient selection processing circuits 625 and 626 (second selection means) are circuits for reading out filter coefficients corresponding to outputs from the ratio extraction circuits 622 and 623 from the coefficient table 628. Filters 624 and 627 are circuits that perform filter processing on the outputs from the independent component analysis circuits 614 and 618 using the filter coefficients from the filter coefficient selection processing circuits 625 and 626.

  Next, an example of an axial chromatic aberration correction procedure performed by the camera signal processing circuit 6 which is a feature of the second embodiment will be described based on a flowchart shown in FIG. In addition, about the process of step S101-S110, since it is the same as that of the process of step S101-S110 shown in FIG. 3 of 1st Embodiment, description is abbreviate | omitted.

  In step S201, the outputs from the independent component analysis circuits 614 and 618 are respectively held in a line memory for 7 lines (the image memory 7 can be used). Then, the orthogonal transform circuits 620 and 621 divide into 8 × 8 blocks and perform 8 × 8 Hadamard transform as orthogonal transform processing. Here, the orthogonal transform may be a discrete cosine transform, and the block size is not limited to 8 × 8 pixels.

  Next, in step S202, the ratio extraction circuits 622 and 623 obtain the ratio between the outputs from the orthogonal transform circuits 620 and 621 and the outputs from the target frequency calculation circuits 610 and 611.

  Next, in step S203, the filter coefficient selection processing circuits 625 and 626 obtain filter coefficients from the coefficient table 628 according to the ratio obtained in step S202.

  Next, in step S204, the outputs from the independent component analysis circuits 614 and 618 are subjected to low-pass filter processing by the filters 624 and 627 using the filter coefficients from the filter coefficient selection processing circuits 625 and 626.

  As described above, in the second embodiment, the ratio of the orthogonal transformation coefficient obtained by orthogonal transformation of the outputs from the independent component analysis circuits 614 and 618 and the output from the target frequency calculation circuits 610 and 611 is obtained. Then, by performing filter processing on the outputs from the independent component analysis circuits 614 and 618 using a filter coefficient corresponding to the ratio, it is possible to correct overcorrection that may occur due to independent component analysis.

<Third Embodiment>
Next, a third embodiment of the present invention will be described with reference to FIGS. 9 and 10. Note that in the configuration of the imaging apparatus of the third embodiment, the description of the same configuration as that of the imaging apparatus described in the first and second embodiments is omitted. The first and second embodiments are different from the first embodiment in that counters 629 and 630 for counting the number of learnings are added to the camera signal processing circuit 6 in addition to the configuration shown in FIG. Second, the procedure for determining whether or not the number of counts has reached a predetermined number and changing the process according to the determination result is added to the axial chromatic aberration correction procedure of the camera signal processing circuit 6.

  FIG. 9 is a block diagram illustrating a functional configuration example of the camera signal processing circuit 6 in the third embodiment.

  In FIG. 9, counters 629 and 630 count the number of learnings of the independent component analysis circuits 614 and 618 and output the count number to the independent component analysis circuits 614 and 618.

  Next, an example of an axial chromatic aberration correction procedure performed by the camera signal processing circuit 6 which is a feature of the third embodiment will be described with reference to a flowchart shown in FIG. In addition, about the process of step S101-S110, since it is the same as that of the process of step S101-S110 shown in FIG. 3 of 1st Embodiment, description is abbreviate | omitted. Further, steps S201 to S204 are the same as the processes of steps S201 to S204 shown in FIG. 8 of the second embodiment, and thus description thereof is omitted.

  In step S109, it is determined whether or not the sharpening accuracy of the image has converged to some extent (for example, to a predetermined level), and if not converged, the process proceeds to step S301.

  In step S301, it is determined whether or not the number of learning has reached a predetermined number of times Cnt (threshold). As a result of the determination, if the number of learning has reached the predetermined number, the process proceeds to step S201, and if the learning number has not reached the predetermined number, the process proceeds to step S302. In step S302, the counters 629 and 630 are incremented by 1, and the process returns to step S108.

  Here, the predetermined number of times Cnt may be held in a memory (not shown) in the camera signal processing circuit 6 or may be set from the outside.

  As described above, according to the third embodiment, whether or not the learning frequency of the independent component analysis circuits 614 and 618 has reached a predetermined number when the image sharpening accuracy is not converged is determined. In response to this, the process of whether to perform learning again or to stop learning and proceed to the next step is changed. Thereby, the processing time can be shortened, and the inconvenience that the processing does not end because the separation accuracy does not converge can be prevented.

  In the first to third embodiments, the axial chromatic aberration correction is performed on the R and B images. However, the present invention is not limited to this, and any color is used as a reference. Other color corrections may be performed.

  In the first to third embodiments, the axial chromatic aberration correction in the case where the image pickup device 3 is covered with the Bayer array color filter and the RGB image signal is output as the image signal has been described. However, the present invention is not limited to RGB image signals, and it goes without saying that axial chromatic aberration that occurs between image signals of different frequency components can be corrected by a similar procedure.

<Other embodiments>
Note that the present invention can be applied to a system (for example, a digital camera) composed of a single device even when applied to a system composed of a plurality of devices (for example, a camera head, a scanner, an interface device, a host computer, etc.). May be.

  The object of the present invention can also be achieved as follows. First, a storage medium (or recording medium) that records a program code of software that implements the functions of the above-described embodiments is supplied to a system or apparatus. Then, the computer (or CPU or MPU) of the system or apparatus reads and executes the program code stored in the storage medium. In this case, the program code itself read from the storage medium realizes the functions of the above-described embodiments, and the storage medium storing the program code constitutes the present invention.

  Further, by executing the program code read by the computer, not only the functions of the above-described embodiments are realized, but also the following can be achieved. That is, when the operating system (OS) running on the computer performs part or all of the actual processing based on the instruction of the read program code, the functions of the above-described embodiments are realized by the processing. It is. Examples of the storage medium for storing the program code include a flexible disk, hard disk, ROM, RAM, magnetic tape, nonvolatile memory card, CD-ROM, CD-R, DVD, optical disk, magneto-optical disk, MO, and the like. Can be considered. Also, a computer network such as a LAN (Local Area Network) or a WAN (Wide Area Network) can be used to supply the program code.

It is a block diagram which shows the function structural example of the imaging device in the 1st Embodiment of this invention. It is a block diagram which shows the function structural example of the camera signal processing circuit in the 1st Embodiment of this invention. It is a flowchart which shows an example of the axial chromatic aberration correction procedure performed with the camera signal processing circuit in the 1st Embodiment of this invention. It is a conceptual diagram which shows the coefficient distribution of orthogonal transformation. It is a figure which shows an example of operation | movement of the band limiting frequency calculation circuit in the camera signal processing circuit in the 1st Embodiment of this invention. It is a figure which shows an example of the filter coefficient selection process in the camera signal processing circuit in the 1st Embodiment of this invention. It is a block diagram which shows the function structural example of the camera signal processing circuit in the 2nd Embodiment of this invention. It is a flowchart which shows an example of the axial chromatic aberration correction procedure performed with the camera signal processing circuit in the 2nd Embodiment of this invention. It is a block diagram which shows the function structural example of the camera signal processing circuit in the 3rd Embodiment of this invention. It is a flowchart which shows an example of the axial chromatic aberration correction procedure performed with the camera signal processing circuit in the 3rd Embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Lens 2 Aperture 3 Image pick-up element 4 Analog front end part 5 Drive circuit 6 Camera signal processing circuit 7 Image memory 8 Compression encoding processing circuit 9 Recording processing circuit 10 Recording medium 11 System control part 12 Non-volatile memory (ROM)
13 Volatile memory (RAM)
14 Bus 600 Synchronization processing circuit 601, 602, 603, 620, 621 Orthogonal transformation circuit 604, 605 Effective band determination circuit 606, 609 Band limit frequency calculation circuit 607, 608 Band limit processing circuit 610, 611 Target frequency calculation circuit 612, 613, 622, 623 Ratio extraction circuit 614, 618 Independent component analysis circuit 615, 617, 625, 626 Filter coefficient selection processing circuit 616, 628 Coefficient table 619 YCMTX circuit 624, 627 Filter 629, 630 Counter

Claims (16)

An image processing device for correcting axial chromatic aberration,
A plurality of input image signals including at least a first image signal in a first optical frequency band serving as a reference and a second image signal in a second optical frequency band other than the first optical frequency band Separating means for separating the image signals in different optical frequency bands;
Analyzing means for performing independent component analysis on the second image signal;
An image processing apparatus comprising: combining means for combining the result of independent component analysis of the second image signal by the analyzing means and the first image signal. First frequency analysis means for obtaining a first frequency analysis value from the first image signal;
Second frequency analysis means for obtaining a second frequency analysis value from the second image signal;
Effective band determining means for determining an effective band based on the second frequency analysis value;
A band limit frequency calculating means for obtaining a band limit frequency from the effective band determined by the effective band determination means and limiting the second frequency analysis value at the band limit frequency;
Band limiting means for limiting the first frequency analysis value at the band limiting frequency obtained by the band limiting frequency calculating means;
Target frequency calculation means for obtaining a target value of the second frequency analysis value from a ratio between the second frequency analysis value limited by the band limit frequency and the first frequency analysis value;
Ratio extraction means for obtaining a ratio between the target value obtained by the target frequency calculation means and the second frequency analysis value restricted by the band restriction frequency;
Selection means for selecting an initial value of independent component analysis according to the ratio obtained by the ratio extraction means;
The image processing apparatus according to claim 1, wherein the analysis unit performs independent component analysis using an initial value selected by the selection unit. Third frequency analysis means for obtaining a third frequency analysis value from the result of independent component analysis of the second image signal by the analysis means;
Second ratio extraction means for obtaining a ratio between the third frequency analysis value and the target value obtained by the target frequency calculation means;
Second selecting means for selecting a filter coefficient in accordance with the ratio obtained by the second ratio extracting means;
Filter means for performing low-pass filtering on the result of independent component analysis of the second image signal using the filter coefficient selected by the second selection means;
The image processing apparatus according to claim 2, wherein the synthesizing unit synthesizes the result of independent component analysis of the filtered second image signal and the first image signal. A counter that counts the number of independent component analyzes performed by the analyzing means;
The analysis means performs the processing on the second image signal until the accuracy of the independent component analysis converges to a predetermined level or the number of times counted by the counter reaches a predetermined threshold. The image processing apparatus according to claim 1, wherein the independent component analysis is repeatedly executed.   The image processing apparatus according to claim 3, wherein the first to third frequency analysis units perform Hadamard transform.   The image processing apparatus according to claim 3, wherein the first to third frequency analysis units perform discrete cosine transform.   The image processing apparatus according to claim 2, wherein the effective band determination unit determines a frequency band whose frequency analysis value corresponding to each frequency is larger than a preset threshold as an effective band.   The image processing apparatus according to claim 2, wherein the band limited frequency is a maximum frequency of the effective band.   The image processing apparatus according to claim 2, wherein the selection unit divides the ratio into a plurality of regions and selects a coefficient that corrects a region having a large ratio among the regions as an initial value.   The second selection unit divides the ratio obtained by the second ratio extraction unit into a plurality of regions, and selects a filter coefficient that corrects a region having a large ratio among the regions. The image processing apparatus according to claim 3. An image processing method for correcting axial chromatic aberration,
A plurality of input image signals including at least a first image signal in a first optical frequency band serving as a reference and a second image signal in a second optical frequency band other than the first optical frequency band A separation step of separating into image signals of different optical frequency bands;
An analysis step of performing independent component analysis on the second image signal;
An image processing method comprising: a combining step of combining the result of independent component analysis of the second image signal in the analyzing step and the first image signal. A first frequency analysis step of obtaining a first frequency analysis value from the first image signal;
A second frequency analysis step of obtaining a second frequency analysis value from the second image signal;
An effective band determining step of determining an effective band based on the second frequency analysis value;
Obtaining a band limited frequency from the effective band determined by the effective band determining step, and limiting the second frequency analysis value at the band limited frequency;
A band limiting step of limiting the first frequency analysis value at the band limiting frequency obtained in the band limiting frequency calculating step;
A target frequency calculating step for obtaining a target value of the second frequency analysis value from a ratio between the second frequency analysis value limited by the band limiting frequency and the first frequency analysis value;
A ratio extraction step for obtaining a ratio between the target value obtained in the target frequency calculation step and the second frequency analysis value limited by the band limitation frequency;
A selection step of selecting an initial value of independent component analysis according to the ratio obtained in the ratio extraction step,
12. The image processing method according to claim 11, wherein in the analysis step, independent component analysis is performed using the initial value selected in the selection step. A third frequency analysis step for obtaining a third frequency analysis value from a result of independent component analysis of the second image signal by the analysis step;
A second ratio extraction step for obtaining a ratio between the third frequency analysis value and the target value obtained in the target frequency calculation step;
A second selection step of selecting a filter coefficient according to the ratio obtained in the second ratio extraction step;
A filter step of performing low-pass filter processing on the result of independent component analysis of the second image signal using the filter coefficient selected in the second selection step;
13. The image processing method according to claim 12, wherein the synthesis step synthesizes the result of the independent component analysis of the filtered second image signal and the first image signal. A counter that counts the number of independent component analyzes performed by the analysis step;
The analysis step is performed on the second image signal until the accuracy of the independent component analysis converges to a predetermined level or the number of times counted by the counter reaches a predetermined threshold. The image processing method according to claim 11, wherein the independent component analysis is repeatedly executed.   The program for making a computer perform each process of the image processing method of any one of Claims 11 thru | or 14.   A computer-readable storage medium storing the program according to claim 15.

Images