JP2008117440A - Digital image position adjusting apparatus and method, data recording / reproducing apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device achieving accurate positioning by two dimensional PLL (phase locked loop) without depending on a marker. <P>SOLUTION: A digital image positioning device is provided with; a vertical interpolation means 105 which compensates displacement in a vertical direction to a digital image and outputs a vertical interpolation digital image; a lateral direction interpolation means 106 which compensates displacement in a lateral direction and outputs a vertical and lateral interpolation digital image; a lateral error detection means 107 which detects a lateral direction phase error of vertical and lateral direction interpolation digital image; a vertical error detection means 108 which detects a vertical direction phase error of the vertical and lateral direction interpolation digital image; a lateral loop filter 109 which outputs a lateral frequency error from the lateral direction phase error; a vertical loop filter 110 which outputs a vertical frequency error from the vertical direction phase error; a lateral frequency phase conversion means 111 to output a lateral tap coefficient from a lateral frequency error; and a vertical frequency phase conversion means 112 to output a vertical tap coefficient from a vertical frequency error. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an apparatus for aligning a two-dimensional digital image, and in particular, can be suitably used for aligning a two-dimensional image of a data recording / reproducing apparatus using holography.

  In recent years, the use of multimedia has progressed and the data handled has increased. In order to record this large amount of data, research has been conducted on a recording device (hereinafter referred to as a hologram memory) using holography.

  Since the hologram memory can record and reproduce two-dimensional page data at a time, the data transfer speed is very high, and it is said that the storage capacity can be increased by using the technique of multiplex recording. Yes.

As a problem in the hologram memory, there is a case where the cell which is the minimum modulation unit of the spatial light modulator and the pixel of the image sensor cannot correspond one-to-one. This is because one pixel may not be sensed for one cell due to lens aberration, disc tilt, mechanical expansion / contraction, and the like. If one pixel cannot be sensed for one cell, the original image cannot be reproduced by the sampling theorem, and the number of errors increases. In order to prevent this, for example, oversampling of 1.5 pixels is performed for one cell. The digital image output from the image sensor is affected by lens aberration, disc tilt, mechanical expansion and contraction, and oversampling. It is necessary to perform alignment so that this corresponds to one pixel for one cell. In order to facilitate this alignment, conventionally, a pattern called an alignment marker (or Sync code) is written together with data. Position alignment can be performed with this marker (see, for example, Patent Document 1).
JP 2004-310957 A (page 4-11, FIG. 1)

  However, in the alignment using the marker, if the marker detection fails, the error increases extremely. In particular, when the S / N ratio is poor and the distortion of the lens is large or when the image is rotated, the marker detection accuracy is lowered, and the marker detection fails, resulting in an increase in error. Also, when the marker peripheral portion is deteriorated by dust or the like, the marker detection accuracy is lowered, and the marker detection fails.

  In addition, a technique called pattern matching that is conventionally used to detect a marker consumes a lot of time for calculation. If parallel processing is used to prevent this, there is a problem that the circuit scale increases.

  SUMMARY OF THE INVENTION An object of the present invention is to provide a digital image position adjusting apparatus and method, and a data recording / reproducing apparatus for performing alignment from a reproduced digital image instead of a marker.

  In order to solve the above-described conventional problems, the digital image position adjustment apparatus of the present invention interpolates each pixel value using a vertical tap coefficient for a two-dimensional digital image in which each pixel is quantized into multiple values. Performing vertical interpolation means for correcting a vertical position shift of the digital image and outputting a vertical interpolation digital image, and interpolating each pixel value using a horizontal tap coefficient for the vertical interpolation digital image, A horizontal interpolation unit that outputs a vertical / horizontal interpolated digital image by correcting a positional deviation in the horizontal direction of the vertical interpolated digital image, and detects a horizontal phase error by comparing each pixel value of the vertical / horizontal interpolated digital image with a horizontal target value. A horizontal error detecting means for comparing each pixel value of the vertical and horizontal interpolation digital image with a vertical target value to detect a vertical phase error; and the horizontal phase. A horizontal loop filter that smoothes the difference and outputs a frequency error in the horizontal direction, a vertical loop filter that smoothes the vertical phase error and outputs a frequency error in the vertical direction, and a horizontal direction output from the horizontal loop filter Horizontal frequency phase converting means for outputting the horizontal tap coefficient from the frequency error of the vertical frequency phase, and vertical frequency phase converting means for outputting the vertical tap coefficient from the frequency error in the vertical direction output from the vertical loop filter. It is what.

  Further, in the digital image position adjusting device, a two-dimensional digital image in which each pixel is quantized to multiple values is equalized to a frequency characteristic adapted to the positional deviation correction in the horizontal direction, and a horizontal equalized digital image is output. Equalization means, vertical equalization means for equalizing the frequency characteristics of the digital image in accordance with vertical position correction and outputting a vertical equalization digital image, the digital image, and the horizontal equalization digital A vertical interpolation digital image obtained by interpolating each pixel value using a vertical tap coefficient and correcting a vertical position shift, and a horizontal equalization_vertical interpolation digital image; Vertical equalization_vertical interpolation means for outputting a vertical interpolation digital image, the vertical interpolation digital image, the horizontal equalization_vertical interpolation digital image, and vertical equality _Vertical and horizontal interpolation digital image obtained by interpolating each pixel value using a horizontal tap coefficient and correcting a positional deviation in the horizontal direction, and horizontal equalization_vertical and horizontal interpolation digital image and vertical equalization_ A horizontal interpolation means for outputting a vertical / horizontal interpolation digital image; a horizontal error detection means for detecting a horizontal phase error by comparing each pixel value of the horizontal equalization_vertical / horizontal interpolation digital image with a horizontal target value; Equalization_Vertical error detecting means for detecting a vertical phase error by comparing each pixel value of a vertical / horizontal interpolation digital image with a vertical target value, and smoothing the horizontal phase error and outputting a horizontal frequency error A horizontal loop filter; a vertical loop filter that smoothes the vertical phase error and outputs a vertical frequency error; and a horizontal frequency error output from the horizontal loop filter A horizontal frequency phase converting means for outputting a tap coefficient, and a vertical frequency phase converting means for outputting the vertical tap coefficient from a vertical frequency error outputted from the vertical loop filter. .

  Further, in the digital image position adjustment apparatus, a vertical two-dimensional digital image in which each pixel is quantized to multiple values is equalized to a frequency characteristic adapted to the positional deviation correction in the vertical direction, and a vertical equalized digital image is output. An equalization means, the digital image, and a vertical interpolation digital image in which each pixel value is interpolated using a vertical tap coefficient to correct a vertical position shift, and vertical equalization. _Vertical interpolation means for outputting a vertically interpolated digital image, and a series of the vertically interpolated digital images that are sequentially output from the vertical interpolator as the horizontal reference position changes, are input in accordance with the lateral displacement correction. Horizontal equalization means for equalizing the frequency characteristics and outputting horizontal equalization_vertical interpolation digital image, the vertical interpolation digital image, and vertical equalization_vertical interpolation digital image. And horizontal equalization_vertical interpolation digital image is input, each pixel value is interpolated using a horizontal tap coefficient, and the horizontal misalignment digital image is corrected for lateral displacement, and horizontal equalization_ A horizontal interpolation means for outputting a vertical / horizontal interpolation digital image and a vertical equalization_vertical / horizontal interpolation digital image; and comparing each pixel value of the horizontal equalization_vertical / horizontal interpolation digital image with a horizontal target value to determine a horizontal phase error. A horizontal error detecting means for detecting; a vertical error detecting means for detecting a vertical phase error by comparing each pixel value of the vertical equalization_vertical / horizontal interpolation digital image with a vertical target value; and smoothing the horizontal phase error. A horizontal loop filter that outputs a frequency error in the horizontal direction, a vertical loop filter that smoothes the vertical phase error and outputs a frequency error in the vertical direction, and a horizontal loop filter that is sequentially output from the horizontal loop filter. A horizontal frequency phase conversion means for outputting a horizontal interpolation position by inputting a frequency error in a horizontal direction and performing a frequency phase conversion; and a horizontal frequency coefficient for outputting the horizontal tap coefficient from a horizontal frequency error output from the horizontal loop filter. A frequency phase conversion unit and a vertical frequency phase conversion unit that outputs the vertical tap coefficient from a vertical frequency error output from the vertical loop filter are provided.

  Further, in the digital image position adjusting device, each pixel value is interpolated using a vertical tap coefficient for a two-dimensional digital image in which each pixel is quantized to multiple values, and the vertical displacement of the digital image is determined. A vertical interpolation means for correcting and outputting a vertical interpolation digital image, and a series of vertical interpolation digital images sequentially output as the horizontal reference position changes from the vertical interpolation means, are input in accordance with the positional deviation correction in the horizontal direction. Horizontal equalization means for equalizing frequency characteristics and outputting horizontal equalization_vertical interpolation digital images, and a series of vertical interpolation digital images sequentially output from the vertical interpolation means as the horizontal reference position changes. Vertical equalization means for outputting a vertical equalization_vertical interpolation digital image equalized to a frequency characteristic in accordance with vertical position correction, the vertical interpolation digital image, Horizontal equalization_vertical interpolation digital image and each series of vertical equalization_vertical interpolation digital image are input, each pixel value is interpolated using horizontal tap coefficient, and vertical / horizontal interpolation in which the horizontal displacement is corrected Horizontal interpolation means for outputting a digital image, horizontal equalization_vertical / horizontal interpolation digital image, vertical equalization_vertical / horizontal interpolation digital image, and each pixel value of the horizontal equalization_vertical / horizontal interpolation digital image as a horizontal target value A horizontal error detecting means for detecting a horizontal phase error by comparing, and a vertical error detecting means for detecting a vertical phase error by comparing each pixel value of the horizontal equalization_vertical / horizontal interpolation digital image with a vertical target value; A horizontal loop filter for smoothing the horizontal phase error and outputting a frequency error in the horizontal direction; a vertical loop filter for smoothing the vertical phase error and outputting a frequency error in the vertical direction; A horizontal frequency phase conversion means for outputting the horizontal tap coefficient from a horizontal frequency error output from the filter, and a vertical frequency phase conversion for outputting the vertical tap coefficient from the vertical frequency error output from the vertical loop filter. And horizontal reference position output means for sequentially outputting the horizontal coordinate of the digital image as the horizontal reference position.

  Further, in the digital image position adjusting device, each pixel value is interpolated using a vertical tap coefficient for a two-dimensional digital image in which each pixel is quantized to multiple values, and the vertical displacement of the digital image is determined. A vertical interpolation means for correcting and outputting a vertical interpolation digital image, and a series of vertical interpolation digital images sequentially output as the horizontal reference position changes from the vertical interpolation means, are input in accordance with the vertical misalignment correction. Vertical equalization means for equalizing frequency characteristics and outputting vertical equalization_vertical interpolation digital image, the vertical interpolation digital image, and the vertical equalization_vertical interpolation digital image series are input, and horizontal taps are input. A horizontal interpolation unit that interpolates each pixel value using a coefficient and outputs a vertical / horizontal interpolation digital image and a vertical equalization_vertical / horizontal interpolation digital image; Horizontal equalization means for inputting a series of vertical and horizontal interpolated digital images to be output next, equalizing them to frequency characteristics in accordance with lateral displacement correction, and outputting horizontal equalization_vertical and horizontal interpolation digital images, and the horizontal equalization _Horizontal error detection means for comparing each pixel value of the vertical / horizontal interpolation digital image with a horizontal target value to detect a horizontal phase error; and vertical equalization_each pixel value of the vertical / horizontal interpolation digital image as a vertical direction target value. A vertical error detecting means for detecting a vertical phase error by comparison, a horizontal loop filter for smoothing the horizontal phase error and outputting a frequency error in the horizontal direction, and smoothing the vertical phase error, A vertical loop filter that outputs a frequency error, a horizontal frequency phase conversion means that outputs the horizontal tap coefficient from a horizontal frequency error output from the horizontal loop filter, and the vertical loop filter Vertical frequency phase conversion means for outputting the vertical tap coefficient from the vertical frequency error output from the horizontal direction, and horizontal reference position output means for sequentially outputting the horizontal coordinate of the digital image as the horizontal reference position. It is what.

  Further, in the digital image position adjusting device, each pixel value is interpolated using a vertical tap coefficient for a two-dimensional digital image in which each pixel is quantized to multiple values, and the vertical displacement of the digital image is determined. Vertical interpolation means for correcting and outputting a vertically interpolated digital image, and using the horizontal tap coefficient for the vertically interpolated digital image, correcting the horizontal misalignment of the first corrected digital image, Horizontal interpolation means for outputting an interpolated digital image; and vertical equalization means for equalizing the vertical and horizontal interpolated digital image to frequency characteristics adapted to correction of positional deviation in the vertical direction and outputting a vertical equalization_vertical and horizontal interpolation digital image; The horizontal and vertical interpolated digital image is equalized to the frequency characteristics adapted to the positional deviation correction in the horizontal direction to the horizontal and vertical interpolated digital image. Horizontal equalizing means for outputting, horizontal error detecting means for detecting a horizontal phase error by comparing each pixel value of the horizontal equalization_vertical / horizontal interpolation digital image with a horizontal target value, and the vertical equalization_vertical / horizontal interpolation A vertical error detection means for detecting a vertical phase error by comparing each pixel value of the digital image with a vertical target value, a horizontal loop filter for smoothing the horizontal phase error and outputting a horizontal frequency error, A vertical loop filter that smoothes the vertical phase error and outputs a vertical frequency error; and a horizontal frequency phase conversion means that outputs the horizontal tap coefficient from a horizontal frequency error output from the horizontal loop filter; Vertical frequency phase conversion means for outputting the vertical tap coefficient from a vertical frequency error output from the vertical loop filter, and the horizontal coordinate of the digital image as the horizontal reference position Is characterized in further comprising a lateral reference position output means for following output.

  Further, in the digital image position adjusting apparatus, the horizontal equalization means and the vertical equalization means are means for equalizing to partial response class 4.

  Further, in the digital image position adjusting device, the vertical equalizing means and the horizontal equalizing means are constituted by the same two-dimensional filter.

  Furthermore, in the digital image position adjusting device, the vertical interpolation means and the horizontal interpolation means are constituted by the same two-dimensional filter.

  Further, in the digital image position adjusting apparatus, the vertical frequency phase conversion means outputs a plurality of lines of vertical interpolation positions simultaneously, and the vertical interpolation means and the horizontal interpolation means can process a plurality of lines simultaneously.

  Further, in the digital image position adjusting device, the vertical frequency phase conversion means outputs a plurality of rows of vertical interpolation positions simultaneously, and the vertical interpolation means, the horizontal interpolation means, the vertical equalization means, and the horizontal equalization means have a plurality of rows. Can be processed simultaneously.

  Further, in the digital image position adjusting apparatus, the vertical frequency phase conversion means outputs a plurality of lines of vertical interpolation positions simultaneously, and at least one of the vertical error detection means and the horizontal error detection means outputs an average of errors of the plurality of lines. It is characterized by doing.

  Further, in the digital image position adjusting apparatus, the vertical frequency phase conversion means outputs M rows of vertical interpolation positions represented by a natural number M simultaneously, and at least one of the vertical error detection means and the horizontal error detection means is a natural number N. An error of N rows represented is averaged and output, and M> N.

  Furthermore, in the data recording / reproducing apparatus of the present invention, a spatial modulator that spatially modulates coherent light, and interference of the light output from the spatial modulator is recorded on the medium and the reproduced light is read from the medium, or in advance An interference optical system for reading the interference of light recorded on the medium as reproduction light, an image sensor for photoelectrically converting reproduction light read from the interference optical system, and digital read by the image sensor 7. A digital image position adjusting apparatus according to claim 1, further comprising: a signal processing unit that inputs an image and performs signal processing, wherein the signal processing unit performs alignment of at least the digital image. And binarization means for binarizing the digital image aligned by the digital image position adjusting device. It is characterized in that to obtain.

  Further, in the data recording / reproducing apparatus, at least a part of the modulation pattern of the spatial modulator is provided with a pull-in pattern for horizontal phase synchronization and a pull-in pattern for vertical phase synchronization. It is what.

  Furthermore, in the data recording / reproducing apparatus, the pull-in pattern added to the modulation pattern of the spatial modulator is added with vertical pull-in pattern data for vertical phase synchronization over a plurality of columns, and horizontal phase synchronization. A horizontal pull-in pattern is added over a plurality of lines.

  Further, in the digital image position adjustment method of the present invention, each pixel value is interpolated using a vertical tap coefficient for a two-dimensional digital image in which each pixel is quantized to multiple values, and the vertical direction of the digital image is adjusted. Correcting the misalignment and outputting a vertically interpolated digital image; and correcting the misalignment in the horizontal direction of the first corrected digital image using a lateral tap coefficient for the vertically interpolated digital image. A step of outputting a vertical / horizontal interpolation digital image, a step of comparing each pixel value of the vertical / horizontal interpolation digital image with a horizontal target value to detect a horizontal phase error, and a vertical direction of each pixel value of the vertical / horizontal interpolation digital image A step of detecting a longitudinal phase error in comparison with a target value; a step of smoothing the lateral phase error and outputting a lateral frequency error; Smoothing the vertical phase error and outputting a vertical frequency error, outputting the horizontal tap coefficient from the horizontal frequency error, and calculating the vertical frequency error from the vertical frequency error. And a step of outputting a tap coefficient.

  Further, in the digital image position adjustment method, the step of equalizing to the frequency characteristic adapted to the lateral displacement correction prior to the step of outputting the lateral frequency error is adapted to the lateral displacement correction. And a step of equalizing the frequency characteristics.

  According to the digital image adjusting apparatus and method and the data recording / reproducing apparatus of the present invention, the phase error of the reproduced digital image is detected and corrected, and the alignment without relying on the marker is performed. It is resistant to deterioration, lens distortion, and rotation of the input image, enabling accurate alignment and improving the error rate.

  In addition, since the calculation amount is small and sequential control is performed as compared with the pattern matching used conventionally, circuitization is easy, and the circuit scale can be suppressed.

  Embodiments of a digital image position adjusting apparatus and method and a data recording / reproducing apparatus according to the present invention will be described below in detail with reference to the drawings.

  FIG. 1 shows the configuration of the alignment apparatus of this embodiment. The memory 101 stores a two-dimensional digital image (initial digital image). The vertical equalization means 103 reads the initial digital image from the memory 101 via the memory interface 102, performs equalization that facilitates phase synchronization in the vertical direction with respect to the initial digital image, and performs the vertical equalization digital image. Is generated. At the same time, the horizontal equalization means 104 reads the initial digital image from the memory 101 via the memory interface 102, performs equalization that facilitates phase synchronization in the horizontal direction with respect to the initial digital image, and performs horizontal equalization. Generate a digital image. The vertical equalizing means 103 and the horizontal equalizing means 104 store the vertical equalized digital image and the horizontal equalized digital image in the memory 101 via the memory interface 102.

  The vertical interpolation means 105 performs the vertical digital pixel, vertical equalization digital image, and horizontal equalization digital image pixels stored in the memory 101 in the vertical direction according to a horizontal reference position set by a horizontal reference position output means (not shown). The number of rows necessary for the direction interpolation is sequentially read out, and the pixels sequentially read out from the respective digital data are interpolated in the vertical direction according to the vertical interpolation position set from the vertical frequency phase conversion means 112. Calculate the appropriate position value. At this time, the interpolated value of each digital image that is sequentially output, the value corresponding to the initial digital image is called the vertical interpolation value, the value corresponding to the vertical equalized digital image is called the vertical equalization_vertical interpolation value, The value corresponding to the horizontal equalized digital image is called horizontal equalization_vertical interpolation value. Each vertical interpolation value sequentially output from the vertical interpolation unit 105 is input to the horizontal interpolation unit 106.

  The horizontal interpolation unit 106 calculates values at appropriate positions in the vertical and horizontal directions by interpolating the series of vertical interpolation values sequentially input according to the horizontal interpolation position set from the horizontal frequency phase conversion unit 111. At this time, the interpolated value of each digital image that is sequentially output, the value corresponding to the initial digital image is called the vertical and horizontal interpolation value, the value corresponding to the vertical equalized digital image is called the vertical equalization_vertical and horizontal interpolation value, The value corresponding to the horizontal equalized digital image is called horizontal equalization_vertical / horizontal interpolation value. The series of vertical / horizontal interpolation values sequentially output from the horizontal interpolation means 106 is a series for which alignment has been completed, and is stored in the memory 101 via the memory interface 102. A series of horizontal equalization_vertical / horizontal interpolation values sequentially output from the horizontal interpolation unit 106 is input to the horizontal error detection unit 107. A series of vertical equalization_vertical / horizontal interpolation values sequentially output from the horizontal interpolation unit 106 is input to the vertical error detection unit 108.

  The horizontal error detection means 107 detects a horizontal phase error from the series of horizontal equalization_vertical / horizontal interpolation values, and outputs it to the horizontal loop filter 109. Further, the vertical error detection means 108 detects a vertical phase error from the vertical equalization_vertical / horizontal interpolation value series and outputs the phase error to the vertical loop filter 110.

  The horizontal loop filter 109 calculates a horizontal frequency error from a horizontal phase error by a PI (Proportion Integral) filter. The vertical loop filter 110 calculates a vertical frequency error from a vertical phase error by a PI filter.

  The horizontal frequency phase conversion means 111 obtains a horizontal interpolation position (phase information) from the horizontal frequency error and the initial frequency error output from the horizontal loop filter 109, and the horizontal interpolation position (phase information) is determined as the horizontal interpolation position (tap). Coefficient). Further, the vertical frequency phase conversion means 112 obtains a vertical interpolation position (phase information) from the vertical frequency error and the initial frequency error output from the vertical loop filter 110, and determines the vertical interpolation position (phase information) as the vertical interpolation position. Convert to (tap coefficient).

  The horizontal frequency phase conversion unit 111 outputs the horizontal interpolation position (tap coefficient) to the horizontal interpolation unit 106. Further, the vertical frequency phase conversion unit 112 outputs the vertical interpolation position (tap coefficient) to the vertical interpolation unit 105.

  As described above, ITR (Interpolation Timing Recovery) that configures PLL (Phase Locked Loop) using interpolation has a double loop configuration, so that resampling can be performed at appropriate positions in the vertical and horizontal directions. . Hereinafter, each component will be described in detail.

  The memory 101 is a memory for storing a digital image, and is configured by an SDRAM (Synchronous Dynamic Random Access Memory) or the like. The memory interface 102 mediates reading and writing of digital images between the memory 101 and other components.

  Now, the initial digital image of FIG. 3A corresponding to the two-dimensional data of FIG. 2 (black is 0, white is 1) is stored in the memory 101. The initial digital image in FIG. 3A is, for example, a digital image obtained by photographing the two-dimensional data in FIG. 2 with a monochrome camera. When the number of pixels of the camera is 1.5 × 1.5 times the two-dimensional data in FIG. 2, an initial digital image as shown in FIG. 3A can be taken. Note that the image of FIG. 3A is a binarized image by applying error diffusion for display, and is actually a multi-valued image.

  The vertical equalization means 103 and the horizontal equalization means 104 respectively filter the initial digital image stored in the memory 101 so that phase synchronization in the vertical and horizontal directions is facilitated. The vertical equalization unit 103 and the horizontal equalization unit 104 are configured by FIR (Finite Impulse Response) filters in the spatial coordinate directions in the vertical direction and the horizontal direction, respectively. For example, when the horizontal equalization means 104 is configured with a 5-tap FIR filter, the value of the pixel after equalization corresponding to a certain pixel P (x, y) in the initial digital image of FIG. As shown in FIG. 3B, it is obtained by applying filter coefficients using pixels P (x−2, y) to P (x + 2, y) as tap inputs.

  The filter coefficients of the vertical equalization unit 103 and the horizontal equalization unit 104 will be described. As described above, the vertical equalizing means 103 and the horizontal equalizing means 104 perform filtering to facilitate phase synchronization. A sequence that can be easily phase-synchronized is a sequence that includes many points that cross the center line, called zero-cross points. When the characteristics of the initial digital image in the spatial coordinate direction are close to the characteristics of PR (1) shown in FIG. 4, it is difficult to detect the zero cross point. However, by performing equalization in the spatial coordinate direction, When filtering to the characteristic of PR (1, 0, −1) in FIG. 4, a zero-cross point that is easy to detect is generated like a point C_b in FIG. 4. When the initial digital image is close to the PR (1) characteristic, the digital coefficient can be changed to the PR (1, 0, -1) characteristic by setting the filter coefficient to PR (1, 0, -1). I can do it. PR (1, 0, -1) is called partial response class 4, and it is said that making a sequence a characteristic of PR (1, 0, -1) is equalized to partial response class 4 (PR4). Since the partial response class 4 can remove low-frequency noise in addition to high frequency, it exhibits a noise removal effect when the low-frequency noise of the digital image is large.

  Now, the initial digital image is close to the PR (1) characteristic, and the two-dimensional data in FIG. 2 is sampled 1.5 times in the spatial coordinate direction. Therefore, by using the values of the points C_i, C_j, C_k, C_l, and C_m shown in FIG. 4 as 5-tap filter coefficients, the initial digital image can have PR (1, 0, −1) characteristics. It is possible to increase the number of zero cross points. In the case of 1.0 times sampling, the 3-tap filter coefficients C_a, C_b, and C_c in FIG. 4 may be used. The vertical equalization means 103 and the horizontal equalization means 104 are 5-tap FIR filters having filter coefficients of points C_i, C_j, C_k, C_l, and C_m in FIG. The vertical equalizing means 103 equalizes the initial digital image of FIG. 3A in the vertical direction, and the horizontal equalizing means 104 equalizes the initial digital image of FIG. 3A in the horizontal direction. The vertical equalized digital image and the horizontal equalized digital image, which are digital images equalized by the vertical equalizing means 103 and the horizontal equalizing means 104, are stored in the memory 101 via the memory interface 102, respectively.

  The vertical interpolation means 105 interpolates the digital image in the vertical direction and obtains appropriate sample points in the vertical direction. The configuration of the vertical interpolation means 105 is shown in FIG. Similar to the vertical equalization means 103, the vertical interpolation means 105 is composed of an FIR filter in the spatial coordinate direction. The vertical interpolation means 105 is a filter that slides the phase in the vertical direction of the spatial coordinates, and the interpolation position (tap coefficient) that is information on the sliding phase is set from the vertical frequency phase conversion means 112. When the number of taps of the vertical interpolation means 105 is 6, to calculate one appropriate value in the vertical direction, vertical x horizontal: 6 × 1 pixels are required.

  The vertical interpolation means 105 reads 6 × 1 pixels at a time in order to calculate a vertical interpolation value that is an appropriate value in the vertical direction of the initial digital image. Further, the vertical interpolation means 105 reads 6 × 1 pixels at a time in order to calculate a horizontal equalization_vertical interpolation value which is an appropriate value in the vertical direction of the horizontal equalized digital image. The vertical interpolation means 105 detects a phase error unlike the other two digital images in order to calculate a vertical equalization_vertical interpolation value which is an appropriate value in the vertical direction of the vertical equalization digital image. In addition, it is necessary to read out not only one appropriate value in the vertical direction but also an appropriate value in the vertical direction on one line or an appropriate value in the vertical direction one line below. Accordingly, an appropriate value in the vertical direction (vertical equalization_vertical interpolation value), an appropriate value in the vertical direction on one line (vertical equalization_vertical interpolation value on one line), and an appropriate value in the vertical direction below one line. When using a filter with 6 taps for each of the three values (vertical equalization_vertical interpolation value in one row), it is necessary to read out 6 * 3 = 18 pixels simultaneously.

  However, since the vertical equalization_vertical interpolation value, vertical equalization on the first line_vertical interpolation value, vertical equalization on the first line_vertical interpolation value, a total of 18 pixels necessary for obtaining the vertical interpolation value include the same pixel. For example, in the case of 1.5 times sampling, the number of pixels required to obtain three values may be eight or nine. Therefore, a total of 21 pixels of 6 × 1 pixels, 6 × 1 pixels, and 9 × 1 pixels may be read at a time from the initial digital image, horizontal equalized digital image, and vertical equalized digital image, respectively. Using these 21 pixels, a vertical interpolation value that is a value of an appropriate vertical position corresponding to the initial digital image, and a horizontal equalization_vertical value that is a value of an appropriate vertical position corresponding to the horizontal equalized digital image are used. Interpolation value, vertical equalization value corresponding to vertical equalized digital image, vertical equalization value_vertical interpolation value, appropriate vertical position corresponding to vertical equalization digital image Five values of vertical equalization_vertical interpolation value on one line, vertical equalization_vertical interpolation value on one line that is the value of an appropriate position one line below the vertical direction corresponding to the vertical equalized digital image Can be generated.

  As shown in FIG. 3A, from the upper left to the lower right, the first row, the second row,... The Nth row, the horizontal reference position output means (not shown) The coordinates from to are sequentially output. The horizontal reference position output means (not shown) counts to the right end of the digital image, for example, 0, 1, 2, 3,... For the first line, and then changes the line to 0, 1, 2, for the second line. 3... In accordance with a horizontal reference position output from a horizontal reference position output unit (not shown), the vertical interpolation unit 105 sequentially reads out 21 pixels from the memory 101 via the interface 102, and performs vertical interpolation values, horizontal equalization_vertical interpolation values, Five values of equalization_vertical interpolation value, vertical equalization on the first line_vertical interpolation value, vertical equalization on the first line and vertical interpolation value are sequentially output. Vertical interpolation value, horizontal equalization_vertical interpolation value is vertical equalization_vertical interpolation value, vertical equalization over one line_vertical interpolation value, vertical equalization under one line_vertical interpolation value is the initial digital image, respectively , A series of pixels that can be aligned only in the vertical direction of the horizontal equalized digital image and the vertical equalized digital image. Vertical interpolation value sequentially output from the vertical interpolation means 105, horizontal equalization_vertical interpolation value, vertical equalization_vertical interpolation value, vertical equalization on one line_vertical interpolation value, vertical equalization below one line_vertical Each series of interpolation values is input to the horizontal interpolation means 106.

  The horizontal interpolation unit 106 performs alignment in the horizontal direction further for the pixel series of each digital image that has been aligned in the vertical direction by the vertical interpolation unit 105. The configuration of the horizontal interpolation means 106 is the same as that of the vertical interpolation means 105 as shown in FIG. The operation performed by the horizontal interpolation unit 106 is almost the same as the operation performed by the vertical interpolation unit 105. The interpolation position (tap coefficient) set by the horizontal frequency phase conversion unit 111 is set to a vertical interpolation value, horizontal equalization_vertical interpolation value, Vertical equalization_Vertical interpolation value, Vertical equalization on one line_Vertical interpolation value, Vertical equalization below one line_Vertical equalization_Vertical interpolation value can be aligned in the horizontal direction. Interpolation value, horizontal equalization_vertical / horizontal interpolation value, vertical equalization_vertical / horizontal interpolation value, vertical equalization on one line_vertical / horizontal interpolation value, vertical equalization below one line_vertical / horizontal interpolation value series are calculated and output. . Since the vertical / horizontal interpolation value is a series corresponding to the initial digital image in which the alignment has been completed both vertically and horizontally, it is sequentially stored in the memory 101 as a series of output digital images. A series of horizontal equalization_vertical / horizontal interpolation values is sequentially output to the horizontal error detection means 107 in order to detect a horizontal phase error. Vertical equalization_Vertical / horizontal interpolation value, Vertical equalization on one line_Vertical / horizontal interpolation value, Vertical equalization_1 vertical line_Vertical / horizontal interpolation value is sent to the vertical error detection means 108 in order to detect a phase error in the vertical direction. Output sequentially.

  The horizontal error detection means 107 sequentially calculates a horizontal phase error with respect to the center line (horizontal target value) from a sequence of horizontal equalization_vertical / horizontal interpolation values, which is a sequence of horizontally equalized digital images that are sequentially input. Output. A method for detecting the phase error will be described. FIG. 6 shows a series in the case where the alignment is completed (the phase error is small). FIG. 6A shows the pixels of the digital image that have been aligned. This value is higher for white. In FIG. 6B, the transition of the luminance of the pixel series of the digital image whose alignment has been completed is sequentially shown from the left to the right, and corresponds to FIG. FIG. 6C shows a pixel series obtained by equalizing FIG. 6B to the partial response class 4. As can be seen from FIG. 6C, the pixels near the center line do not vary.

  Next, look at FIG. FIG. 7 also shows a series in the case where the alignment is not performed (the phase error is large), in addition to the case where the alignment is performed (the phase error is small). Pixels having a phase shift + θ vary in the vicinity of the center line. The phase shift point A is shifted considerably above the center line. The phase shift point B is shifted below the center line. As described above, the shift in the phase direction can be recognized by the shift in the amplitude direction. Pixels in the vicinity of the center line are called zero cross points, which are detected by the threshold shown in FIG. If the detected zero-cross point is a position with a positive slope when viewed from the left to the right (phase shift point A or the like), the amplitude is detected as a phase error as it is, and the zero-cross point is a position with a negative slope ( If it is a phase shift point B or the like, the amplitude is multiplied by a minus sign and detected as a phase error. The lateral phase error calculated by the lateral error detection means 107 is output to the lateral loop filter 109.

  The vertical error detection means 108 is a series of vertically equalized digital images that are sequentially input. Vertical phase with respect to the center line (vertical target value) from three series of vertical equalization_vertical interpolation value, vertical equalization on one line_vertical / horizontal interpolation value, vertical equalization below one line_vertical / horizontal interpolation value Errors are sequentially calculated and output, and can be detected with substantially the same configuration as the lateral error detection means 107. However, although the horizontal error detection means 107 determines the sign of the phase error by determining the plus or minus of the slope of the transition of the series that is sequentially input, the vertical error detection means 108 is the vertical input that is input simultaneously. It is determined whether the slope of the three values of the vertical_horizontal interpolation value, vertical equalization on the first row_vertical / horizontal interpolation value, and vertical equalization_vertical / horizontal interpolation value on the first row is positive or negative. This is for determining the inclination in the vertical direction. The vertical phase error calculated by the vertical error detection means 108 is output to the vertical loop filter 110.

  The horizontal loop filter 109 and the vertical loop filter 110 respectively convert the vertical and horizontal phase errors through the PI filter to convert them into vertical and horizontal frequency errors. The configuration of the horizontal loop filter 109 and the vertical loop filter 110 is shown in FIG. The loop filter is configured to multiply the Proportion and the Integral by a gain and add them by an adder. Proportion is multiplied by gain A. In Integral, gain B is multiplied by the integrator output. The vertical and horizontal frequency errors calculated by the horizontal loop filter 109 and the vertical loop filter 110 are output to the horizontal frequency phase conversion unit 111 and the vertical frequency phase conversion unit 112, respectively.

  The horizontal frequency phase conversion unit 111 and the vertical frequency phase conversion unit 112 add the vertical and horizontal frequency errors and the initial frequency error, and perform frequency_phase conversion to obtain an interpolation position (phase information). Then, the interpolation position (phase information) is converted into an interpolation position (tap coefficient) and output. The configuration diagram is shown in FIG. An adder 901 adds the initial frequency error to the frequency error. The initial frequency error is added when there is a steady frequency error. When 1.5 times oversampling is performed, a value corresponding to 1.5 is added to the frequency error. After adding the initial frequency error, it is input to the integrator 902. It is generally known that frequency-phase conversion can be obtained by integration. The interpolation position (phase information) calculated by the integrator 902 is input to the coefficient selection unit 903. Coefficient selection means 903 selects and outputs an interpolation position (tap coefficient) corresponding to the input interpolation position (phase information). The coefficient selection unit 903 stores a plurality of interpolation positions (tap coefficients) corresponding to a plurality of interpolation positions (phase information) in advance, and selects an interpolation position (tap coefficient) corresponding to the input interpolation position (phase information). You can choose. The interpolation position (tap coefficient) stored in advance in the coefficient selection means 903 is a tap coefficient for sliding the phase, and a filter coefficient generally called a Nyquist filter may be used. The vertical and horizontal interpolation positions (tap coefficients) calculated by the horizontal frequency phase conversion unit 111 and the vertical frequency phase conversion unit 112 are output to the vertical interpolation unit 105 and the horizontal interpolation unit 106.

  As described above, by performing alignment of digital images using a two-dimensional PLL, alignment can be performed with a sequential circuit, and a circuit can be easily realized with a small circuit scale. Also, instead of using some information such as markers for alignment, timing information can be extracted from each pixel of the digital image and spatial coordinates can be recovered, so there are some degraded parts. Does not significantly affect the alignment accuracy. In addition, even if there is a lens distortion or a digital image rotation, the alignment is performed while following, so that an accurate alignment is possible. A conceptual diagram following the rotation of the digital image is shown in FIG. Even when the angle θ is rotated, when reading from the left to the right, the effect of the vertical phase-locked loop can be interpolated to perform the appropriate resampling.

  In the present embodiment, the vertical equalization means 103 and the horizontal equalization means 104 are processed before the vertical interpolation means 105 as shown in FIG. 1, but as shown in FIG. The processing of the equalization means 104 may be performed by the second horizontal equalization means 1001 after the vertical interpolation means 105. In this case, the horizontal equalized digital image output from the horizontal equalization means 104 to the memory 101 is not necessary, so that the memory capacity and the memory bandwidth can be reduced. The second horizontal equalization means 1001 receives the vertical interpolation values sequentially input from the vertical interpolation means 105, and equalizes the series with an FIR filter.

  In the present embodiment, the vertical equalization means 103 and the horizontal equalization means 104 are processed before the vertical interpolation means 105 as shown in FIG. 1, but as shown in FIG. The processing by the equalizing unit 103 and the horizontal equalizing unit 104 may be performed by the second vertical equalizing unit 1101 and the second horizontal equalizing unit 1001 after the vertical interpolation unit 105. In this case, the vertical equalized digital image output from the vertical equalizing means 103 to the memory 101 is not necessary, and therefore the memory capacity and the memory bandwidth can be reduced. Since the second vertical equalizing means 1101 is configured after the vertical interpolating means 105, sampling is performed 1.0 times in the vertical direction, so that the PR (1, 0, −1) and the like are equal. Conversion is performed by C_a, C_b, and C_c in FIG. 4 to form a 3-tap FIR filter. In this case, the vertical interpolation value sequentially input from the vertical interpolation means 105, the vertical interpolation value on the first line, the vertical interpolation value on the second line, the vertical interpolation value on the lower line, the vertical interpolation value on the lower line, Input 5 series, generate vertical equalization_horizontal interpolation value on 1 line from 3 values of vertical interpolation value on 1 line, vertical interpolation value on 1 line, vertical interpolation value, and on 1 line Generates vertical equalization_horizontal interpolation value from three values: vertical interpolation value, vertical interpolation value, vertical interpolation value in the first row, vertical interpolation value, vertical interpolation value in the lower row, vertical interpolation value in the upper row The vertical equalization_horizontal interpolation value one row below is generated from these three values, and the generated three sequences are sequentially output. Further, the second horizontal equalization means 1001 and the second vertical equalization means 1101 may be constituted by the same two-dimensional filter. A configuration example of the two-dimensional filter is shown in FIG.

  In the present embodiment, the vertical equalization means 103 and the horizontal equalization means 104 are processed before the vertical interpolation means 105 as shown in FIG. 1, but as shown in FIG. The equalizing means 103 may be performed by the second vertical equalizing means 1101 after the vertical interpolation means 105, and the horizontal equalizing means 104 may be performed by the third horizontal equalizing means 1201 after the horizontal interpolation means 106. In this case, since the number of sequences to be interpolated by the horizontal interpolation means 106 is reduced, the circuit scale can be reduced. Since the third horizontal equalization means 1201 is configured after the horizontal interpolation means 106, sampling is performed 1.0 times in the horizontal direction, so that the PR to (1, 0, −1) is equal. Conversion is performed by C_a, C_b, and C_c in FIG. 4 to form a 3-tap FIR filter. In this case, by inputting and equalizing a series of vertical and horizontal interpolation values sequentially input from the horizontal interpolation means 106, horizontal equalization_vertical and horizontal interpolation values are generated and sequentially output.

  In the present embodiment, the vertical equalization means 103 and the horizontal equalization means 104 are processed before the vertical interpolation means 105 as shown in FIG. 1, but as shown in FIG. The equalizing unit 103 may be performed by the third vertical equalizing unit 1301 after the horizontal interpolating unit 106, and the horizontal equalizing unit 104 may be performed by the third horizontal equalizing unit 1201 after the horizontal interpolating unit 106. In this case, since the number of sequences to be interpolated by the horizontal interpolation means 106 is reduced, the circuit scale can be reduced. Since the third vertical equalization means 1301 is configured after the horizontal interpolation means 106, sampling is performed 1.0 times in the horizontal direction, so that it is equal to PR (1, 0, −1), etc. Conversion is performed by C_a, C_b, and C_c in FIG. 4 to form a 3-tap FIR filter. In this case, vertical equalization_vertical / horizontal interpolation is performed by inputting and equalizing a series of vertical / horizontal interpolation values sequentially input from the horizontal interpolation means 106, vertical / horizontal interpolation values on one line, and vertical / horizontal interpolation values below one line. A value is generated and output sequentially. Further, the third horizontal equalization means 1201 and the third vertical equalization means 1301 may be constituted by the same two-dimensional filter. A configuration example of the two-dimensional filter is shown in FIG.

  As described above, equalization for facilitating phase synchronization performed by the horizontal equalization means 104, 1001, 1201 and the vertical equalization means 103, 1101, 1301 is similar to the horizontal error detection means 107 that detects the phase error. What is necessary is just to perform before the error detection means 108, and this invention is not limited by the position which performs equalization. The configurations of FIGS. 1 and 10 to 13 have different circuit scales and performances, and the designer may determine which one is to be superior.

  In this embodiment, the vertical interpolation unit 105 and the horizontal interpolation unit 106 are described as separate filters. However, the tap coefficients of the two are configured in a two-dimensional manner by convolution so that the vertical interpolation unit 105 and the horizontal interpolation unit are combined. It is also possible to configure 106 as a two-dimensional spatial filter. In this case, since the two interpolation processes described in this embodiment are performed by one circuit, the calculation speed can be improved. FIG. 14 shows a configuration example of the two-dimensional filter.

  In this embodiment, it has been described that the order of reading data is sequentially read in the first row, the second row,... N rows using FIG. A plurality of interpolated positions such as the 8th line are collectively output, and the vertical interpolation unit 105 reads a plurality of lines such as the 1st to 8th lines, and configures an alignment apparatus configured to be capable of parallel processing. May be. In this case, since the vertical interpolation calculation can be performed efficiently, the circuit scale does not increase with an increase in the calculation speed due to the parallelization, and the cost performance can be improved.

  When the above-mentioned plural lines are collectively read and aligned, the phase error calculated by the horizontal error calculating means 107 and the vertical error calculating means 108 is obtained by averaging the phase errors of the plural lines as a new phase error. Should be output as In this case, since the phase errors of a plurality of rows are averaged, the accuracy of the error is improved, and highly reliable alignment can be performed.

  When the above-mentioned plurality of M lines are collectively read and aligned, the phase error calculated by the horizontal error calculating means 107 and the vertical error calculating means 108 is obtained by averaging the phase errors of N lines. Good. At this time, M> N and N> = 1. This is a method of performing interpolation of a larger number of rows at a time with a phase error of a smaller number of rows. For example, an average of phase errors for three rows is output as a new phase error, and this is used. It is configured to perform interpolation for 8 lines. In this case, the number of rows required for error calculation is reduced, and the circuit scale can be reduced.

  The vertical and horizontal directions described in this embodiment may be interchanged. For example, the vertical equalizing means may be read as the horizontal equalizing means, and is limited by the reading direction of the digital data and the processing order of the vertical and horizontal directions. It is not something.

  FIG. 16 shows the configuration of the data recording / reproducing apparatus of the present embodiment. Light emitted from a coherent light source (not shown) is divided by a prism beam splitter (not shown), and one of the lights is irradiated to the spatial light modulator 161. The spatial light modulator 161 performs spatial modulation on the amplitude and phase of light. The data modulator 162 modulates the input sequence and outputs it to the spatial light modulator 161. The spatially modulated light is input to an interference optical system 163 for recording an interference pattern on the medium, and data is recorded on the medium 164. In the present embodiment, a method of recording on the medium 164 by causing two lights to interfere with each other by spatially modulating one of them divided by a prism beam splitter (not shown) as signal light and using the other as reference light. At the time of reproduction, only the reference light is input to the interference optical system 163, and the diffracted light is read from the medium 164 as reproduction light. The reproduction light from the medium 164 is irradiated to the image sensor 165, and the image sensor 165 outputs a digital image by photoelectric conversion. The digital image is binarized by the signal processing means 166. The coherent light source is a laser light source, and its wavelength bandwidth is very narrow in order to record interference.

  The spatial light modulator 161 that spatially modulates the amplitude and phase of light is composed of a liquid crystal panel or the like, and can set data “1” that is transmitted and data “0” that is blocked for each cell. As long as the amplitude of light can be spatially modulated, a DMD (Digital Mirror Device) or the like may be used, and the present invention is not limited by the ON / OFF structure of the spatial light modulator 161.

  The modulation pattern of the spatial light modulator 161 is set from the data modulator 162. The data modulator 162 applies, for example, a modulation in which only one cell out of four cells called “1-4 code (1 Light Bit on 4 cells Coding)” is “1” data to the input sequence, and is described in the second embodiment. Outputs a pattern with a phase pull-in pattern.

  The interference optical system 163 is an optical system that causes the light modulated by the spatial light modulator 161 to interfere with the reference light and records it on the medium 164. In this embodiment, light from a coherent light source (not shown) divided by a prism beam splitter (not shown) is used as reference light, and the other is applied to the spatial light modulator 161 and modulated light is used as signal light. Realizes light beam interference.

  The interference optical system 163 includes a Fourier transform lens and an aperture. The signal light collected by the Fourier transform lens 1631 is applied to the transmission medium 164. Further, the reference light is also applied to the medium 164. The interference between the two lights is recorded on the medium 164. The aperture 1633 removes high frequency components by restricting the aperture on the Fourier plane constructed by the Fourier transform lens 1631. The medium 164 is made of a material such as a photoreactive polymer and can record light interference. During reproduction, the medium 164 is irradiated with only reference light, and the diffracted light is subjected to inverse Fourier transform by the inverse Fourier transform lens 1632, and this light can be read from the medium 164 as reproduction light.

  The image sensor 165 photoelectrically converts the reproduction light read from the medium 164 to generate a digital image, and the digital image is input to the signal processing unit 166. The signal processing unit 166 performs at least alignment, spatial filtering, and binarization. I do. The output of the image sensor 165 is as shown in FIG. However, although there is no pull-in pattern in FIG. 3A, when the pull-in pattern is added, the pattern can also be photographed at the same time. The pull-in pattern will be described later. The configuration of the signal processing means 166 is shown in FIG.

  The digital image output from the image sensor 165 is stored in the memory 1662 via the memory interface 1661. The digital image stored in the memory 1662 is input to the alignment device 1663 via the memory interface 1661.

  The alignment device 1663 is the alignment device described in the first embodiment, applies a two-dimensional ITR to the digital image output from the image sensor 165, resamples it with appropriate coordinates, and stores it again in the memory 1662. Usually, the pixel pitch of the image sensor 165 does not correspond to the cell pitch of the spatial modulator 161 on a one-to-one basis. This is because one pixel may not be sensed for one cell due to the effects of lens aberration, media tilt, mechanical expansion and contraction, and the like. When one pixel cannot be sensed for one cell, the original image cannot be reproduced by the sampling theorem, and the number of errors increases. In order to prevent this, oversampling of, for example, 1.5 pixels is performed on one cell. The alignment device 1663 aligns the influence of aberration or the like, or sampling shifted due to oversampling so as to correspond to one pixel per cell.

  Next, a case where a pull-in pattern is added will be described. FIG. 18 shows an example of a data modulation method with a pull-in pattern. White represents 1 and black represents 0. In FIG. 18, a margin for phase pull-in is added to the left side of the two-dimensional data. To the left end, 4T continuous data in the vertical direction is added over a plurality of columns as the vertical pull-in pattern. Here, the 4T signal is an arrangement of data “1000”, and represents an arrangement of data “100” for 3T and “1010” for 2T2T. Further, on the right side of the vertical pull-in pattern, continuous 4T data in the horizontal direction is added as a horizontal pull-in pattern over a plurality of rows. Then, on the right side of the horizontal pull-in pattern, “1100” data that is 1T3T is added over a plurality of lines as a pull-in end pattern. Two-dimensional data is arranged on the right side of the drawing end pattern.

  In this way, by adding a pull-in pattern for phase pull-in to the modulation data, the pull-in range for phase synchronization can be expanded. The 4T continuous pattern is a sequence that has the largest number of zero crossings when equalized to the partial response class 4, and is a pattern suitable for detecting a phase error.

  When reading a digital image while performing phase synchronization from left to right, first, vertical phase synchronization is locked by a vertical pull-in pattern. Next, the horizontal phase synchronization is locked by the horizontal pull-in pattern. Thereafter, the pull-in end pattern is detected, data after the pull-in end pattern is resampled at an appropriate position, and alignment can be performed.

  Further, FIG. 19 shows a more efficient method of adding a 4T continuous pattern. FIG. 19 shows a pull-in pattern in which a horizontal 4T continuous pattern and a vertical 4T continuous pattern are efficiently combined. Since the efficiency is higher than the drawing pattern of FIG. 18, the area to which the drawing pattern is added may be narrow. Further, as shown in FIG. 20, a 4T pattern may be arranged, and a region to which a drawing pattern is added may be narrowed. In this way, by adding the vertical and horizontal phase acquisition patterns to the modulation data, the phase synchronization acquisition range can be expanded.

  The spatial filter 1664 is a filter for equalizing a digital image. Digital image equalization refers to filtering the spatial frequency characteristics of a digital image to predetermined characteristics so that data can be easily read. The spatial filter 1664 can be realized by a two-dimensional FIR filter. The filter coefficient of the spatial filter 1664 is set by a filter coefficient setting unit (not shown). The digital image filtered by the spatial filter 1664 is stored again in the memory 1662 via the memory interface 1661.

  Although the Nyquist filter for interpolation used in the alignment unit 1663 and the spatial filter 1664 are described as different units, the filter coefficients of both are convolved and the result is used as a new filter coefficient. By inputting to one filter, it is possible to realize both the Nyquist filter and the spatial filter 1664 for interpolation with one filter. In this case, there are effects such as reduction in circuit scale and memory access.

  The filtered digital image is input from the memory 1662 to the binarization unit 1665 via the memory interface 1661. The binarizing means 1665 demodulates 1-4 modulation, and then performs error correction and outputs data. The data output from the binarizing means 1665 is transferred to the host.

  As described above, by using the alignment measure 1663 using the two-dimensional ITR, the circuit scale can be suppressed, the alignment can be performed with high accuracy, and as a result, data reproduction with less errors and high cost performance can be performed. it can.

  Further, since the data to be recorded on the medium 164 is provided with a phase synchronization pull pattern by the data modulator 162, the phase lock pull range is large, and the playability can be improved.

  The digital image position adjusting apparatus and method and the data recording / reproducing apparatus according to the present invention are characterized by adjusting the position of the reproduced two-dimensional digital image without using a marker, and are useful as a hologram recording / reproducing apparatus or the like.

1 is a configuration diagram of an alignment apparatus in Embodiment 1 of the present invention. Fig. 2 is an image diagram of two-dimensional data in Embodiment 1 of the present invention. The digital image figure in Example 1 of this invention The figure explaining equalization in Example 1 of this invention Configuration diagram of interpolation means in Embodiment 1 of the present invention The figure explaining the waveform of a series without a phase error in Example 1 of the present invention. The figure explaining the waveform of a series with a phase error in Example 1 of the present invention. 1 is a configuration diagram of a loop filter according to a first embodiment of the present invention. Configuration diagram of frequency phase conversion means in Embodiment 1 of the present invention Configuration diagram of second alignment apparatus in Embodiment 1 of the present invention Configuration diagram of a third alignment apparatus in Embodiment 1 of the present invention Configuration diagram of a fourth alignment apparatus in Embodiment 1 of the present invention Configuration diagram of fifth alignment apparatus in Embodiment 1 of the present invention Configuration diagram of two-dimensional interpolation means in Embodiment 1 of the present invention Digital image diagram with rotation in Embodiment 1 of the present invention Configuration diagram of data recording / reproducing apparatus in Embodiment 2 of the present invention Configuration diagram of signal processing means in Embodiment 2 of the present invention Data image diagram having a phase pull-in pattern in Embodiment 2 of the present invention FIG. 6 is a data image diagram having a second phase pull-in pattern in Embodiment 2 of the present invention. Data image diagram having a third phase pull-in pattern in Embodiment 2 of the present invention

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 Memory 102 Memory interface 103 Vertical equalization means 104 Horizontal equalization means 105 Vertical interpolation means 106 Horizontal interpolation means 107 Horizontal error detection means 108 Vertical error detection means 109 Horizontal loop filter 110 Vertical loop filter 111 Horizontal frequency phase conversion means 112 Vertical frequency Phase conversion means 901 Adder 902 Integrator 903 Coefficient selection means 1001 Second horizontal equalization means 1101 Second vertical equalization means 1201 Third horizontal equalization means 1301 Third vertical equalization means 161 Spatial light modulator 162 Data Modulator 163 Interference Optical System 1631 Fourier Transform Lens 1632 Inverse Fourier Transform Lens 1633 Aperture 164 Medium 165 Image Sensor 166 Signal Processing Unit 1661 Memory Interface 1662 Memory 1663 Positioning Device 166 Spatial filter 1665 binarizing means

Claims (18)

A two-dimensional digital image in which each pixel is quantized into multiple values is interpolated with each pixel value using a vertical tap coefficient, and the vertical misalignment digital image is corrected by correcting the vertical displacement of the digital image. Vertical interpolation means to output;
Horizontal interpolation means for interpolating each pixel value using a horizontal tap coefficient for the vertical interpolation digital image, correcting a horizontal position shift of the vertical interpolation digital image, and outputting a vertical / horizontal interpolation digital image;
A horizontal error detection means for detecting a horizontal phase error by comparing each pixel value of the vertical and horizontal interpolation digital image with a horizontal target value;
Vertical error detection means for detecting a vertical phase error by comparing each pixel value of the vertical and horizontal interpolation digital image with a vertical target value;
A horizontal loop filter that smoothes the lateral phase error and outputs a lateral frequency error;
A longitudinal loop filter that smoothes the longitudinal phase error and outputs a longitudinal frequency error;
A lateral frequency phase converting means for outputting the lateral tap coefficient from a lateral frequency error output from the lateral loop filter;
Vertical frequency phase conversion means for outputting the vertical tap coefficient from a vertical frequency error output from the vertical loop filter;
A digital image position adjusting device comprising: A horizontal equalization means for equalizing the frequency characteristics in accordance with the positional deviation correction in the horizontal direction and outputting a horizontal equalized digital image to a two-dimensional digital image in which each pixel is quantized to multiple values;
Vertical equalization means for equalizing the digital image with frequency characteristics adapted to vertical positional deviation correction and outputting a vertical equalized digital image;
For the digital image, the horizontal equalized digital image, and the horizontal equalized digital image, a vertical interpolation digital image obtained by interpolating each pixel value using a vertical tap coefficient and correcting a positional deviation in the vertical direction, Vertical interpolation means for outputting horizontal equalization_vertical interpolation digital image and vertical equalization_vertical interpolation digital image;
The vertical interpolation digital image, the horizontal equalization_vertical interpolation digital image, and the vertical equalization_vertical interpolation digital image are interpolated for each pixel value using a horizontal tap coefficient to correct the positional deviation in the horizontal direction. Horizontal interpolation means for outputting the vertical / horizontal interpolation digital image, horizontal equalization_vertical / horizontal interpolation digital image, and vertical equalization_vertical / horizontal interpolation digital image;
Horizontal error detection means for detecting a horizontal phase error by comparing each pixel value of the horizontal equalization_vertical / horizontal interpolation digital image with a horizontal target value;
Vertical error detection means for detecting a vertical phase error by comparing each pixel value of the horizontal equalization_vertical / horizontal interpolation digital image with a vertical target value;
A horizontal loop filter that smoothes the lateral phase error and outputs a lateral frequency error;
A longitudinal loop filter that smoothes the longitudinal phase error and outputs a longitudinal frequency error;
A lateral frequency phase converting means for outputting the lateral tap coefficient from a lateral frequency error output from the lateral loop filter;
Vertical frequency phase conversion means for outputting the vertical tap coefficient from a vertical frequency error output from the vertical loop filter;
A digital image position adjusting device comprising: Vertical equalization means for equalizing the frequency characteristics according to the positional deviation correction in the vertical direction and outputting the vertical equalized digital image to the two-dimensional digital image in which each pixel is quantized to multiple values;
The digital image and the vertical equalization digital image are interpolated with pixel values using vertical tap coefficients to correct the vertical misalignment, and the vertical equalization_vertical interpolation digital image. Vertical interpolation means for outputting
The vertical interpolation digital image series sequentially output from the vertical interpolation means according to the change of the horizontal reference position is input, equalized to the frequency characteristics adapted to the positional deviation correction in the horizontal direction, and horizontal equalization_vertical interpolation digital Horizontal equalization means for outputting an image;
Each series of the vertical interpolation digital image, vertical equalization_vertical interpolation digital image, and horizontal equalization_vertical interpolation digital image is input, and each pixel value is interpolated using a horizontal tap coefficient to obtain a horizontal direction Horizontal interpolation means for outputting a vertical / horizontal interpolation digital image corrected for misalignment, horizontal equalization_vertical / horizontal interpolation digital image, and vertical equalization_vertical / horizontal interpolation digital image;
Horizontal error detection means for detecting a horizontal phase error by comparing each pixel value of the horizontal equalization_vertical / horizontal interpolation digital image with a horizontal target value;
Vertical error detection means for comparing each pixel value of the vertical equalization_vertical / horizontal interpolation digital image with a vertical target value to detect a vertical phase error;
A horizontal loop filter that smoothes the lateral phase error and outputs a lateral frequency error;
A longitudinal loop filter that smoothes the longitudinal phase error and outputs a longitudinal frequency error;
A horizontal frequency phase conversion means for inputting a horizontal frequency error sequentially output from the horizontal loop filter and performing a frequency phase conversion to output a horizontal interpolation position;
A lateral frequency phase converting means for outputting the lateral tap coefficient from a lateral frequency error output from the lateral loop filter;
Vertical frequency phase conversion means for outputting the vertical tap coefficient from a vertical frequency error output from the vertical loop filter;
A digital image position adjusting device comprising: A two-dimensional digital image in which each pixel is quantized into multiple values is interpolated with each pixel value using a vertical tap coefficient, and the vertical misalignment digital image is corrected by correcting the vertical displacement of the digital image. Vertical interpolation means to output;
A series of vertically interpolated digital images that are sequentially output from the vertical interpolating means in accordance with the change of the horizontal reference position is input, equalized to frequency characteristics adapted to the positional deviation correction in the horizontal direction, and horizontally equalized_vertically interpolated digital image Horizontal equalization means for outputting
A series of vertically interpolated digital images that are sequentially output from the vertical interpolating means in accordance with the change of the horizontal reference position is input, equalized to frequency characteristics adapted to the positional deviation correction in the vertical direction, and vertically equalized_vertically interpolated digital image Vertical equalization means for outputting
Each series of the vertical interpolation digital image, the horizontal equalization_vertical interpolation digital image, and the vertical equalization_vertical interpolation digital image is input, and each pixel value is interpolated using a horizontal tap coefficient, and the horizontal direction A horizontal interpolation means for outputting a vertical / horizontal interpolation digital image, a horizontal equalization_vertical / horizontal interpolation digital image, and a vertical equalization_vertical / horizontal interpolation digital image, in which the positional deviation is corrected,
Horizontal error detection means for detecting a horizontal phase error by comparing each pixel value of the horizontal equalization_vertical / horizontal interpolation digital image with a horizontal target value;
Vertical error detection means for detecting a vertical phase error by comparing each pixel value of the horizontal equalization_vertical / horizontal interpolation digital image with a vertical target value;
A horizontal loop filter that smoothes the lateral phase error and outputs a lateral frequency error;
A longitudinal loop filter that smoothes the longitudinal phase error and outputs a longitudinal frequency error;
A lateral frequency phase converting means for outputting the lateral tap coefficient from a lateral frequency error output from the lateral loop filter;
Vertical frequency phase conversion means for outputting the vertical tap coefficient from a vertical frequency error output from the vertical loop filter;
Lateral reference position output means for sequentially outputting the abscissa of the digital image as the lateral reference position;
A digital image position adjusting device comprising: A two-dimensional digital image in which each pixel is quantized into multiple values is interpolated with each pixel value using a vertical tap coefficient, and the vertical misalignment digital image is corrected by correcting the vertical displacement of the digital image. Vertical interpolation means to output;
A series of vertically interpolated digital images that are sequentially output from the vertical interpolating means in accordance with the change of the horizontal reference position is input, equalized to frequency characteristics adapted to the positional deviation correction in the vertical direction, and vertically equalized_vertically interpolated digital image Vertical equalization means for outputting
Each series of the vertical interpolation digital image and the vertical equalization_vertical interpolation digital image is input, and each pixel value is interpolated using a horizontal tap coefficient, and the vertical / horizontal interpolation digital image and vertical equalization_vertical / horizontal interpolation are performed. Horizontal interpolation means for outputting a digital image;
Horizontal equalization means for inputting a series of vertical / horizontal interpolation digital images sequentially output from the horizontal interpolation means, equalizing them to frequency characteristics adapted to lateral displacement correction, and outputting horizontal equalization_vertical / interpolation digital images; ,
Horizontal error detection means for detecting a horizontal phase error by comparing each pixel value of the horizontal equalization_vertical / horizontal interpolation digital image with a horizontal target value;
Vertical error detection means for comparing each pixel value of the vertical equalization_vertical / horizontal interpolation digital image with a vertical target value to detect a vertical phase error;
A horizontal loop filter that smoothes the lateral phase error and outputs a lateral frequency error;
A longitudinal loop filter that smoothes the longitudinal phase error and outputs a longitudinal frequency error;
A lateral frequency phase converting means for outputting the lateral tap coefficient from a lateral frequency error output from the lateral loop filter;
Vertical frequency phase conversion means for outputting the vertical tap coefficient from a vertical frequency error output from the vertical loop filter;
Lateral reference position output means for sequentially outputting the abscissa of the digital image as the lateral reference position;
A digital image position adjusting device comprising: A two-dimensional digital image in which each pixel is quantized into multiple values is interpolated with each pixel value using a vertical tap coefficient, and the vertical misalignment digital image is corrected by correcting the vertical displacement of the digital image. Vertical interpolation means to output;
Horizontal interpolation means for correcting the positional deviation in the horizontal direction of the first corrected digital image using a horizontal tap coefficient with respect to the vertical interpolation digital image and outputting a vertical / horizontal interpolation digital image;
Vertical equalization means for equalizing the frequency characteristics in accordance with the vertical displacement correction to the vertical / horizontal interpolation digital image and outputting the vertical / horizontal interpolation digital image;
Horizontal equalization means for equalizing the frequency characteristics in accordance with the positional deviation correction in the horizontal direction with respect to the vertical / horizontal interpolation digital image, and outputting horizontal equalization_vertical / horizontal interpolation digital image;
Horizontal error detection means for detecting a horizontal phase error by comparing each pixel value of the horizontal equalization_vertical / horizontal interpolation digital image with a horizontal target value;
Vertical error detection means for comparing each pixel value of the vertical equalization_vertical / horizontal interpolation digital image with a vertical target value to detect a vertical phase error;
A horizontal loop filter that smoothes the lateral phase error and outputs a lateral frequency error;
A longitudinal loop filter that smoothes the longitudinal phase error and outputs a longitudinal frequency error;
A lateral frequency phase converting means for outputting the lateral tap coefficient from a lateral frequency error output from the lateral loop filter;
Vertical frequency phase conversion means for outputting the vertical tap coefficient from a vertical frequency error output from the vertical loop filter;
Lateral reference position output means for sequentially outputting the abscissa of the digital image as the lateral reference position;
A digital image position adjusting device comprising: 3. The digital image position adjusting apparatus according to claim 2, wherein the horizontal equalizing means and the vertical equalizing means are means for equalizing to a partial response class 4. 7. The digital image position adjusting apparatus according to claim 2, wherein the vertical equalizing unit and the horizontal equalizing unit are configured by the same two-dimensional filter. 7. The digital image position adjusting apparatus according to claim 1, wherein the vertical interpolation unit and the horizontal interpolation unit are configured by the same two-dimensional filter. 2. The digital image position adjusting apparatus according to claim 1, wherein the vertical frequency phase conversion means outputs a plurality of rows of vertical interpolation positions simultaneously, and the vertical interpolation means and the horizontal interpolation means can process a plurality of rows simultaneously. The vertical frequency phase conversion means outputs a plurality of rows of vertical interpolation positions simultaneously, and the vertical interpolation means, the horizontal interpolation means, the vertical equalization means, and the horizontal equalization means can process a plurality of lines simultaneously. The digital image position adjusting device according to claim 2. The vertical frequency phase conversion means simultaneously outputs a plurality of rows of vertical interpolation positions, and at least one of the vertical error detection means and the horizontal error detection means outputs an average of errors of a plurality of rows. The digital image position adjusting device according to any one of 10 and 11. The vertical frequency phase conversion means simultaneously outputs M rows of vertical interpolation positions represented by a natural number M, and at least one of the vertical error detection means and the horizontal error detection means outputs an error of N rows represented by a natural number N. 12. The digital image position adjusting apparatus according to claim 10, wherein the digital image position adjusting apparatus outputs an average value and M> N. A spatial modulator that spatially modulates coherent light;
An interference optical system for recording the interference of the light output from the spatial modulator on the medium and reading the reproduction light from the medium, or for reading the interference of the light previously recorded on the medium as the reproduction light;
An image sensor for photoelectrically converting reproduction light read from the interference optical system;
Signal processing means for inputting a digital image read by the image sensor and performing signal processing;
The digital image position adjustment apparatus according to any one of claims 1 to 6, wherein the signal processing unit performs at least the alignment of the digital image.
Binarization means for binarizing the digital image that has been aligned by the digital image position adjustment device;
A data recording / reproducing apparatus. 15. The modulation pattern of the spatial modulator is at least a part, and a pull-in pattern for phase synchronization in the horizontal direction and a pull-in pattern for phase synchronization in the vertical direction are added. Data recording / reproducing apparatus.   The pull-in pattern to be added to the modulation pattern of the spatial modulator is obtained by adding vertical pull-in pattern data for vertical phase synchronization over a plurality of columns, and horizontal pull-in pattern for horizontal phase synchronization. The data recording / reproducing apparatus according to claim 15, wherein a plurality of lines are added. A two-dimensional digital image in which each pixel is quantized into multiple values is interpolated with each pixel value using a vertical tap coefficient, and the vertical misalignment digital image is corrected by correcting the vertical displacement of the digital image. Output step;
For the vertically interpolated digital image, using a horizontal tap coefficient, correcting the lateral displacement of the first corrected digital image and outputting a vertically and horizontally interpolated digital image;
Comparing each pixel value of the vertical and horizontal interpolation digital image with a horizontal target value to detect a horizontal phase error;
Comparing each pixel value of the vertical and horizontal interpolation digital image with a vertical target value to detect a vertical phase error;
Smoothing the lateral phase error and outputting a lateral frequency error;
Smoothing the longitudinal phase error and outputting a longitudinal frequency error;
Outputting the lateral tap coefficient from the lateral frequency error;
Outputting the vertical tap coefficient from the frequency error in the vertical direction;
A digital image position adjustment method comprising: Performing equalization to the frequency characteristics adapted to the lateral displacement correction prior to outputting the lateral frequency error;
Performing equalization to frequency characteristics in accordance with lateral displacement correction,
18. The digital image position adjusting method according to claim 17, wherein:

Images