JPH0773305A - Recording and reproducing device - Google Patents

Abstract

PURPOSE:To provide a recording and reproducing device which can efficiently extract only recording infiormation from image informations affected by unevenness on the surface of a recording medium. CONSTITUTION:In the recording and reproducing device for utilizing a signal to be generated by a physical mutual operation between a probe electrode 116 and a recording medium 111, by using a predictive means to calculate the predictive picture element values of picture elements in two-dimensional image data to produce predictive image data composed of prodictive picture elements provided with the predictive picture element values while using predictive parameters decided in advance, a learning means to optimize the predictive parameters while using the respective picture element values of image data and the respective predictive picture element values of the predictive image data, the two-dimensional image data are binarized by calculating the respective predictive picture element values of image data at the predictive means again while using thr predictive parameters optimized by the learning means.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a recording / reproducing apparatus and an image processing apparatus used in the recording / reproducing apparatus, and more particularly, to a recording medium while two-dimensionally scanning probe electrodes relative to the recording medium. The present invention relates to a recording / reproducing apparatus that records information and reproduces the recorded information as image information, and an image processing apparatus that extracts the recorded information from the image information reproduced by the recording / reproducing apparatus.

[0002]

2. Description of the Related Art In recent years, the recording capacity of data in a recording device tends to increase. In such a tendency, it is essential that the size of the recording unit becomes smaller and the density becomes higher. For example, in a digital audio disc by optical recording, the size of a recording unit reaches about 1 μm ² .

On the other hand, recently, a scanning tunneling microscope (Sc
anning Tunneling Microscope; hereinafter abbreviated as STM)
Has been developed and has the following advantages, so that it is expected to have a wide range of applications.

1. High resolution measurement of real space images can be performed regardless of single crystal or amorphous.

2. It can be observed at low power without damaging the recording medium due to electric current.

3. It works not only in ultra-high vacuum but also in air and solution.

4. It can be used for various materials.

The STM utilizes that a tunnel current flows between the probe electrode (metal probe) and a conductive substance when a voltage is applied to bring them close to a distance of about 1 nm. The tunnel current is very sensitive to changes in the distance between the probe electrode and the conductive material. Therefore, surface information in the real space can be obtained by scanning the probe in the in-plane direction while performing control such that the tunnel current or the average distance between the two is kept constant. At this time, the resolution in the in-plane direction is 10 nm or less.

By applying this STM principle and using a material having a memory effect for the voltage-current switching characteristics as a recording medium, for example, a thin film layer of a π-electron organic compound or chalcogenide, a recording unit becomes Information recording of 0.01 μm ² or less is possible.

Further, if a method of changing the surface state of the medium using an electromagnetic wave such as an electron beam or light is used, the recording unit becomes large due to the limitation of the focusing degree of the beam, but it is equivalent to the current optical recording. It is possible to record and reproduce information at the recording density of.

However, in the conventional recording / reproducing apparatus applying the above-mentioned STM principle, when recording / reproducing high-density information over the entire surface (recording surface) of the recording medium having a predetermined area, The ruggedness on the recording surface deteriorates the SN ratio and error rate of the signal at the time of reproduction, which is a major obstacle to achieving a high recording capacity. As one method for reducing the physical unevenness of the surface of the recording medium, a method of growing a single crystal of an Au thin film on the cleavage plane of mica to obtain a smooth surface can be applied.

However, in the recording medium by such a method,
Normally, there is almost no unevenness of the same size as the recording bit size with a diameter of 10 nm or less, but there are also atomic-order steps and ridges. The problem that the shape of the recording bit is changed by the unevenness of the atomic order and the SN ratio and error rate of the signal at the time of reproduction is lowered remains.

In order to solve this problem, after reading the information recorded on the recording medium as image information, it is necessary to perform image processing for extracting only the recording information from the read image information. For example, by extracting the recording bits from the reproduced image by measuring the state of the surface of the recording medium in which the recording bits have not been written in advance and determining the recorded information from the difference between the surface state and the surface state shown by the reproduced image. It is possible to do it.

[0014]

However, in the recording / reproducing apparatus to which the conventional STM principle is applied, there is a problem that the SN ratio of the reproduced signal is lowered and the error rate is increased due to the unevenness on the surface of the recording medium. In order to solve the problem, as described above, when the state of the surface of the recording medium in which the recording bits are not written is measured in advance to create an image before recording, and the reproduced image is compared with the image before recording. Requires that the surface condition of the recording medium before recording is stored in the recording medium. Is required, the recording area used for recording is greatly limited, and there is no point in recording information on the medium.

The present invention has been made in view of the problems of the above-mentioned prior art, and it is possible to efficiently extract only the recording information from the image information affected by the unevenness of the surface of the recording medium. An object is to provide a reproducing device.

[0016]

The recording / reproducing apparatus of the present invention physically scans the probe electrode between the probe electrode and the recording medium when the probe electrode is two-dimensionally scanned relative to the recording surface of the recording medium. In a recording / reproducing apparatus for recording information on the recording medium using a signal generated by the interaction and reproducing the recorded information as image information, a pixel of two-dimensional image data using a predetermined prediction parameter. Of the prediction pixel value of a pixel in the vicinity of the pixel and predicting image data including prediction pixels having the prediction pixel value, and pixel value of the image data and prediction of the prediction image data A learning means for optimizing a prediction parameter by using the pixel value and a prediction parameter optimized by the learning means are used to re-determine each prediction pixel value of image data by the prediction means. Ri, characterized by binarizing the two-dimensional image data.

A recording / reproducing apparatus according to another aspect of the present invention is
When the probe electrode is two-dimensionally scanned relative to the recording surface of the recording medium, information is recorded on the recording medium by using a signal generated by a physical interaction between the probe electrode and the recording medium. In a recording / reproducing apparatus that reproduces the information as image information, a predicted pixel value of a pixel of two-dimensional image data is obtained from pixel values of pixels in the vicinity of the pixel using a predetermined prediction parameter, and the predicted pixel A predicting means for creating predictive image data composed of predictive pixels having a value; a learning means for optimizing a prediction parameter using each pixel value of the image data and each predictive pixel value of the predictive image data; It is characterized in that two-dimensional image data is binarized by re-obtaining each prediction pixel value of the image data by the prediction means using the optimized prediction parameter.

In any of the above recording / reproducing devices,
The generation of the predicted image data by the prediction unit and the optimization of the prediction parameter by the learning unit may be performed before recording, and the unrecorded area is used except when recording or reproducing recorded information. May be performed.

As a concrete configuration example, the absolute value of the difference between each predicted pixel value obtained by the prediction means and each pixel value of the image data is calculated, and if the absolute value is smaller than a predetermined threshold value, Create a binary image having "0" as a pixel value and "1" if the absolute value is larger than a predetermined threshold value 2
If the pixel value of the binary image created by the binary image created by the binary image created by the binary image created by the binary image created by the binary image created by the binary image created by the binary image created by the binary image creating means is "0". For example, if the pixel value of the image data is the pixel value of the binary image is '1', the predicted pixel value of the predicted image data is selected to create a composite image, and then the created composite image is output to the prediction means. A binary image reconstructing unit including a composite image creating unit is provided, and the predicting unit uses the composite image instead of two-dimensional image data when the composite image is input from the composite image creating unit, New prediction image data is created using the prediction parameters optimized by the learning means, and the binary image creation means creates a new binary image from the new prediction image data created by the prediction means, The binary image reconstructing means is equivalent to the binary image storing means. The binary image is reconstructed by obtaining the logical sum of the pixel value of the stored binary image and the pixel value of the new binary image created by the binary image creating means, and creating the binary image. An example is one that outputs the reconstructed binary image to the outside when the values of all the pixels of the binary image sent from the means are "0".

[0020]

In the image processing apparatus of the present invention, the image data before recording can be predicted from the image data after recording by using the optimized prediction parameter obtained by the learning means. The prediction parameter serves as a calculation parameter when the value of the pixel to be predicted is obtained from the values of pixels around the pixel to be predicted, and is effective for all areas of the recording medium.

In the present invention, since the pixel value is predicted by using the above-mentioned prediction parameter, the bits recorded on the recording medium can be efficiently extracted.

The binary image creating means subtracts the predicted image data before recording from the recorded image data to obtain difference data indicating the difference between the predicted image data before recording and the image data after recording. In addition, it is possible to obtain the recording information (binary image) to be obtained by binarizing the difference data.

However, since the predicted image data before recording does not always match the image data before recording, the binary image created by the binary image creating means is stored in the binary image storage means. After that, in the composite image creating means, if the pixel value of the binary image created by the binary image creating means is '0', the pixel value of the image data is set to the pixel value of the binary image. For example, a predicted pixel value of predicted image data is selected to create a composite image, thereby creating a composite image that is close to the image data before recording. Subsequently, the predicting means similarly predicts the image data before recording from the composite image data, and the binary image creating means subtracts the predicted image data before recording from the composite image data to thereby post-record. The difference data indicating the difference from the image data before recording predicted from the image data of 1 can be obtained, and the obtained difference data is binarized to obtain a binary image. In the binary image thus obtained, when the predicted pre-recording image data matches the pre-recording image data, the values of all the pixels are “0”, so that all the binary image The above operation is repeated until the pixel value becomes "0".

Further, in each pixel of the binary image obtained by repeating the above operation, one having a pixel value of "1" corresponds to a pixel in which information is recorded and has a pixel value of "0". The ones correspond to the pixels for which no information is recorded. Therefore, each time the operation is performed,
The binary image reconstructing means obtains a logical sum of the pixel value of the binary image stored in the binary image storing means and the pixel value of the new binary image created by the binary image creating means. By reconstructing the binary image by using the above method, a binary image corresponding to the record information can be finally obtained.

In the recording / reproducing apparatus of the present invention, the image processing means is constructed similarly to the above-mentioned image processing apparatus, and the learning means optimizes the prediction parameter by using the surface state image data before recording. It is not necessary to store the surface state image data before recording, and the surface area in which the state has changed due to the recording operation is extracted and binarized.

However, also here, the predicted surface condition image data before recording does not always match the surface condition image data before recording. When predicting a pixel value of a pixel inside a recording bit area in which a recording bit, which is a recording unit of information, is written from pixel values of pixels in the vicinity thereof, when the neighboring pixel is also inside the recording bit area, the predicting means Since the predicted pixel value is predicted from the pixel values of the neighboring pixels, the predicted pixel value at this time does not become the pixel value of the pixel before recording. However, the surface state image data before recording can be predicted from the pixel value of the pixel at the boundary portion of the recording bit area. From the surface condition image data including the recording bit area and the predicted image thereof, the binary image creating means determines from the pixel which gives “1” to the pixel value of the boundary portion where the surface condition image data before recording can be predicted. Then, the composite image creating means creates composite image data in which the boundary portion is replaced with the surface state image data before recording. By repeating the above-mentioned process by using this composite image data as an input image of the prediction means, the binary image creation means is made up of binary values each having a pixel value "1" at the newly formed boundary portion of the recording bit area. The images are sequentially created, and the composite image creating means sequentially creates the composite image data in which the newly formed boundary portion is replaced with the surface state image data before recording. When all the pixel values of the binary image become '0', the pixel values of the pixels in the recording bit area are all replaced by the pixel values of the surface state image data before recording, which is the binary value. The image reconstructing means obtains the logical sum of the pixel values of the binary image obtained up to this point to obtain a binary image in which the recording bit area corresponds to "1".

In the recording / reproducing apparatus provided with the scanning control means and the learning control means, even if a part of the surface of the recording medium is already used for recording, there is a free time during which the apparatus does not perform recording / reproducing operation. The learning control means compares the difference data with a predetermined threshold value from a plurality of areas on the surface of the recording medium arbitrarily selected by the scanning control means, and the unused value is not used. Since the area can be found, learning of the prediction parameter is performed at any time using the pixel value of the pixel of the surface state image data corresponding to this area and the predicted pixel value thereof.

[0028]

Embodiments of the present invention will now be described with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a recording / reproducing apparatus according to an embodiment of the present invention.

The main body of the recording / reproducing apparatus 101 of the present invention includes a structure 113 for supporting a fine motion control mechanism 117 for finely moving the probe electrode 116 and a coarse motion control mechanism 112 for coarsely moving the recording medium 111, and a structure 111. It is composed of a large coarse movement control mechanism 114 on which is mounted and a vibration isolation table 115 on which the large coarse movement control mechanism 114 is mounted.

The structure 113 supports each structural part and is made of an Invar alloy. The coarse movement control mechanism 112 is attached to the bottom surface of the structure 113, and the recording medium 111 is placed on the coarse movement control mechanism 112. The recording medium 111 is a medium on which information is actually written, and squarylium-bis-6-octylazulene, which has a memory effect on current-voltage switching characteristics, is formed on a graphite substrate by the Langmuir-Blodgett method (LB
Method) is used to accumulate 8 layers. The recording medium 111 is in a non-recording state (off state) in the initial state. The coarse motion control mechanism 112 has a parallel spring that uses an elastic hinge, and is shown on the recording medium 111 in the drawing X.
This is for performing coarse movement control in the axial direction and the Y-axis direction. A fine movement control mechanism 117 is attached to the inside of the upper surface of the structure 113 so that the probe electrode 116 faces the recording medium 111. The probe electrode 116 is
It is used for recording and reproducing information. The probe electrode 116 uses a tungsten needle whose tip is mechanically polished or electrolytically polished in order to improve the recording / reproducing resolution. The probe electrode 116 is made of Pt-Ir or Pt.
Etc., and the processing method is not limited to this. The fine movement control mechanism 117 is composed of a cylindrical piezoelectric element,
The probe electrode 116 is moved in the in-plane direction (X,
It is for moving in the Y direction) and in the direction between the recording medium 111 and the probe electrode 116 (Z direction). During normal operation, the distance between the tip of the probe electrode 116 and the surface of the recording medium 111 is such that a tunnel current flows.

The large coarse movement mechanism 114 has a fine movement control mechanism 117.
Also, the size of the coarse motion control mechanism 112 is out of the control range. The anti-vibration table 115 is for preventing external vibration from being transmitted to the recording / reproducing apparatus 101 and preventing malfunction of the recording / reproducing apparatus 101 due to external vibration.

The electrical and control system of the recording / reproducing apparatus 101 includes a voltage applying device 119 and a probe current amplifier 120.
And an arithmetic unit 121 and an XY controller 118.

The voltage applying device 119 stores information on the recording medium 1
11, a bias voltage for recording, reproducing and erasing is set to −10 between the probe electrode 116 and the recording medium 111.
It is given with an arbitrary size in the range of V to + 10V.

The probe current amplifier 120 is connected to the probe electrode 116, detects a tunnel current flowing through the probe electrode 116 by a bias voltage,
The detected tunnel current is converted into a voltage signal, amplified, and output to the arithmetic unit 121.

The XY controller 118 controls the fine movement control mechanism 117 and the coarse movement control mechanism 112 to displace the probe electrode 116 arbitrarily relative to the recording medium 111.

The arithmetic unit 121 is the probe current amplifier 1
20 for inputting the output of the voltage applying device 119
And control the XY controller 118. Arithmetic unit 12
1 is an XY controller 118 and a voltage application device 11
9 for controlling each of them, an A / D converter for digitally converting and quantizing the voltage signal sent from the probe current amplifier 120, a tunnel current converted into a digital signal, or movement of the probe electrode 116 in the Z direction. The image processing apparatus stores a reproduced image having a pixel value of the voltage applied to the fine movement control mechanism 117 via the XY controller 118 and extracts only the recording bit area from the reproduced image.

In this recording / reproducing apparatus, in addition to the normal STM operation, the probe electrode 116 is moved to an arbitrary position in the space on the recording medium 111, and -10 V is applied even when feedback is applied or feedback is turned off.
A bias voltage of any magnitude can be applied within the range of + 10V.

Although the actual recording / reproducing operation of the recording / reproducing apparatus 101 will not be described in detail here, the recording on the recording medium 111 is larger than the threshold value of the voltage which causes the memory effect of the recording medium 111. It is performed by applying a voltage of +10 V to the recording medium 111 as a recording voltage.
In the area where the memory effect is generated, when the reproduction is performed with a bias voltage lower than the threshold value after the application of the recording voltage, for example, in the constant current mode in STM, an image having an upward convex appears, so that the recording information ( It can correspond to a recording bit ('1') of '1' or '0').

By the way, the recording / reproducing apparatus 101 of the present embodiment.
Then, as described above, the image processing device is incorporated in the arithmetic device 121. The recording bit, which is the minimum unit of recording information, appears as a convex region on the image as described above, but in the actual image, the influence of the irregularities on the surface of the recording medium is further superimposed. This image processing apparatus uses a general threshold logic to avoid the influence of the unevenness of the surface of the recording medium from the reproduced image in which the pixel value is the value corresponding to the tunnel current value or the displacement amount of the probe electrode 116 in the Z direction. Recording bit area image (binary image) in which recording information is extracted by a method different from binarization and the recording information is used as a pixel value
Is output.

FIG. 2 is a block diagram showing the structure of the image processing apparatus.

Predictor 212 used in the image processing apparatus
Has a learning function of predicting the value of each pixel of a two-dimensional image from the value of each pixel in the vicinity of the two-dimensional image, and adjusting the prediction parameter by comparing the input image that is input with the predicted image that is predicted. ing. First, the predictor 212 will be described. FIG. 3 is a principle diagram showing the learning process of the predictor 212, and FIG. 4 is an example of the learning process.

Since the learning process is to learn the unevenness of the surface of the recording medium 111, it is necessary that the recording bit is not written, and the surface within the range covered by the scanning of the fine movement control mechanism 117 is used at the time of shipment. This is done as follows.

In the learning process, a reproduced image 302 obtained from the recording / reproducing apparatus 101 (however, as described above, an image of the surface in the non-recorded state at the time of shipment, for example, the image shown in FIG. 4A) predictor 212 And input. For example, the predictor 21
If 2 is composed of an adaptive AR model, the adjusted prediction parameter is a linear coefficient, and the adjustment method is
Mean square error 30 between the reproduced image 302 and the predicted image 304
The steepest descent method based on 5 is achieved. If the predictor is composed of a neural network, the adjusted prediction parameter is the coupling coefficient (weight) between units.
The adjustment method is also achieved by the steepest descent method based on the mean square error 305 between the reproduced image 302 and the predicted image 304. Through such a learning process, the predictor 212 can output the predicted image shown in FIG. 4B with respect to the input image shown in FIG. 4A as shown in FIG. It can be expected that the absolute value 403 of the difference between the pixels will fall within a certain range.

The image processing apparatus includes an input image storage unit 211,
A prediction unit 210 including a predictor 212 and a prediction image storage unit 213, a pixel processing unit 220 including an absolute value circuit 221, a comparison circuit 222, output limiting circuits 223 and 224, and a combining unit 225, and a binary image storage unit 231. The image reconstructing unit 230 includes a binary image synthesizing unit 232 and a pixel value summing unit 233.

The input image storage section 211 is used for the recording / reproducing apparatus 1.
01 reproduced image or the pixel processing unit 2 as described later.
The composite image input from 20 is temporarily stored as an input image to be output to the predictor 212, and the value of each pixel forming the input image is sequentially stored in the absolute value circuit 2 of the pixel processing unit 220.
21 and the output limiting circuit 223. Predictor 21
2 outputs the predicted image predicted from the input image as described above to the predicted image storage unit 213. Predicted image storage unit 213
Temporarily stores the predicted image predicted by the predictor 212, and sequentially outputs the value of each pixel forming the predicted image to the absolute value circuit 221 and the output limiting circuit 224 of the pixel processing unit.

The absolute value circuit 221 includes the input image storage unit 21.
The difference between the value of each pixel of the input image from 1 and the value of each pixel of the predicted image from the predicted image storage unit 213, that is, the prediction error is obtained, and the absolute value of this prediction error is output to the comparison circuit 222. The comparison circuit 222 compares the absolute value of the difference data output from the absolute value circuit 221 with a preset threshold value, and if the absolute value of the difference data is larger than the threshold value, it is '1'.
Is output to the output limiting circuits 223 and 224, the binary image storage unit 231 of the image reconstructing unit 230, and the pixel value summing unit 233. Output limiting circuits 223, 22
4 has three terminals, an input terminal, an output terminal and an output restriction terminal. If the input of the output restriction terminal is' 1 ', the input is directly output from the output terminal and the input of the output restriction terminal is' If it is 0 ', the circuit outputs nothing from the output terminal regardless of the input value. Output limiting circuit 223
Inputs the value of each pixel of the input image from the input image storage unit 211 to the input terminal, inverts the binary output of the comparison circuit 222 to the output limiting terminal, and outputs the value of each pixel whose output is limited to the combining unit. To 225. The output limiting circuit 224 inputs the value of each pixel of the predicted image from the predicted image storage unit 213 to the input terminal, inputs the binary output of the comparison circuit 222 to the output limiting terminal as it is, and outputs each pixel of the output-limited pixel. The value is synthesized part 2
Output to 25. The combining unit 225 uses the output limiting circuit 223.
And each pixel value of the output of the output limiting circuit 224 is added and output to the input image storage unit 211 of the prediction unit 210.
Therefore, this output forms a composite image composed of pixels of the input image or the predicted image in the input image storage unit 211.

The binary image storage unit 231 has a comparison circuit 222.
The binary output of 1 is stored until one image is formed, and the binary is output to the binary image forming unit 232 at an appropriate reading speed. Each pixel value of the binary image stored in the binary image storage unit 231 is updated to a pixel value newly input for each image. The pixel value summing unit 233 includes “1” included in one image when the binary output of the comparison circuit 222 forms one image.
If the number of pixels in '1' becomes 0,
An output start signal is output to the binary image construction unit 232. The binary image synthesis unit 232 stores the binary output of the binary image storage unit 231. Each time a new pixel value is input, the value of each pixel to be stored is updated to the logical sum of the input pixel value and the value already stored. Upon receiving an output start signal from the pixel value summing unit 233, the stored binary image is output as a recording bit area image which is a final output image of the image processing apparatus.

Next, the operation of the image processing apparatus will be described.

The predictor 212 outputs a predictive image for the input image based on the predictive parameters adjusted as described above. As the input image of the predictor 212, the reproduced image from the recording / reproducing apparatus 101 is given as an initial value.

The absolute value circuit 221 calculates the absolute value of the difference (difference data) between the input image and the predicted image for each pixel, and the comparison circuit 222 checks the magnitude relationship between the absolute value of this difference data and the threshold value. . Therefore, the output of the comparison circuit 222 is a binary image in which the pixels whose absolute value of the difference data is larger than the threshold value are “1” in the binary image storage unit 231.

The output of the comparison circuit 222 is output to the output restriction terminals of the output restriction circuits 223 and 224. However, since the output restriction terminal of the output restriction circuit 223 is an inverting input, the output of the comparison circuit 222 is If the output is "1", the pixel of the predicted image is output to the combining unit 225, and if the output of the comparison circuit 222 is "0", the pixel of the input image is
It is output to the combining unit 225. Therefore, the composite image in the input image storage unit 211 is an input image in which the pixels of which the absolute value of the difference data is larger than the threshold value are replaced with the pixels of the prediction image. The prediction unit 210 and the pixel processing unit 220 create a prediction image and a composite image by using this composite image as a new input image.

A binary image is sent to the image reconstruction unit 230 every time the predictor 212 outputs a predicted image.
The value image synthesis unit 232 calculates the set sum of the binary images, and the pixel value summing unit 233 calculates the sum of the pixel values. When all the pixel values of the pixels of the binary image are “0”, the binary image synthesizing unit 232 receives the output start signal from the pixel value summing unit 233 and records the binary image obtained by taking the set sum. Finally output as a bit domain image.

FIG. 5 shows a specific example of this operation. For simplicity, the reproduced image shown in FIG. 5A is one-dimensional data. The predictor 212 outputs the predicted image as shown in FIG. 5B for the reproduced image shown in FIG. Absolute value of these difference data The image shown in FIG. 5C is generated, and based on this, the comparison circuit 122 creates the binary image shown in FIG. 5D by the threshold value, and the output limiting circuit 223. , 224 and the combining unit 225 are shown in FIG.
The combined image shown in (e) is combined. In the composite image shown in FIG. 5 (e), pixels having a pixel value of 1 in the binary image shown in FIG. 5 (d) corresponding to pixels in the reproduced image shown in FIG. 5 (a) are shown in FIG. 5 (b). ) Are replaced with the pixels of the prediction image shown in FIG. The composite image shown in FIG. 5E synthesized in this way is used as an input image, and the above process is repeated.
The binary image and composite image shown in (h) and (i) are obtained. Figure 5
The above process is performed again using the binary image shown in (h) as an input image to obtain the binary image shown in FIG. Since the pixel values of this binary image are all 0, the composite image and binary image obtained by performing the above process again are as shown in FIG.
The composite image shown in FIG. 5 is the same as the binary image shown in FIG. The binary image synthesizing unit 232 calculates the logical sum of each pixel of the binary images shown in FIGS. 5 (d) and 5 (h), which are obtained in this way, to obtain the binary image shown in FIG. 5 (m). Create a value image. The image processing device is finally obtained in FIG.
The binary image shown in (m) is output as a recorded bit area image.

Although the learning process is performed at the time of shipping in the above embodiment, the learning process is performed at any time after shipping in order to accurately predict even if the surface condition of the recording medium 111 changes with time. It is conceivable to update the prediction parameter by updating. The update of the prediction parameter is performed when the recording / reproducing apparatus 101 does not perform the recording / reproducing operation, for example, before the recording / reproducing apparatus 101 is powered on to perform the recording / reproducing operation, or the recording / reproducing apparatus 101 is powered off. After the previous recording / reproducing operation, it is necessary to use the area of the recording medium 111 where the recording bits are not written.

To this end, a scanning control unit is newly provided in the arithmetic unit 121 and a learning control unit is newly provided in the prediction unit 210 of the image processing apparatus. The scanning control unit is for automatically scanning the probe electrode 116 over an arbitrary area of the recording medium 111 when the recording / reproducing apparatus is in a predetermined state, for example, after power-on to obtain a reproduced image. The learning control unit operates in conjunction with the scanning control unit, and controls the learning function of the predictor 212 as follows based on the reproduced image obtained by the scanning control unit.

Reproduced images of the area on the surface of the recording medium 111 randomly selected by the scanning control unit are sequentially sent to the image processing apparatus. The image processing apparatus performs the above processing, and at this time, the learning control unit outputs the absolute value circuit 221.
That is, the difference between the pixel values of the reproduced image and the predicted image is monitored. The learning control unit confirms that this difference is smaller than a preset value, that is, confirms that there are no recording bits in this area, and then uses the pixel value of the pixel having this difference and the pixel values of its neighboring pixels. Predictor 21 for learning process
Let 2 do it.

In the above-described embodiments, the recording / reproducing apparatus is described as the STM for detecting the tunnel current generated between the probe and the recording medium, but the configuration characteristic of the present invention is not limited to this. Instead of an object, an atomic force microscope (AFM) that detects the atomic force between the probe and the recording medium may be used.

[0058]

Since the present invention is configured as described above, it has the following effects.

In the recording / reproducing apparatus, since the recording bit area is extracted by taking the difference between the surface state image data after recording and the surface state image data before recording predicted from this image data, the surface of the recording medium is extracted. Recording can be reproduced accurately even if there are irregularities on the surface. Further, since the prediction of the surface state image data before recording may have a prediction parameter,
It is not necessary to use an extra recording area on the surface of the recording medium,
There is an advantage that there is no need to have any other storage means and it does not burden the apparatus itself.

In the recording / reproducing apparatus provided with the scanning control means and the learning control means, the learning of the prediction parameter can be performed at any time by using the unused area on the surface of the recording medium, so that the surface of the recording medium changes with time. However, the recording can be reproduced accurately.

Since the image processing apparatus can handle general image data, it can be used particularly when detecting a change with respect to the background, for example, when counting cells.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a recording / reproducing apparatus according to an embodiment of the present invention.

FIG. 2 is a block diagram showing the configuration of an image processing apparatus according to an embodiment of the present invention.

FIG. 3 is a principle diagram showing a learning process of a predictor.

FIG. 4 is a diagram showing an example of a learning process.

FIG. 5 is a diagram showing a specific example of the operation of the image processing apparatus.

[Explanation of reference numerals] 101 recording / reproducing device 111 recording medium 112 coarse movement control mechanism 113 structure 114 large coarse movement control mechanism 115 seismic isolation table 116 probe electrode 117 fine movement control mechanism 118 XY controller 119 voltage application device 120 probe current amplifier 121 calculation Device 210 Prediction unit 211 Input image storage unit 212 Predictor 213 Prediction image storage unit 220 Pixel processing unit 221 Absolute value circuit 222 Comparison circuit 223, 224 Output restriction circuit 225 Compositing unit 230 Image reconstruction unit 231 Binary image storage unit 232 2 Value image synthesizing unit 233 Pixel value summing unit 302 Reproduced image 304 Predicted image 305 Mean square error 401 Input image 402 Predicted image 403 Absolute value of pixel difference 501 Reproduced image 502, 506, 510 Predicted image 503, 507, 511 Difference data The absolute Value 504,508,511,512 binary image, 505, 509 synthetic image

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical display location G11B 20/18 560 J 9074-5D H01J 37/28 Z H04N 5/92

Claims (5)

[Claims] 1. A recording medium utilizing a signal generated by a physical interaction between the probe electrode and the recording medium when the probe electrode is two-dimensionally scanned relative to the recording surface of the recording medium. In a recording / reproducing apparatus for recording information in a recording medium and reproducing the recorded information as image information, a prediction pixel value of a pixel of two-dimensional image data is calculated by using a predetermined prediction parameter. Prediction means that obtains predicted image data composed of predicted pixels having the predicted pixel value and pixel values of the image data and predicted pixel values of the predicted image data, and uses the prediction parameter Learning means for optimizing, and using the prediction parameters optimized by the learning means,
A recording / reproducing apparatus, wherein the two-dimensional image data is binarized by recalculating each predicted pixel value of the image data by the prediction means. 2. Information is recorded on the recording medium by using a signal generated by a physical interaction between the probe electrode and the recording medium when the probe electrode is two-dimensionally scanned relative to the recording surface of the recording medium. In a recording / reproducing apparatus for recording the recorded information and reproducing the recorded information as image information, the predicted pixel value of the pixel of the two-dimensional image data is set to the pixel value of the pixel in the vicinity of the pixel using a predetermined prediction parameter. The prediction parameter is optimized by using a prediction unit that generates prediction image data including prediction pixels having the prediction pixel value and each pixel value of the image data and each prediction pixel value of the prediction image data. By using the learning means to be converted and the prediction parameters optimized by the learning means,
A recording / reproducing apparatus, wherein the two-dimensional image data is binarized by recalculating each predicted pixel value of the image data by the prediction means. 3. The recording / reproducing apparatus according to claim 1 or 2, wherein creation of the predicted image data by the prediction means and optimization of the prediction parameter by the learning means are performed before recording. Recording / playback device. 4. The recording / reproducing apparatus according to claim 1, wherein the prediction image data is created by the prediction means and the prediction parameters are optimized by the learning means when the recording information is not recorded or reproduced. The recording / reproducing apparatus is characterized by being performed using an unrecorded area. 5. The recording / reproducing apparatus according to claim 1, wherein an absolute value of a difference between each predicted pixel value obtained by the prediction means and each pixel value of the image data is calculated. A binary image creating means for creating a binary image having a pixel value of "0" when the absolute value is smaller than a predetermined threshold value and "1" when the absolute value is larger than the predetermined threshold value. A binary image storage means for storing the binary image created by the binary image creating means; and a pixel value of the binary image created by the binary image creating means being '0'. If the pixel value of the image data is 2
If the pixel value of the value image is '1', the predicted pixel value of the predicted image data is selected to create a composite image, and then the created composite image is output to the prediction means. When the synthetic image is input from the synthetic image creating unit, the predicting unit uses the synthetic image instead of the two-dimensional image data and the learning unit. New prediction image data is created using the prediction parameters optimized by, and the binary image creation means creates a new binary image from the new prediction image data created by the prediction means. The binary image reconstructing means stores the value of the pixel of the binary image stored in the binary image storage means and the value of the pixel of the new binary image created by the binary image creating means. Binary image for logical sum of Is reconstructed, and the reconstructed binary image is output to the outside when all the pixel values of the binary image sent from the binary image creating means are "0". Recording / playback device.