Signal processing device, signal processing method and information reading apparatus

The present invention relates to an information reading apparatus having a sensor. More particularly, the invention relates to an information reading apparatus suitable for reading pattern information, such as an image printed with magnetic ink on paper currency or other such paper notes, by using a line sensor, in which a plurality of sensing elements are disposed in a line.

As one information reading apparatus of this type, there is a magnetic sensing apparatus having a line sensor with a plurality of magnetic sensors whose resistance values change when a magnetic field is applied, and which are disposed in a line. The magnetic sensing apparatus is used in, e.g., a paper currency identification apparatus that identifies paper currency. By means of the magnetic sensors, the paper currency identification apparatus detects an image printed with magnetic ink on the paper currency to be identified. The paper currency identification apparatus also makes determination on the validity, denomination, and so on of the paper currency on the basis of similarity between the detected printed image and a stored reference image.

Because the magnetic field due to the magnetic ink used to print the image on the paper currency or the like is quite weak, the sensing signal obtained by the magnetic sensing apparatus is quite weak. Moreover, a direct current component (offset component) is added to the sensing signal in the magnetic sensing apparatus. The sensing signal must accordingly be amplified and the offset component must be removed. The offset component varies, however, with environmental changes such as changes in the ambient temperature, and with aging changes and the like in the magnetic sensors. It is therefore not easy to separate the offset component in the sensing signal.

A technique for correcting non-uniformity in the level of the sensing signal due to the temperature of the magnetic sensing elements and so on is disclosed in the magnetic pattern sensing apparatus described in patent reference 1. The magnetic pattern sensing apparatus in patent reference 1 performs an offset adjustment during each interval between the passage of one piece of paper currency through the magnetic sensing apparatus and the passage of the next piece of paper currency through the magnetic sensing apparatus.

Patent References

Patent reference 1: Japanese Patent Application Publication No. 2012-128683 (paragraphs 0049-0062, FIGS. 4-6)

Problem to be Solved by the Invention

The offset adjustment method in patent reference 1, however, calculates the offset adjustment value only in the interval between one piece of paper currency and the next. Throughout the interval in which it is reading a piece of paper currency, the offset adjustment method in patent reference 1 applies the same offset adjustment value. A consequent problem is that when the sensing signal level varies during the interval in which a piece of paper currency is being read, non-uniformity in the level of the sensing signal cannot be adequately corrected.

There has been a similar problem in apparatus that reads information from image patterns other than magnetic patterns.

This problem is addressed by the present invention, which reduces the effect of variations in the level of the sensing signal that occur during the reading period of each object being sensed. By quickly correcting the non-uniformity in the level of the sensing signal, the invention can read pattern information accurately.

Means for Solving the Problem

A signal processing device according to the present invention comprises:

a pattern presence/absence evaluation value generation unit for calculating, on a basis of sensing data which includes a signal component of an object being sensed and which is obtained by sequential sensing, by a sensing element, of the object being scanned, a degree of inclusion of the signal component in the sensing data;

a correction data updating unit for taking a weighted sum of correction data representing a direct current component of the sensing element and the sensing data to generate updated correction data, with weight given to the correction data being greater as said degree increases, and weight given to the sensing data being greater as said degree decreases; and

a correction data subtraction unit for subtracting the updated correction data from the sensing data to generate corrected sensing data.

Effect of the Invention

The invention can reduce the effect of variations in the level of a sensing signal that occur during the reading period of each object that is sensed.

FIG. 1 is a block diagram illustrating an information reading apparatus in a first embodiment of the invention.

FIG. 2 illustrates the schematic structure and the operation of the pattern sensing unit 1 in FIG. 1 .

FIG. 3 illustrates the operation of the sensing enable generation unit 2 in FIG. 1 .

FIGS. 4( a ) to 4( e ) are timing diagrams illustrating signals output from the A/D conversion unit 3 in FIG. 1 .

FIG. 5 is a block diagram showing an example of the structure of the correction data updating unit 42 in FIG. 1 .

FIG. 6 illustrates the relation between the input and the output of the weighting coefficient generation unit 421 in FIG. 5 .

FIG. 7 is a flowchart illustrating the operation of the direct current component correction unit 4 in FIG. 1 .

FIG. 8 is a flowchart illustrating details of the weighting coefficient generation step S 31 in FIG. 7 .

FIGS. 9( a ) to 9( d ) show other examples of changes, in the secondary scanning direction, in the signals of the individual units of the direct current component correction unit 4 in FIG. 1 .

FIGS. 10( a ) to 10( d ) show other examples of changes, in the secondary scanning direction, in the signals of the individual units of the direct current component correction unit 4 in FIG. 1 .

FIG. 11 is a block diagram illustrating an information reading apparatus in a third embodiment of the invention.

FIG. 12 is a block diagram showing an example of the structure of the correction data updating unit 42 b in FIG. 11 .

FIG. 1 shows the structure of an information reading apparatus 100 according to a first embodiment of the invention. The information reading apparatus 100 has a pattern sensing unit 1 , a sensing enable generation unit 2 , an A/D conversion unit 3 , a direct current component correction unit 4 , a signal conditioning unit 5 , and a system control unit 6 . The direct current component correction unit 4 has a pattern presence/absence evaluation value generation unit 41 , a correction data updating unit 42 , a correction data subtraction unit 43 , and a correction data line memory 44 . The direct current component correction unit 4 may also have a delay unit 45 . The direct current component correction unit 4 is a signal processing device.

First Embodiment

FIG. 1 shows the structure of an information reading apparatus 100 according to a first embodiment of the invention. The information reading apparatus 100 has a pattern presence/absence evaluation value generation unit 41 , a correction data updating unit 42 , and a correction data subtraction unit 43 . The information reading apparatus 100 may also have a pattern sensing unit 1 , a sensing enable generation unit 2 , an A/D conversion unit 3 , a signal conditioning unit 5 , and a system control unit 6 , or a delay unit 45 . The direct current component correction unit 4 has a pattern presence/absence evaluation value generation unit 41 , a correction data updating unit 42 , a correction data subtraction unit 43 , and a correction data line memory 44 . The direct current component correction unit 4 may also have a delay unit 45 . The direct current component correction unit 4 is a signal processing device.

The pattern sensing unit 1 senses an image pattern of an object that is sensed (an object that is read). The pattern sensing unit 1 outputs a sensing signal d 1 in analog form.

The sensing enable generation unit 2 outputs a sensing enable signal d 2 . The sensing enable signal d 2 distinguishes intervals in which image pattern sensing is and is not performed. For example, it indicates whether or not the object is in the sensing area. The ‘sensing area’ is an area in which pattern information is read.

The A/D conversion unit 3 receives, as its input, the sensing signal d 1 . The A/D conversion unit 3 converts the sensing signal d 1 to sensing data d 3 in digital form.

A sensing data generation unit 10 includes the pattern sensing unit 1 and the A/D conversion unit 3 . The sensing data generation unit 10 senses the image pattern. The sensing data generation unit 10 outputs the sensing data d 3 , representing the sensing result.

The direct current component correction unit 4 receives, as its inputs, the sensing data d 3 and the sensing enable signal d 2 . The direct current component correction unit 4 outputs corrected sensing data d 4 in which a direct current component in a secondary scanning direction Y has been removed from the sensing data d 3 . The direction of motion (direction of transport) of the paper currency 13 with respect to the pattern sensing unit 1 is referred to as the secondary scanning direction.

The signal conditioning unit 5 receives, as its input, the corrected sensing data d 4 . The signal conditioning unit 5 performs signal conditioning on the corrected sensing data d 4 . The signal conditioning unit 5 outputs conditioned sensing data d 5 .

A more detailed description will be given below, taking as an example a case in which the object being read is paper currency, and the information reading apparatus is a magnetic sensing apparatus that detects a magnetic pattern from an image printed with magnetic ink on the paper currency.

The system control unit 6 outputs a system reset signal SRT, a line synchronization signal LSY, a system clock signal SCL, and data indicating a target value TGT.

The line synchronization signal LSY is a signal that aligns timings in one-line cycles. The line synchronization signal LSY is generated once per line cycle. The line synchronization signal LSY is a synchronization signal of a negative polarity.

The system reset signal SRT is a signal for placing the information reading apparatus 100 in its initial state.

The target value TGT is a value of the direct current component of the signal that undergoes direct current component correction in the direct current component correction unit 4 .

The system clock signal SCL is a signal that aligns the timing of the operation of a plurality of circuits by cyclically assuming a high voltage state and a low voltage state.

The line synchronization signal LSY and the system clock signal SCL are supplied to the pattern sensing unit 1 , the A/D conversion unit 3 , the direct current component correction unit 4 , and the signal conditioning unit 5 (the signal lines by which they are supplied are omitted from the drawings). The line synchronization signal LSY and the system clock signal SCL are used to synchronize processing in the pattern sensing unit 1 , the A/D conversion unit 3 , the direct current component correction unit 4 , and the signal conditioning unit 5 .

FIG. 2 illustrates the schematic structure and the operation of the pattern sensing unit 1 . As shown in FIG. 2 , the pattern sensing unit 1 has a line sensor 12 . The line sensor 12 has a plurality of magnetic sensors (sensing elements) 11 which are disposed in a line and whose resistance values change when a magnetic field is applied. The line sensor 12 senses a magnetic pattern provided on the paper currency 13 . The line sensor 12 outputs the sensing signal d 1 .

The plurality of sensing elements 11 of the line sensor 12 sense the magnetic pattern in their order of arrangement. The sequential output of the sensing signal d 1 from the sensing elements 11 is referred to as primary scanning. A wire extends from the bottom of each sensing element 11 in FIG. 2 . The sensing signal d 1 is output through these wires. The longitudinal direction of the line sensor 12 is referred to as a primary scanning direction. The longitudinal direction is the left-right direction in FIG. 2 . The paper currency 13 has regions MG 1 , MG 2 , and MG 3 with magnetic patterns. The region MG 3 is positioned on a sensing line La in the primary scanning direction X in FIG. 2 .

The paper currency 13 is transported by a transport belt 14 or the like. The paper currency 13 passes through the sensing area. The sensing area is the area in which the magnetic pattern is sensed by the pattern sensing unit 1 . The transport belt 14 is shown schematically by dashed lines in FIG. 2 . The paper currency 13 is oriented so that its short side direction is the secondary scanning direction Y and its long side direction is the primary scanning direction X. The direction of motion (direction of transport) of the paper currency 13 with respect to the pattern sensing unit 1 is referred to as the secondary scanning direction. Here, the transport direction of the paper currency 13 is the secondary scanning direction Y. Depending on the device configuration, the pattern sensing unit 1 may move with respect to the paper currency 13 .

The primary scanning is performed taking each sensing element 11 as a unit. The secondary scanning is performed taking each sensing line La as a unit. A ‘sensing line’ is a line-like area that is read in one primary scan. The sensing area is the range within which the magnetic pattern is sensed as the sensing line moves in the secondary scanning direction.

The sensing data generation unit 10 scans the paper currency 13 in the primary scanning direction X with the line sensor 12 while the paper currency 13 is moved in the transport direction with respect to the line sensor 12 . As a result, the sensing data generation unit 10 can obtain data representing a two-dimensional image of the paper currency 13 . Here, the data representing the two-dimensional image the sensing data d 3 .

The paper currency 13 has the regions MG 1 and MG 2 with magnetic patterns. The regions MG 1 and MG 2 are disposed on lines MSa and MSb extending in the secondary scanning direction Y.

The sensing of the magnetic pattern on the paper currency 13 by the pattern sensing unit 1 is carried out in synchronization with the line synchronization signal LSY output from the system control unit 6 . That is, sensing of one sensing line La is performed each time the line synchronization signal LSY occurs. The sensing of one sensing line La consists of sensing by all the sensing elements 11 in the line sensor 12 . Concurrently with the sensing of one sensing line La, the paper currency 13 is transported by one sensing line La in the secondary scanning direction Y.

The sensing signals for a plurality of scanning lines La are output sequentially by repetitions of this process.

When the dimension of the paper currency 13 in the primary scanning direction X is shorter than the longitudinal dimension of the line sensor 12 , the sensing of the magnetic pattern may be carried out only by the sensing elements 11 positioned within the extent of the dimension of the paper currency 13 in the primary scanning direction X. The sensing of the magnetic pattern is the reading of the magnetic pattern. In this case, the primary scanning of one sensing line La is completed by the sensing by all of the sensing elements 11 that are positioned within the extent of the width of the paper currency 13 .

As shown in FIGS. 1 and 2 , the sensing enable generation unit 2 has a paper currency detector 21 . On the basis of the output of the paper currency detector 21 , the sensing enable generation unit 2 generates the sensing enable signal d 2 as shown in FIG. 3 . The sensing enable generation unit 2 outputs the sensing enable signal d 2 .

FIG. 3 illustrates the operation of the sensing enable generation unit 2 . To make it easy to understand the relation between the paper currency 13 and the sensing enable signal d 2 , FIG. 3 depicts both the paper currency 13 and the sensing enable signal d 2 in a single drawing.

The paper currency 13 is depicted in the upper part of FIG. 3 . The primary scanning direction X is the up-down direction in FIG. 3 ; the secondary scanning direction Y is the right-left direction in FIG. 3 . Four pieces of paper currency 13 are arranged in a row in the secondary scanning direction Y.

The sensing enable signal d 2 is depicted in the lower part of FIG. 3 . The left-right direction in FIG. 3 indicates the reading intervals in the secondary scanning direction Y. The up-down direction in FIG. 3 indicates the output value of the sensing enable signal d 2 .

The sensing enable signal d 2 is a rectangular signal. The sensing enable signal d 2 is ‘1’ in the intervals in the secondary scanning direction Y in which the paper currency 13 is positioned. The sensing enable signal d 2 is ‘0’ in the intervals in the secondary scanning direction Y in which the paper currency 13 is not positioned.

The sensing enable signal d 2 is a signal that indicates whether sensing is valid or invalid in the reading intervals in the secondary scanning direction Y. The sensing enable signal d 2 is ‘1’ in the intervals in which the paper currency 13 is passing through the sensing area (the intervals in which the magnetic pattern is sensed). The sensing enable signal d 2 is ‘0’ in the intervals in which the paper currency 13 is not passing through the sensing area.

The intervals in which the paper currency 13 is passing through the sensing area are intervals in which the magnetic pattern sensing is valid. The intervals in which the paper currency 13 is not passing through the sensing area are intervals in which the magnetic pattern sensing is invalid.

FIGS. 4( a ) to 4( e ) are timing diagrams illustrating signals output from the A/D conversion unit 3 in FIG. 1 . In FIGS. 4( a ) to 4( e ) , the horizontal axis indicates time, and the vertical axis indicates the signal level.

Receiving the sensing signal d 1 as its input, the A/D conversion unit 3 operates in synchronization with the system reset signal SRT, the system clock signal SCL, and the line synchronization signal LSY. The system reset signal SRT is shown in FIG. 4( a ) . The system clock signal SCL is shown in FIG. 4( b ) . The line synchronization signal LSY is shown in FIG. 4( c ) . The A/D conversion unit 3 performs A/D conversion on the sensing signal d 1 and outputs the sensing data d 3 in digital form. The sensing data d 3 in digital form is shown in FIG. 4( d ) . The sensing enable signal d 2 is shown in FIG. 4 ( e ) .

When the system is reset, the system reset signal SRT changes from ‘1’ to ‘0’ and then back to ‘1’.

The line synchronization signal LSY changes from ‘1’ to ‘0’ and then back to ‘1’ when the sensing data d 3 for a sensing line La is output. The sensing data d 3 for respective lines La are output in turn. In FIG. 4( d ) , a sensing line La 1 , a sensing line La 2 , and a sensing line La 3 are shown, being aligned in the time axis direction.

The sensing enable signal d 2 is ‘0’ in the sensing lines La 1 and La 2 .

The sensing enable signal d 2 is ‘1’ in and after the sensing line La 3 .

To emphasize that the sensing data d 3 indicates a value for each sensing element 11 , reference characters d 4 ( i ) will sometimes be used below. Reference character i indicates that the data is associated with the i-th sensing element 11 in the line sensor 12 . The same is true of other data to which reference character i is affixed. To indicate the data of each sensing line La, a ‘(j)’ will sometimes be added. Reference character j indicates that the data is associated with the j-th sensing line La in the secondary scanning direction Y. The ‘(i)’ or ‘(j)’ will be omitted when it is not necessary to emphasize that the data is associated with individual sensing elements 11 or sensing lines La.

The direct current component correction unit 4 receives, as its inputs, the sensing data d 3 ( i ) and the sensing enable signal d 2 . The direct current component correction unit 4 outputs the corrected sensing data d 4 ( i ) from which the direct current component has been removed.

The direct current component correction unit 4 has the pattern presence/absence evaluation value generation unit 41 , the correction data updating unit 42 , the correction data subtraction unit 43 , and the correction data line memory 44 . The direct current component correction unit 4 may also have the delay unit 45 .

The pattern presence/absence evaluation value generation unit 41 compares the sensing data d 3 ( i ) and correction data d 44 ( i ) and decides whether a pattern is present or absent. The sensing data d 3 ( i ) is sensing data obtained by each sensing element 11 (the i-th sensing element). The correction data d 44 ( i ) is data pertaining to the i-th sensing element 11 , held in the correction data line memory 44 . The pattern presence/absence evaluation value d 41 represents the degree of inclusion of signal components corresponding to a magnetic pattern. The pattern presence/absence evaluation value generation unit 41 outputs the pattern presence/absence evaluation value d 41 . In the following description, the correction data line memory 44 is used to hold the correction data, but the memory need not be a line memory as long as it is a memory that stores data; what follows is an example.

The decision as to pattern presence or absence is a decision as to the degree of inclusion of signal components corresponding to a magnetic pattern. On the basis of the pattern presence/absence decision, the pattern presence/absence evaluation value generation unit 41 outputs an evaluation value (the pattern presence/absence evaluation value d 41 ) representing the degree of inclusion of signal components corresponding to a magnetic pattern.

A value indicating a variation component of the sensing data d 3 of each sensing element 11 , for example, is used as the pattern presence/absence evaluation value d 41 . More specifically, as the value indicating the variation component of the sensing data of each sensing element 11 , the difference of the data value of the sensing data d 3 with respect to the correction data d 44 or the amount of change in the data value is used. Alternatively, as the value indicating the variation component of the sensing data of the sensing element 11 , a value obtained by summarizing the differences of the data values of the sensing data d 3 with respect to the correction data d 44 or the amounts of change in the data value is used. The value obtained by summarization is, for example, the mean value of the absolute values of the differences over the entire sensing line La. The ‘difference’ is a difference between two values.

The pattern presence/absence evaluation value generation unit 41 determines, for example, the mean value MAD(j) of the absolute values of the differences between the sensing data d 3 ( i ) and the correction data d 44 ( i ) of the same sensing element 11 for each sensing line La (the mean value thus determined will be referred to as the line mean value MAD(j) below). The pattern presence/absence evaluation value generation unit 41 outputs this line mean value MAD(j) as the pattern presence/absence evaluation value d 41 ( j ). The ‘absolute value of the difference’ is the absolute value of the difference between two values.

As described above, the absolute value of the difference between the sensing data d 3 ( i ) and the correction data d 44 ( i ) pertaining to the same sensing element 11 could be said to represent the amount of change of the sensing data d 3 ( i ) with respect to the correction data d 44 ( i ). The correction data d 44 ( i ) represents the direct current component included in the sensing data d 3 for each sensing element 11 , determined from the sensing results up to the preceding sensing line La. The ‘preceding sensing line’ is the sensing line La sensed one line before the sensing line La currently being sensed.

When the amount of change in the sensing data d 3 ( i ) with respect to the correction data d 44 ( i ) representing the direct current component is large, it can be inferred that many signal components corresponding to a magnetic pattern are included in the sensing line La currently being sensed. Conversely, when the amount of change in the sensing data d 3 ( i ) with respect to the correction data d 44 ( i ) representing the direct current component is small, it can be inferred that only a few signal components corresponding to a magnetic pattern are included in the sensing line La currently being sensed.

The correction data updating unit 42 takes a weighted sum of the sensing data d 3 ( i ) and the correction data d 44 ( i ). The correction data updating unit 42 takes this weighted sum on the basis of the pattern presence/absence evaluation value d 41 ( j ) and the sensing enable signal d 2 . The pattern presence/absence evaluation value d 41 ( j ) is the value output by the pattern presence/absence evaluation value generation unit 41 . The correction data updating unit 42 outputs the result of this addition as new correction data d 42 ( i ). The ‘new correction data’ is updated correction data.

The correction data updating unit 42 takes the weighted sum of the sensing data d 3 ( i ) and the correction data d 44 ( i ) when the sensing enable signal d 2 is ‘1’. The correction data updating unit 42 obtains the updated correction data d 42 ( i ). When, on the basis of the pattern presence/absence evaluation value d 41 ( j ), there are many signal components corresponding to the magnetic pattern, the correction data updating unit 42 increases the weight of the correction data d 44 . When there are few signal components corresponding to the magnetic pattern, the correction data updating unit 42 increases the weight of the sensing data d 3 ( i ).

The interval in which the sensing enable signal d 2 is ‘1’ is the interval in which the magnetic pattern sensing is valid. The pattern presence/absence evaluation value d 41 ( j ) is the result of the decision output by the pattern presence/absence evaluation value generation unit 41 .

When the sensing enable signal d 2 is ‘0’ (the interval in which the magnetic pattern sensing is invalid), the correction data updating unit 42 increases the weight of the sensing data d 3 ( i ) regardless of the pattern presence/absence evaluation value d 41 ( j ) of the pattern presence/absence evaluation value generation unit 41 . The pattern presence/absence evaluation value d 41 ( j ) is the result of the pattern presence/absence decision of the pattern presence/absence evaluation value generation unit 41 .

The correction data updating unit 42 takes a weighted sum of the weighted sensing data d 3 ( i ) and the correction data d 44 ( i ). The correction data updating unit 42 thereby obtains the updated correction data d 42 ( i ).

FIG. 5 is a block diagram showing an example of the structure of the correction data updating unit 42 .

The correction data updating unit 42 has a weighting coefficient generation unit 421 , a correction data computation unit 422 , and a parameter setting unit 423 .

The weighting coefficient generation unit 421 receives, as its inputs, the pattern presence/absence evaluation value d 41 ( j ) and the sensing enable signal d 2 . The weighting coefficient generation unit 421 outputs a sensing data weighting coefficient K 3 ( j ) and a correction data weighting coefficient K 44 ( j ) to the correction data computation unit 422 .

When the sensing enable signal d 2 is ‘1’ (the interval in which the magnetic pattern sensing is valid), the weighting coefficient generation unit 421 calculates the correction data weighting coefficient K 44 ( j ) from the relation shown in the following mathematical expression (1).

[Mathematical Expression 1]

The following definitions apply here in expression (1).

TH 1 is a first threshold value.

TH 2 is a second threshold value.

Kmax is the maximum value of the correction data weighting coefficient.

Kmin is the minimum value of the correction data weighting coefficient.

The first threshold value TH 1 and the second threshold value TH 2 satisfy the relation TH 1 ≤TH 2 .

The maximum value Kmax of the correction data weighting coefficient and the minimum value Kmin of the correction data weighting coefficient satisfy the relation 0≤Kmin≤Kmax≤1.

That is, when the pattern presence/absence evaluation value is less than the first threshold value, the weighting coefficient generation unit 421 outputs the minimum value of the correction data weighting coefficient as the correction data weighting coefficient.

When the pattern presence/absence evaluation value is greater than the second threshold value, the weighting coefficient generation unit 421 outputs the maximum value of the correction data weighting coefficient as the correction data weighting coefficient.

When the pattern presence/absence evaluation value is equal to or greater than the first threshold value and equal to or less than the second threshold value, the weighting coefficient generation unit 421 outputs, as the correction data weighting coefficient, a value that increases monotonically with respect to the pattern presence/absence evaluation value in the interval from the minimum value of the correction data weighting coefficient to the maximum value of the correction data weighting coefficient.

The relation in expression (1) is shown in FIG. 6 . FIG. 6 illustrates the relation between the input and the output of the weighting coefficient generation unit 421 . The horizontal axis represents the pattern presence/absence evaluation value d 41 . The vertical axis represents the weighting coefficient K 44 .

The operation of the weighting coefficient generation unit 421 illustrated by expression (1) and FIG. 6 will now be described. The interval in which the sensing enable signal d 2 is ‘1’ is the interval in which the magnetic pattern sensing is performed (the valid interval of the magnetic pattern sensing). The interval in which the sensing enable signal d 2 is ‘0’ is the interval in which the magnetic pattern sensing is not performed (the invalid interval of the magnetic pattern sensing).

First the operation when the sensing enable signal d 2 is ‘1’ will be described.

In the operation of the weighting coefficient generation unit 421 , there are three cases, as follows.

As the first case, when the pattern presence/absence evaluation value d 41 ( j ) is less than the first threshold value TH 1 , the weighting coefficient generation unit 421 finds that there is no magnetic pattern. The weighting coefficient generation unit 421 outputs the minimum value Kmin of the correction data weighting coefficient as the correction data weighting coefficient K 44 ( j ). ‘No magnetic pattern’ means that negligibly few signal components corresponding to a magnetic pattern are included.

As the second case, when the pattern presence/absence evaluation value d 41 ( j ) is greater than the second threshold value TH 2 , the weighting coefficient generation unit 421 finds that a magnetic pattern is present. The weighting coefficient generation unit 421 outputs the maximum value Kmax of the correction data weighting coefficient as the correction data weighting coefficient K 44 ( j ). ‘A magnetic pattern is present’ means that many signal components corresponding to a magnetic pattern are included.

As the third case, when the pattern presence/absence evaluation value d 41 ( j ) is equal to or greater than the first threshold value TH 1 and equal to or less than the second threshold value TH 2 , the weighting coefficient generation unit 421 outputs a value determined by a characteristic that increases monotonically from the minimum value Kmin to the maximum value Kmax of the correction data weighting coefficient as the correction data weighting coefficient K 44 ( j ).

Next the operation when the sensing enable signal d 2 is ‘0’ will be described.

The weighting coefficient generation unit 421 calculates the correction data weighting coefficient K 44 ( j ) from the relation indicated by the following expression (2).

K 44( j )= Krst   (2)

Here, Krst is a parameter indicating the value of the correction data weighting coefficient K 44 ( j ) during invalid magnetic pattern sensing intervals. Krst satisfies the relation 0≤Krst≤Kmin.

As shown by expression (2), when the sensing enable signal d 2 is ‘0’, the weighting coefficient generation unit 421 outputs the correction data weighting coefficient K 44 ( j ) without regard to the value of the pattern presence/absence evaluation value d 41 ( j ). The value of the correction data weighting coefficient K 44 ( j ) at this time is the value Krst. The value Krst is equal to or less than the value Kmin. The sensing enable signal d 2 is ‘0’ in intervals in which the magnetic pattern sensing is invalid.

The weighting coefficient generation unit 421 also calculates the sensing data weighting coefficient K 3 ( j ) that sums to 1 when added to the correction data weighting coefficient K 44 ( j ), from the relation shown by expression (3)

K 3( j )=1− K 44( j )  (3)

As described above, the weighting coefficient generation unit 421 generates the correction data weighting coefficient K 44 ( j ) that increases when the pattern presence/absence evaluation value d 41 ( j ) is large, and the sensing data weighting coefficient K 3 ( j ) whose sum with the correction data weighting coefficient K 44 ( j ) is 1.

Next, the correction data computation unit 422 takes a weighted sum of the sensing data d 3 ( i ) and the correction data d 44 ( i ) according to the expression (4) below. The correction data computation unit 422 then outputs the updated correction data d 42 ( i ). In taking the weighted sum, the correction data computation unit 422 uses the weighting coefficients K 3 ( j ) and K 44 ( j ).

d 42( i )= K 44( j )× d 44( i )+ K 3( j )× d 3( i )  (4)

As shown in the above expression, the correction data computation unit 422 multiplies the correction data d 44 ( i ) by the correction data weighting coefficient K 44 ( j ), multiplies the sensing data d 3 ( i ) by the sensing data weighting coefficient K 3 ( j ), and performs addition (weighted addition) on the multiplication results.

The parameter setting unit 423 holds the parameters TH 1 , TH 2 , Kmax, Kmin, and Krst, which are set by the user in advance. The parameter setting unit 423 supplies the parameters TH 1 , TH 2 , Kmax, Kmin, and Krst to the weighting coefficient generation unit 421 . Setting of the parameters TH 1 , TH 2 , Kmax, Kmin, and Krst is not limited to the user. They may be set by the manufacturer on the production line, or elsewhere. That is, the parameters TH 1 , TH 2 , Kmax, and Krst are set from the outside.

The first threshold TH 1 and the second threshold TH 2 are preset in consideration of the signal-to-noise ratio of the information reading apparatus. The purpose of this is to do the utmost to prevent the occurrence of incorrect signal decisions or failure to detect the signal because of noise.

The maximum value Kmax of the correction data weighting coefficient is preset to 1 or a value close to 1. The purpose of this is to give a large weight to the correction data d 44 ( i ) in the updating of the correction data when it is decided that signal components corresponding to a magnetic pattern are present. Here, ‘signal components are present’ means that sufficiently many signal components are present.

The minimum value Kmin of the correction data weighting coefficient is preset in consideration of the rapidity of changes in the sensing signal level and the frequency characteristic of the signal. The purpose of this is to ensure that while changes in the sensing signal level are reflected in the updating of the correction data when it is decided that signal components corresponding to a magnetic pattern are absent, the low frequency components of the sensing signal are not removed. Here, ‘signal components are absent’ means that negligibly few signal components are present.

The correction data weighting coefficient Krst in the invalid intervals of the magnetic pattern sensing is preset to a value equal to or less than Kmin. The purpose of this is to give the sensing data d 3 ( i ) greater weight over the correction data.

The apparatus is configured to enable these parameters to be updated from the outside by, for example, the user.

The correction data updating unit 42 outputs the updated correction data d 42 ( i ) by means of the above configuration.

The correction data subtraction unit 43 subtracts the correction data d 42 ( i ) updated by the correction data updating unit 42 from the sensing data d 3 ( i ). The correction data subtraction unit 43 outputs the result of this subtraction as corrected sensing data d 43 ( i ).

The correction data line memory 44 holds the updated correction data d 42 ( i ) for one sensing line La. The updated correction data d 42 ( i ) is data that have been updated by the correction data updating unit 42 . The updated correction data d 42 ( i ) held in the correction data line memory 44 are used in the processing of the next sensing line La by the pattern presence/absence evaluation value generation unit 41 and the correction data updating unit 42 as the correction data d 44 ( i ).

By means of the above configuration, the direct current component correction unit 4 outputs the corrected sensing data d 43 ( i ) as the direct current component corrected sensing data d 4 ( i ).

The direct current component corrected sensing data d 4 ( i ) output from the direct current component correction unit 4 are supplied to the signal conditioning unit 5 . The signal conditioning unit 5 includes, for example, a gain adjustment circuit that adjusts the gain or a noise reduction circuit that performs noise reduction. The signal conditioning unit 5 carries out signal conditioning such as gain adjustment or noise reduction on the corrected sensing data d 4 ( i ). Then the signal conditioning unit 5 outputs the conditioned sensing data d 5 ( i ).

The operation of the direct current component correction unit 4 in the information reading apparatus 100 according to the first embodiment will be described below with reference to FIG. 7 .

FIG. 7 is a flowchart illustrating the operation of the direct current component correction unit 4 .

The processing by the direct current component correction unit 4 is carried out one sensing line La at a time. First, the sensing data d 3 ( i ) for one sensing line La is input in a step S 0 .

In a correction data reading step S 1 , the correction data d 44 ( i ) for one sensing line La held in the correction data line memory 44 is read.

Next, in a pattern presence/absence decision step S 2 , the line mean value MAD(j) of the absolute values of the differences between the sensing data d 3 ( i ) and the correction data d 44 ( i ) for the same sensing elements 11 is determined. In the pattern presence/absence decision step S 2 , this line mean value MAD(j) is output as the pattern presence/absence evaluation value d 41 ( j ) of the sensing line La that is undergoing signal processing. The ‘line mean value’ is a value obtained by dividing a sum of values based on the outputs of the sensing elements arranged in a line, by the number of the sensing elements. That is, it is the mean value of the values sensed in the sensing line La.

Next, in a correction data updating step S 3 , the correction data d 44 ( i ) based on the pattern presence/absence evaluation value d 41 ( j ) are updated. This update is performed for all of the sensing elements in the sensing line La that is undergoing signal processing.

In the correction data updating step S 3 , first, the weighting coefficients K 3 ( j ) and K 44 ( j ) are generated based on the pattern presence/absence evaluation value d 41 ( j ) and the sensing enable signal d 2 in a weighting coefficient generation step S 31 .

FIG. 8 is a flowchart illustrating details of the weighting coefficient generation step S 31 .

In a step S 310 , the sensing enable signal d 2 of the sensing line La undergoing signal processing is checked. If, in the step S 310 , the sensing enable signal d 2 of the sensing line La undergoing signal processing is ‘1’ (a valid magnetic pattern sensing interval; Yes in the conditional branch in the step S 310 ), the process proceeds to a step S 311 . In the step S 311 , the weighting coefficient K 44 ( j ) is generated from the relational expression (expression (1)) dependent on the pattern presence/absence evaluation value d 41 ( j ), and the process proceeds to a step S 312 . Next, in the step S 312 , the weighting coefficient K 3 ( j ) is generated from expression (3).

Conversely, if the sensing enable signal d 2 of the sensing line La undergoing signal processing is ‘0’ (an invalid magnetic pattern sensing interval; No in the conditional branch in the step S 310 ), the weighting coefficient K 44 ( j ) is generated from the relational expression (expression (2)) that does not depend on the pattern presence/absence evaluation value d 41 ( j ), and the process proceeds to the step S 312 . Then in the step S 312 , the weighting coefficient K 3 ( j ) is generated from expression (3).

The description now returns to FIG. 7 .

In a correction data computation step S 32 , the generated weighting coefficients K 3 ( j ) and K 44 ( j ) are used to take a weighted sum of the sensing data d 3 ( i ) and the correction data d 44 ( i ) according to expression (4). The value of the weighted sum is output as the updated correction data d 42 ( i ).

Here, the weighted sum is taken of the sensing data d 3 ( i ) and the correction data d 44 ( i ) for the same sensing element 11 . The processing of one sensing line La in the correction data computation step S 32 is completed by taking the above weighted sum for all sensing elements 11 in one sensing line La.

After the correction data for one sensing line La have been updated in the correction data update step S 3 , the updated correction data d 42 ( i ) is written into the correction data line memory 44 in a correction data writing step S 4 .

Next, the updated correction data d 42 ( i ) is subtracted from the sensing data d 3 ( i ) in a step S 5 . The corrected sensing data d 4 ( i ) is obtained by this subtraction in the step S 5 . The subtraction is performed here on the sensing data d 3 ( i ) and the updated correction data d 42 ( i ) for the same sensing element 11 .

Next, the corrected sensing data d 4 ( i ) obtained in the step S 5 is output in a step S 6 . The processing in the step S 6 is also carried out one sensing line La at a time. The corrected sensing data for one sensing line La is output in the step S 6 . ‘One sensing line La at a time’ means for all sensing elements in the sensing line La undergoing signal processing.

When the processing of one sensing line La is completed, a return is made to the step S 0 and the processing of the next sensing line La begins. When similar processing has been completed for all sensing lines La (Yes in the conditional branch in a step S 7 ), the process is terminated.

The delay unit 45 indicated by the dashed line in FIG. 1 may further be provided.

In FIG. 1 , the delay unit 45 (indicated by the dashed line) is located on the input side of the correction data subtraction unit 43 . That is, the sensing data d 3 ( i ) may be input to the correction data subtraction unit 43 through the delay unit 45 . The purpose of this is to subtract the updated correction data d 42 ( i ) from the sensing data d 3 ( i ) that was used to update them. In this case, the delay time due to the delay unit 45 is equivalent to the time needed for the processing done by the pattern presence/absence evaluation value generation unit 41 and the correction data updating unit 42 .

The above is a description of the operation of the direct current component correction unit 4 in the information reading apparatus 100 according to the first embodiment of the present invention.

Next, the effects due to the direct current component correction unit 4 in the information reading apparatus 100 according to the first embodiment of the present invention will be described by use of FIGS. 9( a ) to 9( d ) and 10( a ) to 10( d ) .

FIGS. 9( a ) to 9( d ) and 10( a ) to 10( d ) show profiles of data for different lines MSa and MSb extending in the secondary scanning direction Y on the paper currency 13 in FIG. 2 . In each of FIGS. 9( a ) to 9( d ) and 10( a ) to 10( d ) , the horizontal axis indicates the secondary scanning direction Y and the vertical axis indicates the signal value.

FIGS. 9( a ) and 10( a ) show the sensing data d 3 ( i ) (i=a, b) for the respective lines MSa, MSb in the secondary scanning direction Y. As shown in FIG. 2 , the sensing element 11 positioned on the line MSa is a different sensing element 11 from the sensing element 11 positioned on the line MSb. That is, FIGS. 9( a ) and 10( a ) represent the sensing data d 3 of different sensing elements 11 , in correspondence to position in the secondary scanning direction Y. These sensing data d 3 ( i ) are respective sequences of the sensing data d 3 ( i ) read by the different sensing elements 11 ( a ), 11 ( b ) in the line sensor 12 .

FIGS. 9( b ) and 10( b ) show the updated correction data d 42 ( i ) (i=a, b) for the lines MSa and MSb in the secondary scanning direction Y.

FIGS. 9( c ) and 10( c ) show the direct current component corrected sensing data d 4 ( i ) (i=a, b) for the lines MSa and MSb in the secondary scanning direction Y.

FIGS. 9( d ) and 10( d ) show the sensing enable signal d 2 .

In the lines MSa and MSb in the secondary scanning direction Y shown in FIG. 2 , there are the regions MG 1 and MG 2 having magnetic patterns due to printing with magnetic ink on the paper currency 13 . As shown in FIGS. 9( a ) and 10( a ) , signal components corresponding to magnetic patterns appear at positions SG 1 and SG 2 corresponding to these regions.

In the line MSa in the secondary scanning direction Y, the level of the sensing signal tends to increase gradually from one end to the other. In the line MSb in the secondary scanning direction Y, the level of the sensing signal tends to decrease gradually from one end to the other.

In FIG. 2 , ‘from one end to the other’ means ‘from the top end to the bottom end’. In FIGS. 9( a ) to 9( d ) and 10( a ) to 10( d ) , ‘from one end to the other’ means ‘from the left end to the right end’.

The level of the sensing signal from a sensing element 11 (a magnetic sensor element) may vary because of temperature and other factors during the interval in which one piece of paper currency is read. The way in which the level of the sensing signal varies may also differ from one sensing element 11 to another. Therefore, the non-uniformity in the sensing signal level in the primary scanning direction X appears as stripe noise in the secondary scanning direction Y in the sensing data d 3 ( i ).

The offset correction method shown in patent reference 1 calculates an offset correction value only in the intervals between one piece of paper currency and another. That is, the offset correction method shown in patent reference 1 calculates an offset correction value in the invalid magnetic pattern sensing intervals. Therefore, when the output level varies during the reading of one piece of paper currency, even if the non-uniformity in the sensing signal level in the primary scanning direction X is suppressed near the leading edge of the paper currency, output levels that are not uniform in the primary scanning direction X may occur near the trailing edge of the paper currency.

Here ‘near the leading edge of the paper currency’ denotes the region of a piece of paper currency in which reading is begun. ‘Near the trailing edge of the paper currency’ denotes the region of a piece of paper currency in which reading is terminated. That is, there has been the problem of lowered paper currency identification accuracy because the offset correction method in patent reference 1 is unable to fully eliminate stripe noise.

The present invention seeks to solve this problem by updating the offset correction value during the valid magnetic pattern sensing intervals, as described above. The offset correction value is weighted according to the presence or absence of signal components corresponding to a magnetic pattern.

This point will be explained below.

The updated correction data d 42 ( i ) shown in FIGS. 9( b ) and 10( b ) represents the result of determining the direct current component of the sensing data on the basis of the sensing data d 3 ( i ) from the start of reading up to the sensing line La (a line in the primary scanning direction X) one line before.

The direct current component does not change greatly between the sensing line La in the primary scanning direction X one line before and the current sensing line La in the primary scanning direction X. The direct current component of the sensing data d 3 ( i ) is therefore the same in the sensing line La in the primary scanning direction X one line before and the current sensing line La in the primary scanning direction X. Here, the ‘direct current component’ is the value obtained by subtracting the signal components corresponding to the magnetic pattern from the sensing data d 3 ( i ) of the current sensing line La in the primary scanning direction X.

In the direct current component correction unit 4 in the information reading apparatus 100 in the first embodiment, when the sensing enable signal d 2 is ‘0’ (in an invalid magnetic pattern sensing interval), the weighting of the sensing data d 3 ( i ) in the generation of the updated correction data d 42 ( i ) is increased. At the starting point of the reading of the paper currency, the value of the updated correction data d 42 ( i ) therefore agrees with the value of the sensing data d 3 ( i ).

There are times when it is already known, from an external signal such as the sensing enable signal d 2 , that there are no signal components corresponding to a magnetic pattern. At these times the correction data updating unit 42 increases the weighting of the sensing data d 3 ( i ), regardless of the pattern presence/absence evaluation value d 41 ( j ) from the pattern presence/absence evaluation value generation unit 41 . This improves the tracking characteristics of the updated correction data d 42 ( i ) with respect to the sensing data d 3 . The correction data d 44 ( i ) can also be initialized in the brief time between two pieces of paper currency 13 .

On entry into an interval in which the sensing enable signal d 2 is ‘1’ (a valid magnetic pattern sensing interval), the direct current component correction unit 4 determines the weighting coefficients according to the pattern presence/absence evaluation value d 41 ( j ) from the pattern presence/absence evaluation value generation unit 41 . The pattern presence/absence evaluation value d 41 ( j ) indicates the degree of inclusion of a magnetic pattern in the sensing data d 3 ( i ). The direct current component correction unit 4 carries out an update of the correction data, using the weighting coefficients K 3 and K 44 having been determined.

When no magnetic pattern is present, the direct current component correction unit 4 increases the weighting of the sensing data d 3 ( i ) and takes a weighted sum of the sensing data d 3 ( i ) and the correction data d 44 ( i ). Then it further updates the updated correction data d 42 ( i ). The updated correction data d 42 ( i ) accordingly approach the sensing data d 3 ( i ).

When a magnetic pattern is present, however, the direct current component correction unit 4 increases the weighting of the correction data d 44 ( i ) and takes a weighted sum of the sensing data d 3 ( i ) and the correction data d 44 ( i ). Then it further updates the updated correction data d 42 ( i ). The updated correction data d 42 ( i ) accordingly approach the correction data d 44 ( i ) before the correction. This means that the correction data d 44 ( i ) determined on the basis of the sensing data d 3 ( i ) in previous sensing lines La are maintained.

In cases intermediate between a situation in which ‘a magnetic pattern is present’ and a situation in which ‘there is no magnetic pattern’, the direct current component correction unit 4 determines the weighting coefficients K 3 ( j ) and K 44 ( j ) according to the pattern presence/absence evaluation value d 41 ( j ). “Cases intermediate between a situation in which ‘a magnetic pattern is present’ and a situation in which ‘there is no magnetic pattern’” are cases in which the line mean value MAD(j) of the absolute values is equal to or greater than threshold value TH 1 and equal to or less than threshold value TH 2 . ‘According to the pattern presence/absence evaluation value d 41 ( j )’ means ‘according to the degree of inclusion of signal components corresponding to a magnetic pattern’. The direct current component correction unit 4 uses the determined weighting coefficients K 3 , K 44 to update the updated correction data d 42 ( i ).

A one-bit signal, for example, can express only one of two states.

If the pattern presence/absence evaluation value d 41 ( j ) were to be just one bit, for example, it could only express one of ‘magnetic pattern present’ or ‘magnetic pattern absent’. In the first embodiment, the pattern presence/absence evaluation value d 41 ( j ) is not a value (a one-bit value) that can indicate only either ‘magnetic pattern present’ or ‘magnetic pattern absent’. A value that varies continuously according to the degree of inclusion of signal components corresponding to a magnetic pattern is used as the pattern presence/absence evaluation value d 41 ( j ). The weighting coefficient generation unit 421 is adapted to generate the weighting coefficients K 44 ( j ), K 3 ( j ) that vary continuously according to the pattern presence/absence evaluation value d 41 ( j ). The direct current component correction unit 4 can thereby perform a suitable update according to the degree of inclusion of signal components corresponding to a magnetic pattern.

When there is nothing but either ‘magnetic pattern present’ or ‘magnetic pattern absent’, at locations of switchovers between the two, step edges may occur in the two-dimensional image represented by the corrected sensing data d 4 . Such step edges tend to be avoided in the first embodiment by processing according to the degree of pattern presence/absence.

The direct current component corrected sensing data d 4 ( i ) shown in FIG. 9( c ) is obtained by subtracting the updated correction data d 42 ( i ) shown in FIG. 9( b ) from the sensing data d 3 ( i ) shown in FIG. 9( a ) .

Similarly, the direct current component corrected sensing data d 4 ( i ) shown in FIG. 10( c ) is obtained by subtracting the updated correction data d 42 ( i ) shown in FIG. 10( b ) from the sensing data d 3 ( i ) shown in FIG. 10( a ) .

As shown in FIGS. 9( c ) and 10( c ) , signal components which correspond to the magnetic patterns in the regions MG 1 and MG 2 and which are present in the segments SG 1 and SG 2 are not lost in the direct current component corrected sensing data d 4 ( i ).

As also shown in FIGS. 9( c ) and 10( c ) , in places (other than SG 1 and SG 2 ) where signal components corresponding to the magnetic patterns are absent, the corrected sensing data d 4 ( i ) are aligned at the level of the target value TGT.

The direct current component correction unit 4 thereby suppresses non-uniformity in the level of the sensing signal in the primary scanning direction X in places where signal components corresponding to a magnetic pattern are absent, even when the level of the sensing signal varies within the reading interval of one piece of paper currency. The information reading apparatus 100 according to the first embodiment of the present invention can thereby reduce stripe noise in the secondary scanning direction Y.

As described above, according to the first embodiment of the invention, correction data can be updated within the interval in which one piece of paper currency is being read. The first embodiment of the present invention can therefore reduce the effect of variations in the level of the sensing signal that occur in the interval while each object that undergoes sensing is being read. The first embodiment of the present invention can also obtain the sensing data in which the non-uniformity in the level is adequately corrected.

The signal processing device 4 has the pattern presence/absence evaluation value generation unit 41 , the correction data updating unit 42 , and the correction data subtraction unit 43 .

On the basis of the sensing data d 3 including signal components of the object being sensed, obtained by sequential sensing of the object sensed by the sensing elements 11 , the pattern presence/absence evaluation value generation unit 41 calculates the degree of inclusion of the signal components in the sensing data d 3 .

The correction data updating unit 42 generates updated correction data d 42 by taking a weighted sum of the sensing data d 3 and the correction data d 44 representing a direct current component of the sensing elements, giving greater weight to the correction data d 44 as the degree increases and giving greater weight to the sensing data d 3 as the degree decreases.

The correction data subtraction unit 43 subtracts the updated correction data d 42 from the sensing data d 3 to generate the corrected sensing data.

Second Embodiment

In the first embodiment, the pattern presence/absence evaluation value generation unit 41 uses the line mean value MAD(j) of the absolute values of the differences between the sensing data d 3 ( i ) and the correction data d 44 ( i ) of identical sensing elements 11 as the pattern presence/absence evaluation value d 41 ( j ). In the second embodiment, the pattern presence/absence evaluation value d 41 ( j ) is determined by a different method.

Component elements that are the same as in the information reading apparatus 100 in the first embodiment will be given the same reference characters and descriptions will be omitted. The component elements that are the same as in the first embodiment are the sensing data generation unit 10 (the pattern sensing unit 1 and the A/D conversion unit 3 ), the sensing enable generation unit 2 , the correction data updating unit 42 , the correction data line memory 44 , the correction data subtraction unit 43 , the delay unit 45 , the system control unit 6 , and the signal conditioning unit 5 . The same reference characters 100 as in the first embodiment will also be used in describing the information reading apparatus.

In the second embodiment, the pattern presence/absence evaluation value generation unit 41 uses, for example, the in-line standard deviation s(j) of the differences of the sensing data d 3 ( i ) and the correction data d 44 ( i ) of identical sensing elements 11 as the pattern presence/absence evaluation value d 41 ( j ).

The in-line standard deviation s(j) can be determined from the expression (5) below.

[Mathematical Expression 2]

The following definitions apply in expression (5).

d 3 ( i ) is the sensing data for the i-th sensing element 11 in each sensing line La.

d 44 ( i ) is the correction data for the same i-th sensing element 11 .

I is the total number of sensing elements 11 in the sensing line La.

In the above case, before the information reading apparatus 100 reads information, the pattern presence/absence evaluation value generation unit 41 calculates the standard deviation of the noise. The pattern presence/absence evaluation value generation unit 41 uses the calculated standard deviation of the noise as the above first threshold value TH 1 .

In the information reading operation of the information reading apparatus 100 , when the in-line standard deviation s(j) is equal to or greater than the first threshold TH 1 , it can be inferred that many signal components corresponding to a magnetic pattern are included in this sensing line La.

When the in-line standard deviation s(j) is less than the first threshold TH 1 in the information reading operation of the information reading apparatus 100 , it can be inferred that few signal components corresponding to a magnetic pattern are included in this sensing line La.

The standard deviation is often used to evaluate noise levels. Calculation of the in-line standard deviation s(j) as the pattern presence/absence evaluation value d 41 ( j ) facilitates the setting of the threshold TH 1 for pattern presence/absence decision, taking account of the non-uniformity in the level of the sensing signal due to noise.

Third Embodiment

In the first embodiment, the pattern presence/absence decision by the pattern presence/absence evaluation value generation unit 41 and the generation of the weighting coefficients K 3 , K 44 by the correction data updating unit 42 are carried out for each sensing line La. In the third embodiment, however, the generation of the weighting coefficients K 3 , K 44 is carried out for each sensing element 11 .

In this case, a direct current component correction unit 4 b shown in FIG. 11 is used instead of the direct current component correction unit 4 in FIG. 1 . In addition, a correction data updating unit 42 b shown in FIG. 12 is used instead of the correction data updating unit 42 in FIG. 5 .

FIG. 11 is a block diagram illustrating the information reading apparatus 110 . FIG. 12 is a block diagram showing an example of the configuration of the correction data updating unit 42 b . Component elements that are the same as in the information reading apparatus 100 in the first embodiment will be given the same reference characters and descriptions will be omitted. The component elements that are the same as in the first embodiment are the sensing data generation unit 10 (the pattern sensing unit 1 and the A/D conversion unit 3 ), the sensing enable generation unit 2 , the correction data line memory 44 , the delay unit 45 , the system control unit 6 , and the signal conditioning unit 5 .

The configurations shown in FIGS. 11 and 12 have generally the same block structure and the input-output relationships as in FIGS. 1 and 5 , but differ in the following points.

The pattern presence/absence evaluation value generation unit 41 in the first embodiment outputs the line mean value MAD(j) of the absolute values of the differences between the sensing data d 3 ( i ) and the correction data d 44 ( i ) for the same sensing elements as the pattern presence/absence evaluation value d 41 ( j ).

In contrast, a pattern presence/absence evaluation value generation unit 41 b outputs the absolute value of the difference between the sensing data d 3 ( i ) and the correction data d 44 ( i ) for the same sensing element as the pattern presence/absence evaluation value d 41 ( i ) of this sensing element 11 .

The weighting coefficient generation unit 421 in the correction data updating unit 42 in the first embodiment generates the correction data weighting coefficient K 44 ( j ) and the sensing data weighting coefficient K 3 ( j ) for each sensing line La according to the pattern presence/absence evaluation value d 41 ( j ) of the sensing line La.

In contrast, a weighting coefficient generation unit 421 b in the correction data updating unit 42 b generates a correction data weighting coefficient K 44 ( i ) and a sensing data weighting coefficient K 3 ( i ) for each sensing element 11 according to the pattern presence/absence evaluation value d 41 ( i ) of the sensing element 11 .

The correction data computation unit 422 in the first embodiment performs a computation using the weighting coefficients K 44 ( j ), K 3 ( j ) generated for each sensing line La. In the above computation, the correction data d 44 ( i ) is weighted with the weighting coefficient K 44 ( i ), the sensing data d 3 ( i ) is weighted with the weighting coefficient K 3 ( j ), and the sum of these weighted data is taken.

In contrast, the correction data computation unit 422 b performs a computation using the weighting coefficients K 44 ( i ), K 3 ( i ) generated for each sensing element 11 . The difference from the first embodiment is that the weighting coefficients K 44 ( i ), K 3 ( i ) differ for each sensing element 11 . The computation using the weighting coefficients K 44 ( i ), K 3 ( i ) is the same as in the first embodiment, however. The updated correction data d 42 ( i ) is obtained by the computation performed in the correction data computation unit 422 b.

Since the correction data updating unit 42 b determines the weighting coefficients K 3 , K 44 for each sensing element 11 , when signal components corresponding to a magnetic pattern are obtained only from a region in one part of a sensing line La, different methods of updating the correction data can be used in the region from which signal components corresponding to a magnetic pattern are obtained and in other regions.

When the region from which signal components corresponding to a magnetic pattern are obtained occupies only part of a sensing line La, a decision made for each sensing line La will find little inclusion of signal components corresponding to the magnetic pattern in this sensing line La. In this case, the correction data will be updated with a large weight being given to the correction data.

Such update, however, is inappropriate for the parts that include many signal components corresponding to the magnetic pattern.

This will be explained by use of FIG. 2 . For example, the region MG 3 having a magnetic pattern is positioned on the sensing line La in the paper currency 13 . However, the proportion of the sensing line La occupied by the region MG 3 is small.

Therefore, when pattern presence/absence is evaluated by the method of the first embodiment, a decision that a magnetic pattern is present (greater than TH 2 ) is not made. As a result, suitable processing is not carried out for the region MG 3 , which has a magnetic pattern. The region MG 3 is a part in which many signal components are included.

In the third embodiment, in contrast, the presence or absence of pattern components is decided for each sensing element 11 . Accordingly, if the degree of inclusion of signal components corresponding to the magnetic pattern is large, the sensing elements 11 in the part corresponding to the region MG 3 will be processed on the assumption that the magnetic pattern is present. The sending elements 11 other than those in the part corresponding to the region MG 3 will be processed on the assumption that the magnetic present is absent.

Accordingly, in the place (the region MG 3 ) where the magnetic pattern is present, the updated correction data d 42 ( i ) will have a value close to the correction data d 44 ( i ) determined on the basis of the sensing data d 3 ( i ) in previous sensing lines La.

In places (regions other than the region MG 3 ) where the magnetic pattern is absent, the updated correction data d 42 ( i ) will have a value close to the sensing data d 3 ( i ).

The direct current component corrected sensing data d 4 ( i ) is obtained by subtracting the updated correction data d 42 ( i ) derived in this way from the sensing data d 3 ( i ).

From the above, in places (the region MG 3 ) where a magnetic pattern is present, the information reading apparatus 110 according to the third embodiment does not lose the signal components corresponding to the magnetic pattern. In places (regions other than the region MG 3 ) where the magnetic pattern is absent, the information reading apparatus 110 obtains the corrected sensing data d 4 ( i ) that are aligned at the level of the target value TGT.

In the second and third embodiments, parts other than those described above may be configured in the same way as in the first embodiment.

In the invalid intervals of the magnetic pattern sensing, the first to third embodiments determines the weighting coefficients K 3 , K 44 based on the a signal from the outside, i.e., the sensing enable signal d 2 . The weighting coefficients K 3 , K 44 may however be determined based on the correction data d 44 ( i ) and the sensing data d 3 ( i ) also in the invalid intervals of the magnetic pattern sensing.

In the first embodiment, the paper currency 13 was described as being moved by the transport belt 14 . Instead of the transport belt 14 , however, the line sensor 12 may be moved. That is, it suffices to produce relative motion between the object being sensed and the line sensor 12 , and provide means for producing the relative motion.

This also applies to the second and third embodiments.

Each of the foregoing embodiments was described for the case in which the sensed objects were paper currency. The present invention is also applicable, however, to cases in which the sensed objects are paper notes other than paper currency. The paper notes are not limited to paper material as long as they are similar in shape to paper notes.

The present invention has also been described for the case in which it is used in apparatus for sensing a magnetic pattern. The present invention is also applicable, however, to apparatus for sensing image patterns other than magnetic patterns. The image patterns are not limited to image patterns as long as they are patterns similar to image patterns. For example, the present invention is also applicable to an image sensing apparatus having a plurality of light receiving elements.

The invention has been described above as an information reading apparatus, but the signal processing apparatus constituting part of the information reading apparatus and the signal processing methods executed by the signal processing apparatus also form part of the invention.

Embodiments of the invention have been described as above, but the invention is not limited to these embodiments.

1 pattern sensing unit, 2 sensing enable generation unit, 3 A/D conversion unit, 4 , 4 b direct current component correction unit, 5 signal conditioning unit, 6 system control unit, 11 magnetic sensor, 12 line sensor, 13 paper currency, 21 paper currency detector, 41 , 41 b pattern presence/absence evaluation value generation unit, 42 , 42 b correction data updating unit, 43 , 43 b correction data subtraction unit, 44 correction data line memory, 421 , 421 b weighting coefficient generation unit, 422 , 422 b correction data computation unit, 423 parameter setting unit.