Method for generating a digital image of at least one section of a value document

Abstract

A method for generating a digital image of at least a portion of a value document, comprises transporting the portion of the value document in a specified transport direction at a specified speed, capturing pixel regions, and generating capture pixel data allocated to integration intervals. The capture pixel data of at least one of the sequences of capture pixel data are corrected. This reduces the influences of the movement of the value document in the transport direction during the capture of capture pixel data.

Claims

1. A method for generating a digital image of at least a portion of a value document, the method comprising: transporting the portion of the value document in a specified transport direction at a specified speed through capture pixel regions, each of the capture pixel regions being arranged along a line extending transverse to the transport direction, wherein each of the capture pixel regions extends in the transport direction, capturing, during said transport, for each of said capture pixel regions arranged along the line transverse to the transport direction, during integration intervals of at least one specified sequence of integration intervals, at least one property of optical radiation coming from the respective capture pixel region, generating a time sequence of capture pixel data respectively allocated to the integration intervals, said capture pixel data reproducing the property of the optical radiation from the respective capture pixel region, integrated over the respective integration interval and the respective capture pixel region, and employing the sequences of capture pixel data as pixel data for spatial sequences of pixels along the transport direction for the digital image of the portion of the value document, wherein capture pixel data of the at least one specified sequence of capture pixel data are corrected to reduce an effect on the capture pixel data specifically due to said specified speed of said transporting of the value document in the transport direction during the capture of the capture pixel data, wherein, when correcting capture pixel data for a respective capture region and allocated to an integration interval, capture pixel data of the same specified sequence for the same respective capture region that are allocated to different integration intervals are employed independently from capture pixel data of other sequences, wherein said correcting of said capture pixel data includes correcting said capture pixel data based at least in part on at least one factor depending on two or more of: (i) the specified speed of said transporting of the value document in the transport direction, (ii) a length of the integration interval, and (iii) an extension of the capture pixel region in the transport direction, and wherein said correcting includes subtracting from a first capture pixel, a product of one of said at least one factor and a second capture pixel data of an integration interval which immediately precedes or immediately follows said first capture pixel.

2. The method according to claim 1, wherein, said correcting comprises a first capture pixel data of a sequence being corrected before all capture pixel data of the sequence are available.

3. The method according to claim 1, wherein said correcting comprises a correction of the capture pixel data of the at least one sequence, and a first correction comprising first changing the capture pixel data for at least one of the integration intervals of the sequence of the integration intervals in dependence on the capture pixel data of the same sequence for one of the integration intervals that precedes, preferably immediately precedes, the integration interval or an integration interval that follows, preferably immediately follows, the integration interval, to reduce an integrated property represented by the capture pixel data.

4. The method according to claim 3, wherein the changing of the capture pixel data upon the first correction additionally takes place in dependence on the transport speed and/or the length of the integration interval and/or the extension of the capture pixel region in the transport direction.

5. The method according to claim 3, further comprising a second correction comprising changing the capture pixel data for the at least one sequence and the at least one integration interval which were corrected upon the first correction, in dependence on the capture pixel data and/or the capture pixel data of the same sequence which were corrected upon the first correction for one of the integration intervals that follows, preferably immediately follows, the at least one integration interval, or one of the integration intervals that precedes, preferably immediately precedes, the integration interval, to reduce a second integrated property represented by the capture pixel data.

6. The method according to claim 5, wherein the changing takes place additionally in dependence on the transport speed and/or the length of the integration interval and/or the extension of the capture pixel region in the transport direction.

7. A method for generating a digital image of at least a portion of a value document, the method comprising: transporting the portion of the value document in a specified transport direction at a specified speed through capture pixel regions, each of the capture pixel regions extending in the transport direction and the capture pixel regions being arranged transverse to the transport direction, capturing, during said transport for the capture pixel regions respectively during integration intervals of at least one specified sequence of integration intervals, at least one property of optical radiation coming from the respective capture pixel region, and generating a time sequence of capture pixel data respectively allocated to the integration intervals, said data reproducing the property of the optical radiation from the respective capture pixel region, integrated over the respective integration interval and the respective capture pixel region, and employing the sequences of capture pixel data as pixel data for local sequences of pixels along the transport direction for the digital image of the portion of the value document, wherein capture pixel data of at least one of the sequences of capture pixel data are corrected, while employing the capture pixel data of the same sequence that are allocated to respectively different integration intervals, to reduce an influence of a movement of the value document in the transport direction during the capture of the capture pixel data on the capture pixel data, wherein said correcting of the capture pixel data of the at least one sequence comprises a first correction comprising first changing the capture pixel data for at least one of the integration intervals of the sequence of the integration intervals in dependence on the capture pixel data of the same sequence for one of the integration intervals that precedes, preferably immediately precedes, the integration interval or an integration interval that follows, preferably immediately follows, the integration interval, to reduce an integrated property represented by the capture pixel data, and said correcting further comprises a second correction comprising changing the capture pixel data for the at least one sequence and the at least one integration interval which were corrected upon the first correction, in dependence on the capture pixel data and/or the capture pixel data of the same sequence which were corrected upon the first correction for one of the integration intervals that follows, preferably immediately follows, the at least one integration interval, or one of the integration intervals that precedes, preferably immediately precedes, the integration interval, to reduce a second integrated property represented by the capture pixel data, wherein, after the first and the second correction further correcting the capture pixel data corrected upon the second correction with a factor depending on the transport speed, the length of the integration interval and the extension of the capture pixel region in the transport direction.

8. The method according to claim 1, wherein, said correcting the capture pixel data, comprises filtering at least one, preferably contiguous, part of the sequence of the capture pixel data.

9. The method according to claim 8, wherein the filtering comprises folding the part of the sequence with one, preferably symmetric, filter function.

10. The method according to claim 9, wherein the filter function drops as the distance increases.

11. The method according to claim 9, wherein the filter function changes sign at least once as the distance increases.

12. An apparatus for determining a digital image of at least a portion of a value document from sequences of capture pixel data, obtained by transporting the portion of the value document in a specified transport direction at a specified speed through capture pixel regions extending respectively in the transport direction, and capturing during said transport for the capture pixel regions respectively during integration intervals of a specified sequence of integration intervals at least one property of optical radiation coming from the respective capture pixel region, and generating a time sequence of capture pixel data respectively allocated to the capture time intervals, which reproduces the property of the optical radiation from the respective capture pixel region, integrated over the respective integration interval and the respective capture pixel region, wherein the apparatus comprises a correction device which is configured so as to execute the method according to claim 1.

13. An apparatus for generating a digital image of at least a portion of a value document, having an image capturing device on which the portion of the value document is transportable in a specified transport direction at a specified speed through capture pixel regions extending in the transport direction, said device configured to capture during said transport, for the capture pixel regions respectively during integration intervals of a specified sequence of integration intervals, at least one property of optical radiation coming from the respective capture pixel region, and configured to generate a time sequence of capture pixel data respectively allocated to the capture time intervals, which reproduce the property of the optical radiation from the respective capture pixel region, integrated over the respective integration interval and the respective capture pixel region, and comprising the apparatus for determining a digital image according to claim 12.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The invention will hereinafter be explained further by way of example with reference to the drawings. The figures are described as follows:

(2) FIG. 1 a schematic view of a value-document processing apparatus,

(3) FIG. 2 a schematic view of a part of a sensor device of the apparatus in FIG. 1 in the transport direction,

(4) FIG. 3 diagrams illustrating the movement of value-document pixel regions through a capture pixel region of the sensor device in FIG. 1,

(5) FIG. 4 diagrams which illustrate the overlap F of value-document pixel regions and the capture pixel region in FIG. 3 as a function of the time t and the integration function of a capture element of the sensor device in FIG. 3 as a function of the time t,

(6) FIG. 5 a roughly schematic flowchart for a first method for generating a digital image, and

(7) FIG. 6 a roughly schematic flowchart of a second method for generating a digital image.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

(8) A value-document processing apparatus in FIG. 1, in the example a banknote processing apparatus for sorting banknotes in dependence on their authenticity and/or state, has an input region 10 for receiving stacked value documents 12, in the example banknotes, a singler 14 for singling value documents 12 from the input region 10, a sensor device 16 for capturing at least one specified property of the singled value documents 12 in a capture region 17 of the sensor device 16, in the example of an image, and output pockets 18 for receiving value documents 12 checked by means of the sensor device and a transport device 20 for transporting the value documents from the singler 14 in a transport direction R through the capture region 17 to the output pockets 18. The value-document processing apparatus further comprises a control device 22, which is connected via signal connections with the sensor device 16 and the transport device 20 and controls the transport device 20 in dependence on signals of the sensor device 16. The sensor device 16 captures the image and determines, in dependence on the captured image, whether the value document fulfills at least one specified criterion for the authenticity and/or the state of the value document, and generates a signal that represents the result of the check. The control device 22 captures the signal and, in dependence on this signal, drives the transport device 20, in the example a gate 24, such that the value document captured by the sensor device 16 is transported in accordance with the result of the check to a specified one of the output pockets 18. In other exemplary embodiments at least one further sensor device can be provided, which captures properties of a singled value document transported by the transport device 20, checks a specified criterion for the respective captured property and emits a corresponding signal representing the result of the check to the control device 22. The control device 22 can then drive the transport device 20 additionally in dependence on this signal.

(9) The sensor device 16 represented in FIG. 2 in greater detail, but only schematically, comprises in a sensor housing 26 among other things an illumination device 28 for illuminating the capture region 17 with optical radiation in at least one specified wavelength range, in the example with white light and IR radiation, a line camera 30 for capturing optical radiation coming from the capture region in the specified wavelength rages and an imaging optic 32 for imaging the capture region onto the line camera 30, i.e. for focusing the optical radiation coming from the capture region 17 on the line camera 30. The housing 26 has a window through which the radiation coming from the capture region can get to the imaging optic 32. Further, the sensor device 16 has an evaluation device 34 that serves, among other things, as a correction device within the meaning of the invention, receiving the signals of the line camera 30 and evaluating them. In particular, it is configured, together with the line camera 30, to carry out the method described hereinafter for generating a digital image of at least one portion of a value document, in the example of a portion over the complete extension of the value document in the transport direction R.

(10) The value documents are transported in a transport plane, which in FIG. 1 is orthogonal to the drawing plane parallel to the transport direction R, in which the value documents are transported through the capture region 17. The line camera 30 has at least one line of capture elements 36, in the following referred to as sensor elements, in the example four lines of sensor elements. In front of each line a filter is disposed, by means of which the radiation impinging on the line from the capture region 17 is filtered. In the example filters are provided for red, green and blue light and infrared radiation. Since the lines are configured uniformly apart from the configuration of the not shown filters, in the following only one line and the processing of the signals of the sensor elements of the line will be further described. Each of the sensor elements is connected with the evaluation device 34 via a signal connection via which the evaluation device 34 captures the signals of the sensor elements, or reads them out.

(11) As shown schematically in FIGS. 1 and 2, the equally configured sensor elements 36 of a line are disposed along a straight line extending transversely to the transport direction R and parallel to the transport plane. Through this arrangement such capture pixel regions 38 of the capture region are imaged onto the sensor elements 36 which lie correspondingly parallel transversely to the transport direction R, so that one capture pixel region 38 is imaged respectively exactly onto one of the sensor elements 36. Therefore one sensor element is allocated to each capture pixel region and vice versa.

(12) Since the sensor elements work uniformly and the processing of the signals of the sensor elements takes place in the same fashion, in the following the operation mode and the signal processing need to be described only for one sensor element.

(13) During the transport of the value document past the sensor device 16 the intensities of the optical radiation impinging on the respective sensor elements are captured by the sensor elements within specified, immediately consecutive capture time intervals T, which are equally long in the present example, and are transformed into sensor signals, respectively (cf. FIG. 4).

(14) More exactly each of the sensor elements, over specified integration intervals within the capture time intervals, captures the complete optical radiation reaching the respective sensor element and forms therefrom a sensor signal for the respective integration interval. Therein the intensity of the optical radiation impinging on the sensor element, i.e. of the optical radiation that comes from the allocated capture pixel region and is filtered by one of the mentioned filters, is captured in a locally integrating fashion over the capture pixel region and in a temporally integrating fashion over the respective integration interval, so that the sensor signal reproduces the corresponding radiation energy. The integration intervals in the present exemplary embodiment are equally long and lie at the same place within the corresponding capture time intervals, for example at the end.

(15) Frequently the integration intervals are somewhat shorter than the capture time intervals so as to make a reading out of the sensor elements possible.

(16) By capturing the property of the optical radiation in the specified capture time intervals or integration intervals within the capture time intervals, there arises a time sequence of sensor signals which are allocated to the respective sensor element and thus the corresponding capture pixel region. The evaluation device 34 forms from this sequence of sensor signals, provided that the sensor signals are not already digitized by digitizing them, a time sequence of capture pixel data which represent the property value of the captured optical radiation respectively represented by the sensor signals, in the example of the radiation energy.

(17) Through the simultaneous transport of the value document past the sensor device 16 there corresponds a local sequence of capture pixel data to this time sequence of capture pixel data for one sensor element or to the capture pixel region allocated thereto, said local sequence of capture pixel data corresponding at least approximately to pixel data of a local sequence of pixels on the value document in the transport direction R. Together with the local resolution transversal to the transport direction R, by employing the line of sensor elements, thus an image of the value document can be obtained that is formed by the evaluation device 34.

(18) As illustrated in FIGS. 3 and 4, however, at high transport speeds the conformity decreases between a value of the property of the captured optical radiation represented by capture pixel data and the corresponding value in a stationary value document. This is due to the circumstance that during the integration interval there is captured a region of the value document which is larger, more exactly longer in the transport direction, than the corresponding extension of the capture pixel values region. Partial regions within the region, however, are located within the capture pixel region for different durations.

(19) This is illustrated in FIGS. 3 and 4. In FIG. 4 there are shown, merely for the sake of easier understanding, four value-document pixel regions P.sub.i2, P.sub.i1, P.sub.i and P.sub.i+1 that are disposed immediately subsequently in the transport direction on a value document that is transported past the sensor device and to be captured by said sensor device, at four consecutive times t.sub.1, t.sub.2, t.sub.3 and t.sub.4. i herein designates a natural number. The value-document pixel regions are transported through the capture pixel region 38 illustrated by a dotted area. At the time t.sub.1 the value-document pixel region P, is not yet in the capture pixel region 38 and at the time t.sub.2 a little less than half. At the time t.sub.3 the two pixel regions are congruent, afterwards the value-document pixel region P.sub.i travels out of the capture pixel region 38 again, wherein the next value-document pixel region P.sub.i+1 enters the capture pixel region 38. In the upper diagram in FIG. 4 this is again illustrated. Therein the overlap F for the value document pixels P.sub.i1, P.sub.i and P.sub.i+1 with the capture pixel 38 is represented as a function of the time t; a complete overlap corresponds to the value 1. This value of overlap indicates how much the respective value-document pixel region or the radiation coming from there contributes to the value at the time t captured by the sensor element allocated to the capture pixel region. The sensor element integrates the contributions respectively over the integration intervals .sub.i. This is shown in the lower diagram in FIG. 4, in which the quantity I as a function of the time t indicates whether the sensor element integrates, value 1, or not, value 0. The value determined by the sensor element for the allocated capture pixel region is proportional to the areas under the curves in the F-t diagram during the integration interval. The corresponding areas are illustrated by hatchings, wherein in the figure the contributions of the value-document pixel regions P.sub.i1 and P.sub.i+1 partly cover those of the pixel region P.sub.i.

(20) The time period between two consecutive integration intervals is employed for reading out the sensor elements.

(21) The lengths of the capture time intervals and the integration intervals in the capture time intervals are chosen in dependence on the transport speed and the extension of the capture pixel regions in the transport direction R.

(22) The signals of the sensor element represent a time sequence of capture pixel data D.sub.i, for the integration intervals .sub.i, wherein the index i simultaneously represents the order of capture and corresponds to the index of the respective value-document pixel region, which had the maximum overlap with the capture pixel region during the respective integration interval.

(23) The sequence of the capture pixel data D.sub.i thus represents an approximate image of a track on the value document, which is formed by the captured value-document pixel regions P.sub.i. These capture pixel data D.sub.i can therefore be employed as pixel values of the image.

(24) The data D.sub.i are formed from the sensor signals and further processed in the evaluation device 34. The evaluation device 34 for this purpose has a data processing device with a processor and a memory, in which there is stored a program for carrying out the method described in the following and illustrated in FIG. 5, when the program is executed by means of the processor. Insofar the evaluation device 34 represents a correction device within the meaning of the invention.

(25) Again only the method steps for one capture pixel region are described, those for the other capture pixel regions and colors are correspondingly interlaced or processed successively.

(26) Initially (cf. FIG. 5) capture pixel data are captured in a first portion and subjected to a first correction. This portion comprises in FIG. 5 the partial steps S10 and S12, which are carried out in a mutually interlaced fashion in the present exemplary embodiment.

(27) In a loop over a sequence of capture pixel data initially a value D.sub.i is captured for an integration interval in the location i of the sequence, wherein the value i=1 is the first value of the sequence for the respective value document and i is incremented by 1 for each loop run. The maximum number of values of the sequence is N.

(28) The following partial step S12 is executed only for values i>1. In this step there is subtracted from the value D.sub.i the product of a factor V and the value D.sub.i1 for the immediately preceding integration interval, and the result is stored a new, now corrected value D.sub.i. Therein V is specified in dependence on the transport speed, the length of the capture time interval, the length of the integration interval and the extension of the capture pixel region parallel to the transport direction R and is a positive value smaller than 1. If the lengths of the intervals vary in different exemplary embodiments, the factor can also be varied correspondingly. It is then specified adapted for the lengths of the integration intervals i.

(29) If it is detected in step S10 that the value document has left the capture pixel region, the loop ends. Thereby such influences are compensated at least approximately which are due to the circumstance that during the capture of a value-document pixel region the preceding value-document pixel region partly overlaps with the capture pixel region.

(30) In a subsequent loop over the values i a second correction is carried out. For this purpose the process is started with i=N and i is decremented by 1 upon every loop run. In step S14, which is not carried out for the last element of the sequence, i.e. for i=N, there is subtracted from each corrected value D.sub.i the product of a factor H and the value D.sub.i+1 for the immediately subsequent integration interval and the result is stored as a new, now twice corrected value Di. Therein H is specified in dependence on the transport speed, the length of the capture time interval, the length of the integration interval and the extension of the capture pixel region parallel to the transport direction, and is a positive value smaller than 1. If the lengths of the time duration intervals vary in different exemplary embodiments, the factor can also be varied correspondingly.

(31) By the second correction such influences are corrected at least approximately which are due to the circumstance that already during the capture of a value-document pixel region the subsequent value-document pixel region partly overlaps with the capture pixel region.

(32) V and H do not need to be equal, but in the present example V=H is chosen.

(33) In the following step S16 the twice corrected capture pixel data D.sub.i are scaled by a factor N, which can be specified in dependence on the factors V and H and thus in dependence on the transport speed, the length of the capture time interval, the length of the integration interval and the extension of the capture pixel region parallel to the transport direction. The value of N is preferably larger than 1. Thereby such influences of the two corrections on the capture pixel data can be compensated that result in a reduction of the amounts of the capture pixel data.

(34) The data D.sub.i can now be employed in the evaluation device as pixel data of an image of the track on the value document and further processed using known methods.

(35) A second exemplary embodiment illustrated in FIG. 6 differs from the first exemplary embodiment through the correction of the capture pixel data and the configuration of the evaluation device. The method steps of the correction are now executed in an FPGA forming part of a correction device within the meaning of the invention, said FPGA passing the results on to a processor of the evaluation device. In the following only the processing for one track is described, like in the first exemplary embodiment, and it is assumed that capture pixel data are captured for one track of the value document and one wavelength channel N.

(36) In the method, like in the first embodiment, capture pixel data D.sub.i are captured in steps S20 corresponding to the steps S10.

(37) The captured capture pixel data of the sequence are processed as follows:

(38) The sequence of the capture pixel data D.sub.i is subjected to a filtering, for which vectors F.sub.i are employed, which, depending on the value, have i to 2*n+1 components, with n herein being a specified natural number larger than 0 that is chosen in dependence on the desired exactness of the correction.

(39) If there is valid 2*n<i<Nn1, a multiplication of a (2*n+1)-dimensional vector F.sub.i having the components F.sub.ij (j=1, . . . , 2*n+1) is carried out with a vector d.sub.i formed of the capture pixel data D.sub.in, . . . , D.sub.i, D.sub.i+n, whose result is employed as a corrected value D.sub.i for D.sub.i:
D.sub.i=F.sub.i*d.sub.i

(40) The values of the components of F.sub.i are chosen in dependence on the transport speed, the extension of the capture pixel region in the transport direction, the duration of the capture time interval and the duration of the integration interval and is equal for a given j for n<i<Nn.

(41) For capture pixel data D.sub.i with i<n+1 the vector d.sub.i comprises only the components D.sub.i, . . . , D.sub.i, . . . D.sub.i+n; then there is employed for i<n+1 instead of a vector F.sub.i for n<i<Nn only a vector F.sub.i with the last n+i components of the vector F.sub.i:
D.sub.i=F.sub.id.sub.i

(42) If N is the total number of the capture pixel data, the vector d.sub.i comprises for the capture pixel data D.sub.i with i>N(n+1) only the Ni+n components D.sub.in, . . . , D.sub.i, . . . D.sub.N; instead of the vector F.sub.i for n<i<Nn in the multiplication there is then employed only a vector F.sub.i with the leading 2*n+1i components of the vector F.sub.i:
D.sub.i=F.sub.i*d.sub.i

(43) This procedure corresponds to a multiplication of a vector with the components D.sub.i, i=1, . . . , N with a band-shaped filter matrix f with elements f.sub.ij in the line i in the column j (i, j=1, . . . , N).

(44) The matrix elements f.sub.ij are then given as follows:
For i<n+1
f.sub.ij=F.sub.ij for j=1, i+n, 0 else,
for n<i<Nn
f.sub.ij=F.sub.ij for j=in, . . . , i+n, 0 else, and
for Nn1<i<N+1
f.sub.ij=F.sub.ij for j=in, . . . , N, 0 else.

(45) This matrix is symmetrical in the present exemplary embodiment.

(46) The vectors F.sub.i for n<i<Nn are thus also symmetrical with regard to the middle element. Further, the components in this exemplary embodiment continuously change their sign proceeding from the middle element. The vector F.sub.i can for example be given for n<i<Nn as follows:
(210.sup.4; 10.sup.3; 710.sup.3; 0.04; 0.2; 1.4; 0.2; 0.04; 710.sup.3; 10.sup.3; 210.sup.4)

(47) The vector F.sub.i can then comprise for i<n+1 the i+n right components of the vector F.sub.n+1, and the vector F.sub.i for i>Nn1 can then comprise the i+n left components of the vector F.sub.n+1.

(48) The matrix f can be approximately obtained by assigning a target capture pixel value D.sub.i.sup.0 to each of the value-document pixel regions i. Then the capture pixel values D.sub.i to be expected can be computed as a function of the target capture pixel values D.sub.i.sup.0, wherein the initially described model and specified values for the transport speed, the extension of the capture pixel region in the transport direction, the duration of the capture time intervals and the duration of the integration intervals are employed. If d designates the vector of the capture pixel values D.sub.i to be expected, and d.sup.0 the vector of the target capture pixel values D.sub.i.sup.0, there results an interrelation in the form
d=g*d.sup.0

(49) g therein is a band matrix which represents a transfer function. By approximate or preferably exact inversion of g there results the filter matrix f. Its entries can partly be set to zero in dependence on an acceptable error.

(50) A further exemplary embodiment differs from the second exemplary embodiment in that no FPGA is used and the capture pixel data are stored first in the correction device, i.e. the evaluation device. When all capture pixel data of a sequence are present, the filtering or multiplication with the filter matrix is carried out.