DE10248621A1 - Method and system for processing banknotes - Google Patents

Abstract

The invention relates to a method and a corresponding system for sorting banknotes, in which a sorting class for the banknote is derived from properties of a banknote using a sorting algorithm. DOLLAR A For simplified and more reliable specification of the sorting algorithm and its execution, in particular on different types of banknote processing machines, it is provided that the sorting algorithm is specified using one or more linked decision nodes, each of which has one or more decision node outputs and when deriving the sort class in such a way that a transition from one decision node to another linked node takes place via a decision node output of the decision node, the decision node output depending on the properties of the banknote and / or on predetermined parameters.

Description

The invention relates to a method and a system for processing banknotes according to the preamble of claims 1 and 18th

In the case of the German published application DE 196 46 454 A1 In known methods, a sorting class for sheet material, in particular bank notes, is derived from the measurement results obtained for the sheet material using a sorting algorithm which is specified in the form of a sorting tree. For this purpose, sorting nodes are arranged in different levels of the sorting tree, in which value ranges are defined for one or more measurement results. The value range of a measurement result in a sorting node of a level is contained in the value range of the corresponding measurement result in an assigned sorting node of the level above.

When deriving the sorting class from the available measurement results the sorting node is determined in the deepest level in the sorting tree, in which all measurement results the banknote lies in the corresponding defined value ranges. The Sort class that corresponds to this sort node will eventually become the Sheet material assigned. The sorting node is determined by Recursion, starting from a sorting node in the highest level all in the next lower-level sorting nodes on the existence of a sorting node are searched for, in which all measurement results of the sheet material within the corresponding value ranges of the measurement results of the sorting node lie. If this is the case, those are assigned to this node Sort node in the next lower level checked in the same way, etc. In this way the sorting node is determined in the lowest level, where all measurement results of the sheet material in the corresponding value ranges of the measurement results of this sorting node.

Because of its hierarchical structure with several levels, such a sorting tree gives the sorting algorithm in one for the operator's form, which is generally not easy to understand and difficult to grasp intuitively again. For the specification of complex sorting algorithms is therefore usually a higher level of experience and expert knowledge required. There is also the danger an incorrect specification of value ranges in the numerous, i.a. for every Denomination of banknotes to be sorted separately Sorting node, which can lead to incorrect sorting results. In addition, one for one certain type of banknote processing machines given sorting tree mostly be modified if this is transferred to another type of banknote processing machines shall be. This also requires a high level of expertise and expertise an undesirable Error source.

It is an object of the invention Procedure and a corresponding system for carrying out the Procedure to specify which is a simplified and more reliable Specification and execution the sorting algorithm, especially for different types of Banknote processing machines, allowed.

This task is accomplished through the procedure and the system with the features of independent claims 1 and 18 solved. Advantageous further developments are the subject of the dependent claims.

The method according to claim 1 stands out characterized in that the Sorting algorithm using one or more decision nodes is specified, which each have one or more decision node outputs and via the or the respective decision node exits with one or more further nodes, in particular further decision nodes, are linked. The decision nodes are used when deriving the sort class go through in such a way that a transition from one decision node to another linked to it Node, especially other decision nodes, via a Decision node exit of the decision node takes place. The Selection of the respective decision node exit, via which the transition from the decision node to another node depends on the properties of the banknote and / or predetermined parameters from.

Obtain the specified parameters e.g. on the configuration or machine type of the corresponding Banknote processing machine, can but freely specified by an operator and different Reflect information such as the identity of the operator, the time or other parameters.

By using decision nodes let yourself the sorting algorithm, preferably as a sorting program code or Input graphics, in easy to understand, intuitively graspable and clearer Specify shape, which makes specifying the sorting algorithm easier and becomes more reliable.

In the system according to the invention according to claim 18, the input device, via which the sorting algorithm is specified, is designed such that the sorting algorithm can be specified in the form of a sorting program, in particular a compilable one. In addition, there is at least one translation device, in particular a compiler, for translating the specified one Sorting program provided in an object program executable for the banknote processing machine.

This to carry out the method according to the invention particularly suitable system allows simple and reliable specification the sorting algorithm, preferably as a sorting program code or Input graphic. By translating the Sorting program codes or the input graphic in an object program, which is particularly of different types of banknote processing machines executable is, too the portability one for a banknote processing machine of a certain type Sorting algorithm on other types of banknote processing machines simplified.

The invention is described below of figures closer explained. Show it:

1 an example of a sorting algorithm specified in the form of a decision graph according to the invention;

2 a table with the definition of decision node functions of the in 1 decision node shown;

3 a table with the content of report tags for decision node outputs of the in 1 decision node shown; and

4 a structure of the system for sorting banknotes.

Example of a sorting algorithm

In the following based on the 1 to 3 An example of a sorting algorithm explained is intended to assign a sorting class from two possible sorting classes to each banknote to be processed. A first sorting class is assigned to so-called fit banknotes, which are suitable for further use, while a second sorting class is assigned to so-called Unfit banknotes, which are unsuitable for further use, for example due to dirt, stains or dog-ears. Banknotes that cannot be classified should be classified as reject banknotes, so-called reject banknotes.

The sorting process is supposed to be in two different ways Protocol types are logged, whereby in a first protocol type, the so-called reject report, information about banknotes is logged which are classified as reject banknotes. In one The second type of protocol, the so-called unfitness report, is intended to provide information about banknotes logged, which are classified as Unfit banknotes.

characteristics of banknotes

To derive the sort class The properties of the banknotes are used by sensors Banknote processing machine in which the banknotes are checked and theirs Sorting class are sorted accordingly.

In the example shown, both the denomination, i.e. the face value, as well as the orientation of the Banknote determined by two sensors, the properties Sensorl Denomination and Sensor2_Denomination or Sensor1_Orientation and Sensor2_Orientation can be obtained.

The DoubleFeed property reflects whether a banknote is properly inserted into the Banknote processing machine has been pulled in, and can accordingly only accept the values TRUE or FALSE.

The four transport properties TransportProp1 up to TransportProp4 take numerical values and contain information about the transport the banknote.

Out of three intended for the authenticity check Sensors are the three authenticity properties Authenticityl, Authenticity2 and Authenticity3, which are available in the presence of a corresponding Authenticity feature the value 1 and in the absence of the authenticity feature assume the value 0. Dependent of the denomination of the respective banknote must for a positive result the authentication none, at most one, at most two or all three of the authenticity properties assume the value 1.

The actual classification of the Banknote in a fit or unfit banknote depends on the condition properties Front Soil, Backsoil, Front Stain, Backstain and DogEar1 to DogEar4, each of which is a measure of the pollution or for Stains on the front and back of banknotes and for dog ears at the four corners of the banknote.

decision node

1 shows a sorting algorithm given in the form of a decision graph according to the invention, on the basis of which a sorting class "FIT", "UNFIT" or "REJECT" is derived from the above-mentioned properties of a banknote.

The decision graph has a diamond-shaped start node START, oval decision nodes with decision node functions FCT 1 to FCT_14 and rectangular end nodes TN (1) to TN (3), some of which are linked together.

When deriving the sorting class of a banknote, starting with the start node START and the subsequent first decision node with the decision node function FCT 1 , run through linked decision nodes until one of the end nodes TN (1) to TN (3) is reached. The class "REJECT", "FIT" or "UNFIT" assigned to the respective end node TN (1) to TN (3) then corresponds, preferably in conjunction with further properties of the banknote, such as denomination and / or orientation, to the sorting class of banknote.

Everyone in 1 The decision node shown has various decision node exits, via which a transition from one decision node to another associated decision node or end node can take place. The decision node exit, via which a transition occurs when the decision graph is run through, depends on the properties of the banknote to be processed.

Preferably the choice of decision node exit depends from the function value of the decision node function FCT_1 to FCT_14 of the respective decision node. Here is every decision node exit a comparison value, e.g. TRUE, FALSE or 1 to 4 assigned. A transition takes place over the decision node exit instead of its comparison value corresponds to the function value of the respective decision node function. For example, these are for the decision node with the decision node function FCT_10 the outputs labeled 1, 2, 3 or 4 for all cases in which the function value is the same 1, 2, 3 and 4, respectively. In cases in which the function value assumes a different value, there is a transition via the decision node exit labeled "Remainder" instead of.

Decision node functions

2 shows a table with definitions of the decision nodes in 1 assigned decision node functions FCT_1 to FCT_14. The decision node functions FCT_1 to FCT_14 each assume a function value which depends on the properties of the banknote and / or on predetermined parameters.

The decision node function FCT_1 of the first decision node to be traversed contains only the DoubleFeed property. The functional value of this decision node function can therefore either have the value TRUE or FALSE, which reflects a correct or incorrect entry of the banknote into the banknote processing machine. If the function value is TRUE, there is a transition (see 1 ) to the end node TN (1), to which the sorting class "REJECT" is assigned, ie the banknote is classified as a reject banknote to be rejected. If, on the other hand, the function value is FALSE, there is a transition to the next decision node with the decision node function FCT_2.

The decision node function FCT_2 contains four linked with AND Inequalities to the transport properties TransportPropl to TransportProp4 the banknote. Each of the inequalities checks whether a transport property, e.g. TransportPropl, within one for the respective transport property predetermined comparison value range, e.g. between TR_PROP1_LO and TR_PROP1_HI, lies. Only if all four transport properties TransportPropl until TransportProp4 satisfies the corresponding inequalities the function value of the decision node function FCT_2 the value TRUE and a transition takes place to the decision node with the decision node function FCT_7.

Is the function value of the decision node function FCT_2, however, FALSE, then the decision nodes with the Go through decision node functions FCT_3 through FCT_6 until finally the Terminal node TN (1), to which the sorting class "REJECT" is also assigned, is achieved. The decision node functions FCT_N with N = 3, 4, 5 and 6 assume the function value TRUE if a transport property TransportPropN less than or equal to the corresponding lower limit TR_PROPN_LO or greater or equal to the corresponding upper limit TR_PROPN_HI of the given one Comparative value range. Otherwise the respective function value is FALSE. In this way it is determined which of the four transport properties TransportPropl to Transport-Prop4 the reason for represents a classification of the banknote as a reject banknote.

The decision node function FCT_7 assumes the function values TRUE or FALSE if the determined denomination is identical or different for both denomination sensors. In the latter case, there is a transition (see 1 ) to the "REJECT" end node TN (1).

Takes the function value of the decision node function FCT_7, however, the value TRUE, so there is a transition to the next decision node with the decision node function FCT_8. There it is checked whether the Orientations of the banknote determined by both sensors are compatible are. The property Sensor1_Orientation and Sensor2_Orientation takes one of four possible Values 1, 2, 3 or 9, which correspond to the different orientations the banknote during correspond to their transport in the processing machine, namely "banknote front top "or" back of the banknote top "(" FaceUp "or BackUp") with two possible options Orientations. A compatibility is always given when the sum of the determined for both sensors Values is 5. If this is not the case, the decision node function takes FCT_8 the value FALSE and a transition to the "REJECT" end node TN (1) takes place.

Otherwise the decision node function FCT_8 takes the value TRUF and it takes place Transition to the decision node with the decision node function FCT_9. There, the sum of the authenticity properties Authenticity1, Authenticity2 and Authenticity3, which can each take the value 1 or 0, is compared with a comparison value tblAuthentic. This comparison value is specified as a one-dimensional variable field tblAuthentic [SEN1_DENO] with the field index SEN1_DENO. The field index SEN1_DENO corresponds to the value of the Sensor Denomination property determined by the banknote to be processed. Consequently, the comparison value tblAuthentic [SEN1_DENO] is a comparison value that depends on the denomination of the respective banknote.

Suppose "D1" is the name of a denomination, for which must have at least two authenticity features, then applies to tblAuthentic [D1] = 2. The function value of the decision node function FCT_9 in this case only takes the value TRUE if the sum of the Authenticity1, Authenticity2 and Authenticity3 authenticity properties bigger or is equal to the value 2. Otherwise the function value is FALSE, and there is a transition to the "REJECT" end node TN (1) instead.

The specification of comparison values or comparison value ranges in the form of variable fields Advantage that in a multitude of different comparison values at a decision node or comparative value ranges for a comparison can be made. This is itself for sorting banknotes with a large number of different denominations only a decision graph is required, which ensures a high degree of clarity and reliability guaranteed becomes.

The decision node function FCT_10 only contains the property Sensor1_Orientation and can therefore, as already was explained above in connection with the decision node function FCT_8, one of four possible Accept function values 1, 2, 3 or 4, which represent the different Orientations of the banknote correspond.

How out 1 it can be seen, the decision nodes with the decision node functions FCT_10 to FCT_13 were additionally assigned their own designations: "Sen1_Orient", "SoilNode", "StainNode" and "DogEar-Node".

In addition, the decision node outputs of decision nodes FCT_10 to FCT_13 are assigned output variables and output values. In the case of a transition via a decision node output, the corresponding output variable is assigned the corresponding output value. Some baseline values are 1 In the area of the decision node outputs, decision nodes FCT_10 to FCT_13 are shown as bracketed numerical values (0) or (1). The output variables of the individual decision nodes can be accessed by decision node functions in decision nodes that are subsequently run through. The names of the output variables of the individual decision nodes are made up of the name of the decision node, eg Sen1_Orient, and the extension exitvalue, eg Sen1_Orient.exitValue.

In again 2 As shown in the table shown, the decision node function FCT_11 compares the banknote properties for the contamination of the front and rear sidesoil or backsoil with corresponding predefined comparison values tb1FitFrontsoil or tb1FitBacksoil, which in this case are specified as two-dimensional variable fields with two field indices each. The first field index Sen1_Orient.exitValue corresponds to the above-described output variable of the decision node labeled "Sen1_Orient", while the second field index SEN1_DEMO corresponds to the property Sensor1_Denomination determined during the processing of the banknote.

The one used for the comparison The comparison value tb1FitFrontsoil or tb1FitBacksoil depends on the orientation and denomination of the banknote to be checked. Just in case both of them Banknote properties frontsoil and backsoil are smaller than that respective comparison values, the decision node function FCT_11 takes the value TRUE. In all other cases the function value is FALSE. The initial value (0) or (1) of this decision node is written to the SoilNode.exitValue output variable.

In the decision node function FCT_12, the properties FrontStain and Backstain are specified with Comparison values tblFitFrontstain and tblFitBackstain compared. The comments on Decision node functions FCT_11 apply analogously. The initial value (0) or (1) of this decision node is in the output variable StainNode.exitValue written.

In the decision node function FCT_13 becomes the sum of the for the four corners of the banknote determined properties DogEar1 to DogEar4 with a denomination-dependent comparison value tblFitDogEar [SEN1_DENO], which is given as a one-dimensional variable field. If the sum is smaller than this comparison value, the function value the decision node function is TRUE, otherwise it is the same FALSE. The initial value (0) or (1) of this decision node becomes written in the output variable DogEarNode.exitvalue.

In the decision node function FCT 14 finally the sum of the initial values of the decision nodes labeled "SoilNode", "StainNode" and "DogEarNode" is checked. Only if the sum of the starting values of the respective Output variables SoilNode.exitvalue, StainNode.exitValue and DogEarNode.exitValue result in the value 0, ie all three output values are equal (0), the function value of the decision node function FCT_14 takes the value TRUE and a transition to the one labeled "FIT" takes place Terminal node TN (2). Otherwise there is a transition to the end node TN (3) labeled "UNFIT".

Report tags

Certain decision node outputs of the in 1 The decision graphs shown are each assigned a report variable RT_1 to RT_13, a so-called report tag. In the case of a transition via a decision node exit, a corresponding entry is made in the associated report tag.

As explained above, is the derivation of the sort class in the example shown in two different types of protocol, namely a reject report and an unfitness report. These types of protocol can freely defined by compiling the desired report tags become. While The sorting process then becomes the content of the selected report tag written in a log of the corresponding log type.

3 shows a table with the content of report tags, the individual decision node outputs from in 1 represented decision nodes were assigned. Accordingly, the content of report tags RT_1 to RT_10 is logged in the log of the reject report, while the content of report tags RT_11 to RT_13 is logged in the log of the unfitness report.

The content of the report tags themselves can also be freely defined. So the content of the report Tags RT_1, RT_3 to RT_6 and RT_10 in the respective decision node tested Banknote properties DoubleFeed, TransportProp1 to TransportProp4 or Sensorl_Orientation selected. at the report tags RT_2, RT_7 to RT_9 and RT_11 to RT_13 the respective entry is freely defined according to the operator's wishes; in the example shown, these are the entries "TRANSPORT", "NON-MATCHING DENOMINATIONS "," NON-MATCHING ORIENTATIONS "," AUTHENTICITY "," SOIL "," STAIN "or" DOGEAR ", which one - also for one inexperienced operator - clear Information about a rejected or display as an unfit banknote.

end nodes

To further simplify the specification of the decision graph for individuals Property values can also be defined using systematic names. For the SEN1_DENO values of the Sensor1_Denomination property can do this e.g. the names D1, D2, D3 etc. are defined. The same applies to the SEN1_ORIENTATION values of the Sensor Orientation property, which z. B. can be designated with L1, L2, L3 and L4.

The values SEN1_ORIENTATION and SEN1_DENO a currently processed banknote are the end nodes TN (2) and Assigned to TN (3) and serve as indices of two-dimensional tables So-called end node identifiers tblTn2 and tblTn3, respectively a numeric value for represent the sorting result, which is sent to the processing machine is passed on. In this way it is possible to create a banknote with "FIT; SEN1_ORIENTATION : L3; SEN1_DENO: D47 " is to sort or stack differently in the processing machine as a banknote labeled "FIT; SEN1_ORIENTATION: L1; SEN1_DENO: D43 "is described.

Stapelmodi

The definition of a batch mode, ie which sorting classes of bank notes are to be output in which stackers of the bank note processing machine, can then be carried out in a simple manner. For example, should it be specified that all Unfit banknotes have an orientation 1 or 3 in a first stacker and all fit banknotes are issued in a second stacker, the corresponding instructions for the respective stacker can be as follows:
first stacker: SELECT "UNFIT" AND (SEN1_ORIENTATION: L1 OR SEN1_ORIENTATION: L3)
second stacker: SELECT "FIT".

sorting program

The in 1 Sorting algorithm shown as an example is shown in the form of a decision graph. Alternatively, the sorting algorithm is specified in the form of a corresponding sorting program, that is to say as a program code, in which the individual decision nodes are defined with their decision node functions and are run through program branches or jumps depending on the functional values of the decision node functions. The programming language for the sorting program is preferably an application-oriented, that is to say a compilable language that is specially tailored to the requirements for sorting banknotes.

Sorting system of banknotes

4 shows a structure of the system for sorting banknotes. The sorting algorithm is in an input device 40 entered. In the example shown here, the sorting program 43 directly with a keyboard 49 entered and from a display device 47 , especially a screen or display.

The sorting program 43 is preferably specified in the form of a compilable program code. During the compilation, a complete translation of the program code is first carried out before it is executed in a banknote processing machine 50 can be done. This enables the compiled program code to be executed quickly in real time.

The translation of the sorting program 43 takes place in a translation facility 44 , in particular a compiler, which converts the sort program into an object program 45 converts its program code from the banknote processing machine 50 can be executed. The program code of the object program 45 is done by an interpreter 51 in the banknote processing machine 50 executed in real time. As an interface between the interpreter 51 and the banknote processing machine 50 is an interpreter bowl 52 intended.

The program code of the object program 45 can also be used by banknote processing machines 50 ' other types of construction, for example with a different sensor and / or stacker configuration, if this is an interpreter 51 ' have, which with the interpreter 51 the banknote processing machine 50 is identical. Only the respective interpreter shell 52 ' as an interface between machine and interpreter 51 ' must be adjusted accordingly.

Alternatively or in addition to entering the sorting program directly 43 can the sorting program 43 also indirectly in the form of an input graphic 46 , which in its form in particular that of 1 corresponds to the decision graph shown. The input device 40 has a graphical user interface for this 48 on, for example a touchscreen display or graphic display which can be controlled by touch and / or by means of a keyboard. Via the input device 40 In particular, decision nodes can be inserted, decision node outputs and decision node functions can be defined, and links between individual nodes can be established. The same applies to end nodes, report tags and all others in connection with the example according to 1 to 3 explained structural units and sizes of the decision graph.

This graphical specification of the Sorting program, in particular using those described in more detail above Decision node and decision node functions, becomes a very clear and less prone to errors Specification of the desired one Sorting algorithm enables.

By a first conversion unit 41 becomes the input graphic 46 in the code of the sorting program 43 converted and can by the display device 47 displayed, checked and, if necessary, using the keyboard 47 to be changed. By a second conversion unit 42 can convert the sorting program 43 in the form of the input graphic, which is then on the input surface 48 shown and, if necessary - can be changed graphically there - as already explained in more detail above.

The input graphic 46 can also be used to represent the path through which a banknote to be sorted or already sorted is or has passed through the decision graph. This illustrates the sorting result for the banknote so that the reason for a particular classification, for example REJECT, FIT or UNFIT, can be tracked in detail. The input interface 48 is operated in a special debug mode.

Interpretermodi

The interpreter 51 respectively. 51 ' can in principle be operated in different interpreter modes.

In a first interpreter mode, the computing time required for deriving the sorting class of a banknote in real time is stored in the central processing unit of the banknote processing machine 50 respectively. 50 ' minimized by the fact that the function value of the corresponding decision node function FCT_1 to FCT_14 is determined only for the decision node currently to be run through. If at this point in time all the properties of the banknote required for this are not available - because, for example, the derivation of a required property from the measured values of a sensor has not yet been completed - the interpreter waits until the functional value of the decision node function has been determined until all the required properties are available. As a result, the computing time is minimized, so that computing capacity for other processes in the banknote processing machine is kept free. However, minimizing the computing time is accompanied by an increase in the overall evaluation time required for deriving the sorting class.

A second interpreter mode can therefore alternatively be selected in which the overall evaluation time required for deriving the sorting class of a banknote is minimized. In this mode, the function values of individual decision node functions FCT_1 to FCT_14 are calculated - regardless of the decision node currently to be run through - as soon as the properties of the banknote required by the respective decision node functions are fully available. Although this requires a higher computing time in the central processing unit of the banknote processing machine, the evaluation time half of the derivation of the sorting class of the banknote is reduced, however, since there are no waiting times at individual decision nodes during the passage through the decision nodes.

Claims (40)

Method for sorting banknotes, in which a sorting class for the banknote is derived from the properties of a banknote using a sorting algorithm, characterized in that the sorting algorithm is specified using one or more decision nodes, each of which has one or more decision node outputs, the decision nodes are linked via the respective decision node exit (s) to further nodes, in particular further decision nodes, and the decision nodes are passed through in deriving the sort class in such a way that a transition from one decision node to another node linked to it, in particular other decision nodes, via a decision node exit of the decision node, the decision node exit, via which the transition from the decision node to another node takes place, from the properties n the banknote and / or depends on predetermined parameters. A method according to claim 1, characterized in that at least a decision node has at least one decision node function (FCT_1 to FCT_14), whose function value depends on the properties the banknote and / or depends on the specified parameters. A method according to claim 2, characterized in that the Function value of the decision node function (FCT_1 to FCT_14) by comparing the properties of the banknote or the specified one Parameters with at least one comparison value or range of comparison values depends. A method according to claim 3, characterized in the at least one comparison value or comparison value range in Form of one or more variable fields is specified. Method according to one of claims 1 to 4, characterized in that that everyone Decision node exit of a decision node at least an output value or output value range is assigned and the Decision node exit at transition the decision node to another node from a comparison the output value or output value range with the properties the banknote or with the functional value of the corresponding decision node function (FCT_1 to FCT_14) depends. Method according to one of the preceding claims, characterized characterized that a or more transitions from individual decision nodes to associated other nodes are logged become. A method according to claim 6, characterized in that the Logging in one or more different log types done, for everyone Protocol type can be specified, which transitions from individual decision nodes to be logged to other nodes. Method according to one of the preceding claims, characterized in that the sorting algorithm in the form of a, in particular compilable, sorting program ( 43 ) is specified. Method according to one of claims 1 to 7, characterized in that the sorting algorithm in the form of an input graphic ( 46 ), especially a decision graph. A method according to claim 9, characterized in that the in the form of the input graphic ( 46 ) predefined sorting algorithm in a, in particular compilable, sorting program ( 43 ) is converted. A method according to claim 8, characterized in that the in the form of a sorting program ( 43 ) predefined sorting algorithm in an input graphic ( 46 ) is converted. Method according to one of claims 8 to 11, characterized in that the sorting program ( 43 ) in a for a banknote processing machine ( 50 . 50 ' ) executable object program ( 45 ), especially by means of compilation. Method according to claim 12, characterized in that the object program ( 45 ) in the banknote processing machine ( 50 . 50 ' ) is executed in real time. Method according to one of claims 12 to 13, characterized in that the object program ( 45 ) by an interpreter ( 51 . 51 ' ) in the banknote processing machine ( 50 . 50 ' ) is performed. A method according to claim 14, characterized in that the interpreter ( 51 . 51 ' ) through an interpreter bowl ( 52 . 52 ' ) is encapsulated. Method according to one of claims 14 to 15, characterized in that the interpreter ( 51 . 51 ' ) is operated in a first interpretation mode for minimal computing time. Method according to one of claims 14 to 15, characterized in that the interpreter ( 51 . 51 ' ) is operated in a second interpretation mode for minimum evaluation time to derive the sorting class. System for sorting banknotes with - at least one input device ( 40 ) for specifying at least one sorting algorithm, on the basis of which a sorting class for at least one banknote can be made from properties of the banknote, and - at least one banknote processing machine ( 50 . 50 ' ) for sorting the banknote according to the derived sorting class, characterized in that - the input device ( 40 ) is designed such that the sorting algorithm takes the form of a sorting program, in particular a compilable one ( 43 ) can be specified, and - at least one translation device ( 44 ), especially a compiler, for translating the sorting program ( 43 ) in a for the banknote processing machine ( 50 . 50 ' ) executable object program ( 45 ) is provided. System according to claim 18, characterized in that the input device ( 40 ) a graphical user interface ( 48 ), via which the sorting program ( 43 ) in the form of an input graphic ( 46 ), especially a decision graph, can be specified. System according to claim 19, characterized in that a first conversion unit ( 41 ) to convert the input graphic ( 46 ) in the sorting program ( 43 ) is provided. System according to one of claims 18 to 20, characterized in that a second conversion unit ( 42 ) to convert the sorting program into an input graphic ( 46 ), in particular a decision graph, is provided. System according to one of claims 18 to 21, characterized in that at least one interpreter ( 51 . 51 ' ) is provided, which the object program ( 45 ) on the banknote processing machine ( 50 . 50 ' ) executes. System according to claim 22, characterized in that the interpreter ( 51 . 51 ' ) an interpreter bowl ( 52 . 52 ' ) having. System according to one of claims 22 to 23, characterized in that the interpreter ( 51 . 51 ' ) can be operated in a first interpretation mode for minimal computing time. System according to one of claims 22 to 23, characterized in that the interpreter ( 51 . 51 ' ) can be operated in a second interpretation mode for minimum evaluation time to derive the sorting class. System according to one of claims 18 to 25, characterized in that - the sorting program code ( 43 ) or the input graphic ( 46 ) comprise one or more decision nodes, each of which has one or more decision node outputs, the decision nodes being linked via the respective decision node outputs with one or more further nodes, in particular further decision nodes, and - the decision nodes when deriving the sorting class in such a way can be run through that a transition from a decision node to another linked node, in particular other decision nodes, can take place via a decision node exit of the decision node, the decision node exit, via which the transition from the decision node to another node takes place, from the properties of the banknote and / or depends on predetermined parameters. System according to claim 26, characterized in that at least a decision node has at least one decision node function (FCT_1 to FCT_14) can be assigned, the function value of which Properties of the banknote and / or the specified parameters depends. System according to claim 27, characterized in that the Function value of the decision node function (FCT_1 to FCT_14) by comparing the properties of the banknote or the specified one Parameters with at least one comparison value or range of comparison values depends. System according to claim 28, characterized in that at least a comparison value or comparison value range in the form of a or several variable fields can be specified. System according to one of claims 26 or 29, characterized in that at least one output value or output value range can be assigned to each decision node output of a decision node and the decision node output when the decision node changes to another node by comparing the output value or output value range with the properties of the banknote and / or the specified parameters or with the function value of the corresponding decision node function (FCT_1 to FCT_14). System according to one of claims 18 to 30, characterized in that the existence or more transitions from individual decision nodes to associated other nodes are logged can be. System according to claim 31, characterized in that the Logging in one or more different log types can be done for Each type of protocol can determine which transitions from individual decision nodes to associated other nodes are logged should be. Method according to one of claims 1 to 17, characterized in that that the Parameter one or more properties of one for sorting banknotes trained banknote processing machine, in particular their Play back configuration or machine type. Method according to one of claims 1 to 17, characterized in that that the Parameters are specified by an operator. Method according to one of claims 9 to 17, characterized in that that the Path to a banknote to be processed or an already processed banknote covered or covered by the decision graph has, is represented. A method according to claim 35, characterized in that the Representation of the path of the banknote through the decision graph done in a debug mode. System according to one of claims 18 to 32, characterized in that that the Parameter one or more properties of one for sorting banknotes trained banknote processing machine, in particular their Play back configuration or machine type. System according to one of claims 18 to 32, characterized in that that the Parameters can be specified by an operator. System according to one of claims 19 to 32, characterized in that the input device ( 40 ), especially the graphical user interface ( 48 ) is designed to represent the path that a banknote to be processed or an already processed banknote has traveled or has traveled through the decision graph. System according to claim 39, characterized in that the input device ( 40 ), especially the graphical user interface ( 48 ), can be operated in a debug mode in which the path of the banknote is represented by the decision graph.