JPH0425981A - Data processor and data processing method - Google Patents

Abstract

PURPOSE:To make the processing route of input data the same irrespectively of a data value, and in addition, to simplify that processing route by selecting outputting one of plural codes according to the added result of an adding part. CONSTITUTION:A single control table 13 in which plural thresholds and plural code one input data ought to refer to are stored is provided beforehand. Then, the input data is compared successively with each of plural thresholds recorded in the above-mentioned control table 13 by comparing parts 14, 15, and '0' or '1' is outputted when the input data is smaller or larger than the threshold respectively, and the result is summed by the adding part 16. This numerical value obtained by summing shows how many pieces of the thresholds smaller than that input data there are, and has a different value for each of plural areas classified by plural thresholds. Then, by using this numerical value as an identifier, one of the code corresponding to above-mentioned plural areas can be selected and outputted.

Description

[Detailed Description of the Invention] [Summary] Regarding a data processing device and a data processing method that detect errors in input data, an object of the present invention is to make the processing path of input data the same regardless of the data value and to simplify the processing path. The purpose is to create a management table that stores multiple thresholds and multiple codes corresponding to multiple ranges divided by the multiple thresholds, and a management table that stores input data and the multiple thresholds in the management table. It has a plurality of comparison sections that compare each of the data and outputs data based on the comparison results, and an addition section that adds the data output from the plurality of comparison sections. The configuration is such that one of the multiple codes stored in the area in the management table is selected and output, or in the configuration,
Either a plurality of processing paths consisting of the comparison section and the addition section are provided in parallel, or a plurality of codes corresponding to a plurality of thresholds and a plurality of ranges divided by the plurality of thresholds are provided. The input data is compared with a plurality of threshold values stored in the above area, the comparison results are sequentially output, the above outputs are added, and the result of the above addition is stored in the above area. Configure to select and output one of multiple codes.

[Industrial application field]

The present invention relates to a data processing device and a data processing method that detect errors in input data.

In data processing in which a huge amount of input data is classified based on various criteria and output with an error code attached, similar processing is repeatedly performed on successive input data. Therefore, in order to improve the overall processing speed, it is necessary to simplify the processing path of each input data as much as possible.

[Conventional technology] When performing error detection processing to identify paper sheets such as banknotes and securities, first, the entire standard genuine note is divided into micro-regions and the shading of each micro-region is measured. do. Based on the results of such measurements performed on a large number of genuine notes, a permissible range of shading, that is, a lower threshold value and an upper threshold value, is determined for each minute region.

Next, the paper sheet to be inspected is divided into minute areas using the same method as described above, and the shading thereof is measured, and this measured value is compared with the lower and upper threshold values corresponding to the same minute areas. Then, a lower limit error code is output when the measured value is below the lower limit threshold, an upper limit error code is output when the measured value is above the upper limit threshold, and a normal code is output when the measured value is within the allowable range.

By repeating the above process for each micro-head region, it is possible to obtain a code map in which one of the above codes corresponds to all the micro-regions that make up the paper sheet to be inspected, and paper sheets can be identified using this code map. This can be done by Furthermore, in order to perform more precise discrimination, several intermediate thresholds are set between the lower and upper thresholds to increase the number of error codes.

FIG. 6 shows a block diagram for explaining a conventional method for performing such data processing. The lower limit error code, upper limit error code, and normal code are added to the input data based on the lower limit threshold value, upper limit threshold value, and the magnitude relationship with these threshold values determined for each minute region by the method described above. Lower threshold table 1
8, upper limit threshold table 19, lower limit error code table 20, upper limit error code table 21, and normal code table 22.

The above data processing will be explained below according to the flowchart shown in FIG.

First, first input data is taken into the input interface 11. The first few bits of input data contain information that specifies the minute area corresponding to the input data.
The last few bits contain data values. The first few bits are sent to the address generation unit 12, and the last few bits are sent to the matching unit 1.
Sent to 23rd. The address generation unit 12 has five address generation circuits corresponding to the five tables described above,
Based on the above information, an address where the threshold value and code to be referenced by the first input data are stored is generated for each table and sent to each table. Then, the data stored at the above address is read from each table and the following processing is performed. That is, the matching unit 123 matches the input data with the lower limit threshold value read from the lower limit threshold table 18.

If the input data is smaller than the lower limit threshold, the lower limit error code is read from the lower limit error code table 20 and outputted to the outside through the output interface 17. Then, the output interface 17 sends a processing completion signal to the input interface 11 to start receiving the next input data. If the human input data is greater than the lower threshold, the input data is sent to the matching unit I[24.

The collation unit 1124 collates the input data with the upper limit threshold value read from the upper limit threshold table 19 . If the input data is larger than the upper limit threshold, the upper limit error code is read from the upper limit error code table 21 and outputted to the outside through the output interface 17.

Then, the output interface 17 sends a processing completion signal to the input interface 11 to start receiving the next input data. If the input data is smaller than the upper limit threshold, a normal code is read from the normal code table 22 and outputted to the outside through the output interface 17.

Then, the output interface 17 sends a processing completion signal to the input interface 11 to start receiving the next input data. By repeating the above processing, it is possible to obtain a code map in which codes are attached to all minute regions.

[Problem to be solved by the invention]

However, as is clear from the above description, in the data processing described above, the data processing path differs depending on the value of input data. For example, if the input data is smaller than the lower limit threshold, the collation unit I23 performs the collation process with the lower limit threshold, immediately reads the lower limit error code from the lower limit error code table 20, and starts processing the next input data. However, if the input data is a normal value, the verification process in the verification unit 123 is followed by the verification process in the verification unit [24], and the processing to read the normal error code from the normal code table 22 is added, so it is less expensive than the former. This increases the time required for processing.

Therefore, when parallel processing is performed by inputting multiple data at the same time, the data input later may overtake the data input earlier, and the output data will not correspond to the input data, making it impossible to obtain a correct code map. A problem arises. In order to avoid such inconvenience, processing of the next plurality of input data must be started after processing of all the plurality of input data is completed. However, in this method, the overall processing speed is determined by the input data that has the slowest processing speed among the plurality of input data that are processed in parallel. In addition, in order to obtain a more precise error code map, it is necessary to further increase the number of thresholds, but in this case, the difference in processing speed depending on the value of human data becomes even larger, causing the above-mentioned problems. becomes noticeable.

Furthermore, in the above example, a table is provided for each threshold value and code for processing input data, so an address generation circuit is required for each table, which makes address management complicated and causes a decrease in processing speed. becomes.

Therefore, it is an object of the present invention to make the processing path of input data the same regardless of the data value and to simplify the processing path.

[Means to solve the problem]

The solution to the above problem is to create a management table that stores multiple thresholds and multiple codes corresponding to multiple ranges divided by the multiple thresholds, and a management table that stores input data and the multiple codes in the management table. It has a plurality of comparison sections that compare each of the threshold values and outputs data based on the comparison results, and an addition section that adds the data output from the plurality of comparison sections, and the addition result of the addition section a data processing device that selects and outputs one of a plurality of codes stored in the area of the management table, or a plurality of processing paths each consisting of the comparison section and the addition section; The above-mentioned data processing device is characterized in that it is provided in parallel, or the above-mentioned data processing device stores a plurality of thresholds and a plurality of codes corresponding to a plurality of ranges divided by the plurality of thresholds, and input data is processed into the above-mentioned area. Compare the code with multiple threshold values stored in the area, output the comparison results sequentially, add the above outputs, and then select one of the multiple codes stored in the area based on the result of the addition. This is achieved by a data processing method characterized by outputting the data as follows.

[For production]

In the present invention, a single management table is prepared in advance in which a plurality of threshold values and a plurality of codes to be referred to by one piece of input data are stored. Then, the input data is sequentially compared with each of the plurality of threshold values recorded in the management table, and if the input data is smaller than the threshold value, the input data is zero.
, if it is larger, output 1 and add the result. The value obtained by this addition indicates how many thresholds are smaller than the input data, and takes a different value for each of multiple areas divided by multiple thresholds. . Therefore, by using this numerical value as an identifier, it is possible to select and output one of the codes corresponding to the plurality of areas.

Furthermore, since input data is processed through the same processing path regardless of its data value, the processing time for all input data is the same. Therefore, if multiple processing paths consisting of comparison sections and addition sections are provided in parallel to process multiple input data in parallel, the data input earlier will be overtaken by the data input later. No inconvenience will occur.

〔Example〕

FIG. 1 is a block diagram showing an embodiment of the present invention, and FIG.
Components with the same functions as those in the figure are given the same numbers. In this embodiment, the data processing for identifying paper sheets described in the prior art section is performed. That is, when the input data that is input one after another is below the lower limit threshold, a lower limit error code is issued, and when the input data is greater than the upper limit threshold, the upper limit error code is used. An example will be described in which data processing is executed to output a normal error code when the error code is greater than or equal to the lower threshold and less than the upper threshold. Management table 1
3, as shown in Figure 2, one word each of the lower limit threshold value, upper limit threshold value, lower limit error code, upper limit error code, and normal code for a certain micro area is written from address A to A+4.
Store in the same area specified by consecutive addresses up to address. Then, an offset of 0.1.2 is set for each of the lower limit error code, normal code, and upper limit error code addresses. A similar method is used for other small areas to store threshold values and codes, for example, at addresses A+5 and below. FIG. 3 is a flow diagram illustrating an embodiment of the present invention. FIG. 4 shows the outputs of comparison section 114, comparison section 1115, and addition section 16 in the block diagram shown in FIG. 1.

First, as shown in FIG. 1, input data is taken into the input interface 11. For example, the input data is 3
Consists of 2 bits, the first 4 bits contain information about the minute area to which the input data belongs, and the subsequent bits)
The ffI area contains data values. Therefore, the first bit of the input data is sent to the address generation section 12, and the subsequent bit area is sent to the comparison section 114. The address generating section I2 generates an address indicating the minute area to which the input data belongs, for example, address A, and sends this to the management table 13, and transitions to address A+1. Comparison part 1
In step 14, the input data is compared with the lower limit threshold value read from address A of the management table 13. As shown in FIG. 4, when the input data is smaller than the lower limit threshold value, O1 is generated, and when it is larger, 1 is generated and sent to the next comparison section 1115 together with the input data.

Next, in the address generation section 2, A+ of the management table 13
The address 1 is sent to the management table 13 and the address is transferred to address A+2. The comparison unit ff15 compares the input data sent from the comparison unit 114 with the upper limit threshold value read from address A+1 of the management table 13. Then, as shown in FIG. 4, when the input data is smaller than the upper limit threshold value, 0 is generated, and when it is larger, 1 is generated and inputted to the adder 16 together with the output of the comparator 114. When the results of the comparator 114 and the comparator 1115 are added in the adder 16, as shown in FIG. When the value is greater than or equal to the value, 2 is output. Therefore, this numerical value is used as an offset to be sent to the management table 13, and an address having a corresponding offset attached to address A+4 from address A+2 is selected.

The code stored at the selected address is then read out and taken out to the outside via the output interface 17. Then, the output interface 17 sends a processing end signal to the input interface 11 to start processing the next input data. By repeating the above processing for successive input data, a code map can be obtained, and paper sheets can be identified using this code map.

In the above embodiment, only the lower threshold and upper threshold are set, but an intermediate threshold is provided between the lower threshold and the upper threshold to set these thresholds. A more precise code map can be obtained by increasing the number of regions divided by can be made to do so.

Further, in this embodiment, the case where input data is input and processed one by one has been described, but when performing parallel processing of a plurality of input data, as shown in FIG. Processing paths consisting of adders 16 are connected in parallel. Then, the address generation unit 12 generates an address from each leading bit of the plurality of input data input in parallel, accesses the management table 13, and sends the address to the comparison unit 114, comparison unit 1115, and addition unit 16 connected in parallel. to perform calculations in parallel. At this time, since each processing time is the same regardless of the data value as described above, the output data are output from the output interface 17 at the same time. Therefore, even when a plurality of input data are input and processed one after another, there is no problem that later input data overtakes previous input data, and parallel processing can proceed normally.

〔Effect of the invention〕

As described above, the present invention is extremely effective when detecting errors in a large amount of input data at high speed through parallel processing.

[Brief explanation of drawings]

Fig. 1 is a block diagram showing an embodiment of the present invention, Fig. 2 is a management table, Fig. 3 is a flowchart showing an embodiment of the invention, and Fig. 4 is the output of the comparator section ■, the comparator section ■, and the adder section. FIG. 5 is a block diagram showing another embodiment of the present invention, FIG. 6 is a block diagram showing problems in the conventional example, and FIG. 7 is a flow chart showing problems in the conventional example. In the figure, 11 is an input interface, 12 is an address generation section, 13 is a management table, 14 is a comparison section I, 15 is a comparison section ■, 16 is an addition section, 17 is an output interface, 18 is a lower limit threshold table, 19 20 is an upper limit threshold table, 20 is a lower limit error code table, 21 is an upper limit error code table, 22 is a normal code table, 23 is a collation unit I, and 24 is a collation unit (2). Block diagram showing an embodiment of the present invention Figure Management table Figure Lower limit threshold Upper threshold ← Q: l l → Output of comparison section 1 ← O: 0: 1 → Comparison section ■
Output ← 0, 1: 2→
Outputs of the adder Figure 4 shows the outputs of the comparison unit I, the comparison unit 2, and the adder. Figure 4 is a block diagram illustrating another embodiment of the present invention. Output data is a block diagram illustrating problems with the conventional example.

Claims (3)

[Claims] (1) A management table (13) storing a plurality of thresholds and a plurality of codes corresponding to a plurality of ranges divided by the plurality of thresholds, and input data and the above in the management table (13). A plurality of comparison units (14) and (15) that compare each of the plurality of threshold values and output data based on the comparison results, and a plurality of comparison units (14) and (15) that output data based on the comparison results. an addition section (16) for adding data;
), the data processing device selects and outputs one of the plurality of codes stored in the area in the management table (13) according to the addition result. (2) The comparison sections (14) and (15) and the addition section (1
6) The data processing device according to claim 1, wherein a plurality of processing paths consisting of the following are provided in parallel. (3) Store multiple thresholds and multiple codes corresponding to multiple ranges divided by the multiple thresholds, and compare input data with the multiple thresholds stored in the above areas. and sequentially outputting the comparison results, then adding the above outputs, and then selecting and outputting one of the plurality of codes stored in the above area based on the above addition result.