JPS63238683A - Optical character reader - Google Patents

Abstract

PURPOSE:To reduce a burden on operation by performing the recognition processing of an image obtained from an image sensor again if the recognition obtained from an image sensor again if the recognition result of a format check part does not accord with a prescribed format. CONSTITUTION:After receiving a signal NEW from a new row detecting circuit 10, a recognition processing control circuit 21 continues to output a signal GO for the recognition processing until a signal OK comes from a format check part 20. Thus, the recognition processing operation is repeated at the time of applying a scanner 1 to a form 3 until the recognition result according with the format is obtained. If the recognition result according with the format cannnot be obtained when an operator applies the scanner 1 to the form 3, he slightly shifts the position, to which the scanner 1 is applied, to retry for the correct recognition result. Thus, format check is performed without increasing the burden on the operation.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention is capable of scanning characters, symbols, etc. (hereinafter, only characters will be described as a representative, but the same applies to symbols as well) by scanning paper with a hand-held scanner. This invention relates to an optical character reading device that reads text.

[Conventional technology]

At supermarkets, department stores, etc., POS (Point of Use) collects sales information for each item and manages inventory.
5ales) system is widespread. This POs
The system often uses hand-held optical character readers.

As such a device, the present applicant has filed a patent application for the device shown in FIG. 11 in Japanese Patent Application No. 41474/1983.

In FIG. 11, reference numeral 1 denotes a scanner, which reads characters written on paper 3 by simply placing it on the paper 3 with hand 2. The paper 3 is, for example, a price tag on which information in a POS system is written.

4 is a light source, 5 is a lens system, and 6 is an image sensor, which must have a field of view of at least one line of characters written on paper 3, which is one line horizontally and three times vertically one character. The field of view is about that. Reference numeral 7 denotes a control/binarization circuit that converts the analog signal that is the output signal of the image sensor 6 into binary signals corresponding to the character area and the background area, respectively, and sends the signals to the memory 8. The memory 8 stores the binarized signal of the entire field of view of the image sensor 6.

Recognition processing is performed from the binarized signals stored in the memory 8, but in this case only the binarized signals in which the characters are correctly within the visual field are subjected to the recognition processing. That is, the 12th
The figure shows an example of scanning with the image sensor 6 and converting it into binarization, and this figure shows a case where characters are correctly included in the field of view 17. FIG. 13 shows a case where scanning is performed with characters protruding from the visual field 17. In this case, the recognition process described later is not performed, and as the scanner 1 moves, the image sensor 6 scans again, binarizes, and stores the digitized signal in the memory 8, thereby improving the reading speed. . In addition,
In FIGS. 12 and 13, some vertical lines indicating pixel divisions are omitted.

The scanning position detection circuit 9 determines whether or not to perform recognition processing. That is, the scanning position detection circuit 9 operates in parallel with storing the binary signal into the memory 8, and determines whether or not to perform recognition processing almost at the same time as the storage of the binary signal into the memory 8 is completed. Deciding. Furthermore, when performing recognition processing, the scanning position detection circuit 9 does not process the entire visual field;
The area to be recognized is determined so that only the area where a character exists is recognized, and a SET signal indicating the upper end of the character position and an END signal indicating the lower end are generated. When there is no character, that is, when the SET and END signals are not output, the CLEAR signal is output.

The new line detection circuit 10 determines the character position from the signals SET and END sent from the scanning position detection circuit S, and compares it with the character position determined by the SET and END signals sent in the next scan, and determines the amount of change in the character position (in pixels). number) or signal CLEA
A judgment is made based on whether the character string has protruded once from the field of view 17 using R, and if it is determined that a new character string exists, a signal NEW is sent to the recognition processing circuit 15. After the signal NEW is sent, the next recognition process is performed.

This line breakage detection circuit 10 allows the scanner 1 to be simply moved from top to bottom or from bottom to top when reading a character string with more lines than the field of view. That is, since the scanner 1 is placed on the paper 3 to read characters, if the scanner 1 is left in place for a long time, the same part will be scanned and read many times, but the new line detection circuit 10 prevents this.

Next, cutting out one character will be explained.

Since the binarized signal stored in the memory 8 is the entire field of view of the image sensor 6, it is larger both vertically and horizontally than the number of pixels that can be processed by the recognition processing circuit 15 that recognizes one character. For this reason, the amount of data that can be processed by the recognition processing circuit 15 is converted into a binary signal of the number of pixels, that is, one character at a time.

The number of pixels that can be processed by the recognition processing circuit 15 will be explained assuming that the number of pixels that can be processed by the recognition processing circuit 15 is n horizontally and m vertically (column direction) in FIG.

First, a binarized signal of p in the horizontal direction and pixel number P of the entire field of view 17 in the vertical direction is transferred to the one-digit buffer 11 . That is, the binarized signals from column B1 to column 83 in rows L1 to L2 are single-digit buffer 1.
It is stored in 1.

The vertical position detection circuit 12 detects the vertical position of one character in the binary signal of the single-digit buffer 11, and reduces the binary signal to pixels m in the vertical direction that can be processed by the recognition processing circuit 15. . The vertical position detection circuit 12 determines the vertical position of the character, and determines the number m of vertical pixels that can be processed by the recognition processing circuit 15 (determined so that the character position is located approximately at the center). As a result, the vertical direction m is determined from the binary signal stored in the minus digit buffer 11 and transferred to the one character buffer 14 as one character data mXn. In the example shown in FIG. 12, the binarized signal contained in the area of L+ mustard 1st row, B, column to 87th column is transferred to the 1-character buffer 14 as 1-character data.

The lateral position detection circuit 13 determines the horizontal position of the binarized signal to be stored when the next binarized signal is stored in the negative digit buffer 11.

The recognition processing circuit 15 recognizes one character from the binary data stored in the one-character buffer 14, and may be of a known type, such as that described in Japanese Patent Publication No. 59-6418. That is, this recognition processing circuit 15 extracts feature data of character constituent parts for each row or each column in the vertical and horizontal character areas, and outputs a character recognition result corresponding to the MM and appearance order of this feature data.

[Problem that the invention seeks to solve]

In the character string written on the price tag at PO8, the number of digits and the type of character on the left end are determined in advance (this is called a format). Conventionally, this format inspection (format check) is performed on the recognition result output by the recognition device section. If the format matches, the recognition result is output, and if it does not match the format, the recognition result is output.
Does not output recognition results. This can significantly reduce the rate of misreading. To perform a format check,
It may seem that it is sufficient to simply combine a format checking section with the optical character reading device according to the prior art described above. However, in the conventional optical character reading device, after the recognition process has been performed - times, the recognition process is not performed until the new line detection circuit detects a new line. Additionally, lines that do not match the format due to misreading will never be reread unless the scanner is briefly removed from the paper and reapplied. That is, in the event of a reading error, the operator must remove the scanner from the paper and reapply it, which places a heavy burden on the operator.

The present invention takes this point into consideration and aims to implement format checking without increasing the operational burden.

[Means for solving problems]

FIG. 1 shows the configuration of the present invention. The scanner 1 is applied to the paper 3, the image captured by the image sensor 6 is stored in the memory 8, and then the recognition processing circuit 1 cuts out each character one by one.
The processing steps up to recognition in step 5 are the same as in the conventional method. However, it is different from the conventional method in that it includes a format checker and in the way recognition processing is controlled.

The format check unit 20 performs a format check on the recognition result from the recognition processing circuit 15, and if it satisfies a predetermined format, the recognition result Dout
and outputs an OK signal.

Conventionally, the new line detection circuit 10 controls the recognition process.
This was done only once when the signal was output. However, in the present invention, after the new line detection circuit 10 outputs the NEW signal, the recognition process is repeatedly performed until the format check section outputs the OK signal. 2
1 is a recognition processing control circuit for this purpose.

[Effect]

The recognition processing control circuit 21 receives the NE from the new line detection circuit 10.
After receiving the W signal, the GO signal for performing recognition processing is continued until an OK signal is received from the format check section 20. As a result, when the scanner 1 is applied to the paper 3, the recognition processing operation is repeated until a recognition result that satisfies the format is obtained. When applying scanner 1 to paper 3, if a recognition result that satisfies the format is not obtained, the user can slightly shift the position of scanner 1 (the pixels of the image sensor and the characters on the paper). You can try to obtain the correct 1m result (which may or may not be recognized correctly depending on the positional relationship)

[First Embodiment] The recognition processing WiiT1 circuit 21 can be realized by, for example, an SR flip-flop. The recognition processing control method described above is designed to "perform recognition processing repeatedly when a recognition result that does not satisfy the format is obtained" (A).
``When a recognition result that satisfies the format is obtained, the recognition processing operation is stopped.'' However, (A
) There are several ways to achieve this:

[Second Embodiment] (The recognition processing operation is always repeated, but after the recognition result that satisfies the format is obtained, the recognition processing result is discarded.) Fig. 2 shows the second embodiment. After storing the images captured by the image sensor 6, the recognition processing circuit 1
5, the processing up to the recognition processing is always performed regardless of the results of the new line detection circuit or the format check section. The output control unit 51 outputs N from the new line detection circuit 10.
After the EW signal arrives, the format check section 20
The input In is output to the output Out until the signal K is output, but after the OK signal is received from the format check unit 20, the human input In is not output to the output Out, that is, the output of the recognition processing circuit 15 is discarded. be left.

FIG. 3 shows an embodiment in which the format check section and the order of output control are reversed. The operations from the scanner 1 to the recognition processing circuit 15 are the same as in FIG. 2.

After receiving the NEW signal from the new line detection circuit 10, the output control section 51' outputs only the first input In obtained from the format check section 20' to the output Out, but the second and subsequent inputs In, The output of the format check unit 20' is not output to the output Oat, that is, the output of the format check section 20' is discarded.

[Third Embodiment] (Configuration without new line detection circuit) The new line detection circuit in Fig. 3 detects whether a new line has come into view, but in addition to the format check section, a timer is By using , it is possible to omit the new line detection circuit. Now, the scanner has a field of view of one line and reads one line of characters at a time, as shown in Figure 4 (a).
) →...1- (C) As shown in (C), the price tag with only one line written on it is changed and the price tag is changed to guess which scanner is 1. At this time, it takes time (for example, 1 second) to switch from ta) to (bl, mountain) to (C1), so if no recognition result is obtained for a certain period of time, transition to a new line. Therefore, an embodiment using a timer and omitting the new line detection circuit is shown in FIG.
The timer 81 measures the time since the format check section 20' outputs the output. Then, when a certain period of time (for example, 1 second) has passed, a NEW signal is output to the output control section 51. After giving the NEW signal, again,
When the format check section 20' outputs an output, the NEW signal is stopped. When the NEW signal is not being output, if the format check section 20' continues to output an output within a certain period of time, the NEW signal is not output.

The format check section 20' and output control section 51' are the same as those shown in FIG. Scanning position detection circuit 9 in FIG.
Unlike the scanning position detection circuit 9 in FIG. Only recognition results that meet a predetermined format obtained in the above process are output. In addition, as shown by the dotted line in FIG. 5, an embodiment is also possible in which the recognition processing circuit 15 measures the time elapsed after the recognition result is obtained using a timer 81.

In the above explanation, we used as an example a recognition circuit in which a scanner has a field of view of one line and reads one line of characters at a time. The configuration shown in FIG. 5 is also possible for a recognition processing circuit that reads multiple lines written on a price tag at once.

[Fourth Embodiment] (Improvements of the Third Embodiment) In the third embodiment, when reading a new line, it is necessary to wait at least a certain period of time (for example, 1 second) after reading the previous line. However, when reading multiple lines on a price tag using a scanner that reads one line at a time, the need to wait a certain amount of time (for example, 1 second) between each line is time-consuming. It is uneconomical in this respect.

Now, as shown by the arrows in Figure 6, (a), mountain), (C
) Consider reading each line of a price tag in turn.

As mentioned above, it takes a certain amount of time to transfer the price tag from (al to mountain) to (bl to (C1).

When reading each line in a price tag, when applying the scanner to each line in turn, it takes very little time (for example, 066 seconds) to reapply the scanner between each line. Here, PO3 according to JISB9551
In a commercial price tag, each line in one price tag has different content. Therefore, by comparing the results of the previous recognition and the results of the next recognition, it can be determined whether the same line has been read.

In view of this point, FIG. 7 shows an embodiment of the present invention in which price tags can be read as shown in FIG. 6.

The previous register 101 is for storing the output of the format check section until the output is obtained from the format check section again. The comparator 102 is
The output contents from the format check section and the contents of the previous register are compared, and if they are different, the output control section 51
A NEW signal is output to '. The output control section 51' has N
When the EW signal comes, respond to it. Outputs the input In to the output Out. The timer 81 measures the time from when the output is output from the format check section like the one shown in FIG. Send 101 helicent signal. The output of the format check section is temporarily saved in the previous register, but after a certain period of time is counted by the timer, the contents of the previous register become empty. Since it is definitely different from the content, the comparator outputs a NEW signal, which is output from the output control section. When reading each line in one price tag in order, the data is stored in the previous register, and when a different content from the previous register is output from the format checker, that is, when a new line is read, the comparison is performed. A NEW signal is output from the device 102, and the contents of the new line are output from the output control section. At this time, outputs are obtained from the format check section one after another before the timer finishes counting the fixed time, so the contents of the previous register are not reset.

Note that the time measured by the timer 81 is the same as the dotted line in FIG. It is also possible to use the time since R15 outputs the recognition result.

[Fifth Embodiment] In the embodiment shown in FIG. 5 or 7, the timer 81 is used to measure the time required to change the price tag, that is, the time during which no recognition processing result is obtained, so that the replacement of the price tag is performed. was detected. Therefore, the timer in FIG. 5 or 7 can be replaced by a paper detector that detects the presence or absence of paper 3. FIG. 8 shows an embodiment in which the timer in FIG. 7 is replaced with a paper detector. In FIG. 8, parts with the same numbers as in FIG. 7 have the same functions as in FIG. The paper detector sends a signal to the gate 11 that becomes true when the scanner 1 is applied to paper.
Send to 2. Gate 112 inverts the paper detector signal and sends it to the reset input of previous register 101. That is, when the paper detector does not detect paper, the contents of the previous register will be empty. On the other hand, when the paper detector detects paper, the contents of the previous register will never be empty, so if the price tag is replaced, only the result that passed through the format check section will be displayed. Output. When reading each line of the same price tag, the output will only be output if the content is different from the previous one, that is, if a new line is read.

The paper detector includes, for example, a proximity sensor using ultrasonic waves, a pushbutton switch attached to the tip of the scanner (when the scanner is placed on paper, the pushbutton is pressed and the switch closes), and the output of the image sensor 6. A comparator IC that compares the level with a predetermined threshold value (when a scanner is applied to a sheet of paper, it detects that the output of the image sensor increases corresponding to the white part of the sheet), or an image sensor 6 The image obtained from
This can be achieved by a circuit that detects the presence of fine black and white slams (when a scanner is applied to a sheet of paper, it detects that fine black and white taps such as letters written on the sheet appear in the image).

In addition, in FIG. 5, an example in which a paper detector is used instead of a timer is shown in FIG.
After a NEW signal is sent from the paper detector 111 to the output control section 51', the recognition result that passes through the format check section and is obtained first is outputted via the output control section 61'. After the price tag is replaced, the recognition results that pass the format check section and are obtained from the second time onward are suppressed by the output control section and are not output.

[Sixth Embodiment] FIG. 10 shows an example in which a recognition processing control circuit is used in place of the output control section 51' of FIG. 9.

When the scanner 1 is applied to the paper 3, a pulse-like signal indicating that paper has been detected is sent from the paper detector 111 to the recognition processing control circuit 21, and the recognition processing control circuit outputs a signal GO to perform the recognition processing. be done. As a result, recognition processing starts, and the result of vl recognition in the recognition processing circuit is sent to the format check section. If the predetermined format is not met, the result is suppressed (read and discarded) by the format check section, but if the predetermined format is met, the result is outputted via the format check section. The format check section 20 sends an OK signal to the recognition processing circuit 21 . The recognition processing control circuit 21 repeats the reading and recognition processing until an OK signal is received from the format check section 20, that is, until a recognition result that satisfies the predetermined format is obtained, and when the OK signal is received, the recognition processing is performed. stop. The recognition processing control circuit 21 restarts the recognition processing when
This is when the paper 3 is replaced and a paper detection signal is received again from the paper detector 111.

With the above operation, only the first result obtained after applying the scanner 1 to the paper 3 that satisfies a predetermined format can be outputted from the format check section 20.

[Seventh Embodiment] In the fourth or fifth embodiment, the comparator 102 can compare all the characters in the recognized line. If the function codes of the lines that appear on the price tag (characters such as P1■, Embodiments are also possible. Furthermore, if only the function codes are compared by the comparator, it is sufficient to store only the function codes in the previous register 101. Examples in this case are shown in FIG. 14 as the seventh embodiment and as shown in FIG. 14 as the eighth embodiment. It is shown in Figure 15.

In FIG. 14, the configuration and processing up to the format check section 20' are the same as those in FIG. 7. 1
Reference numeral 41 is a section for extracting a function code, which extracts the function code from the recognition result that has passed through the format check section 20' and outputs it to the comparator 102 and the previous register 101.
send to Timer 81 measures the time from when the function code is output from 141. Last time register 10
1 stores the function code immediately after obtaining it from 141, but timer 8
After the time measured by 1 has elapsed, the previous register is initialized. The operation of the comparator 102 and the output control unit 51' is as follows except that the function code obtained from 141 and the contents of the previous register 101 are compared in the comparator.
The operation is the same as in the embodiment.

[Eighth Embodiment] In FIG. 15, the configuration and processing up to the format check section 20' and the configuration and processing up to the gate 121 are the same as those in FIG. Reference numeral 141 is a section for extracting the function code, and the format check section 20'
Among the recognition results passed through, the function code is sent to the comparator 102 and the previous register 101. Previous register 1
01, the comparator 102, and the output control section 51' operate in the same manner as in FIG. 14, except that the previous register 101 is initialized by the output of the gate 112.

In FIGS. 14 and 15, function codes are directly compared. However, rather than directly comparing the function codes, an embodiment is also possible in which the function codes are once converted into another code and then compared indirectly.

〔Effect of the invention〕

According to the present invention, it is possible to realize an optical reading plan that has an easy-to-use formatting function! Wear. Thereby, the misreading rate can be dramatically reduced.

[Brief explanation of drawings]

Figure 1 is a block diagram showing the configuration of the present invention, and Figure 2 is a block diagram showing the configuration of the present invention.
FIG. 3 is a partially omitted block diagram of a modification of the second embodiment, FIG. 4 is an explanatory diagram of how to read a price tag, part 1, and FIG. 5 is a partially omitted block diagram of a modification of the second embodiment. Block diagram with parts omitted, Figure 6 is the second explanatory diagram of how to read the price tag, and Figure 7 is the fourth diagram.
FIG. 8 is a partially omitted block diagram of the fifth embodiment; FIG. 9 is a partially omitted block diagram of a partially omitted modification of the fifth embodiment; FIG. 10 is a partially omitted block diagram of the sixth embodiment. FIG. 11 is a block diagram of the conventional example, FIGS. 12 and 13 are conceptual explanations of the field of view of the image sensor, and FIG. 14 is a partially omitted block diagram of the seventh embodiment. Figure, No. 15
The figure shows a partially omitted block of the eighth embodiment. 1...Scanner, 2...Hand,
3...Paper, 4...Light source, 5...Lens, 6...
Image sensor, 7...Control/binarization circuit, 8
...Memory, 9.9'...Scanning position detection circuit, 10...New line detection circuit, 11...
...-Digit bath sofa, 12...Vertical position detection circuit, 13...Horizontal position detection circuit, 14...-Character buffer, 15...Recognition Processing circuit, 16...Control circuit, 17...Field of view, 18...Number of horizontal pixels in character recognition processing,
20.20'...Format check section, 2
1... Recognition processing control circuit, 51.51'... Output control section, 81...
・Timer, 101...Previous register, 10
2... Comparator, 111... Paper detector, 112... Gate, 141...
Function code extraction part. Patent applicant: Sumitomo Electric Industries, Ltd. Agent: Kama 1) Text 2 Figure 12 Figure 13 Procedural amendment. picture. 1. Display of case Patent application No. 56293 of 1988 2, name of invention Optical character reader name (8 system (213) Sumitomo Electric Industries Co., Ltd.
Correct content (2) The unpublished "Optical reading number proposal" on page 24 is corrected to "Optical reading device J.

Claims (7)

[Claims] (1) An optical character reading device that includes an image sensor that fits multiple characters, symbols, etc. for one line within its field of view and a means for recognizing each character on the screen captured by the image sensor, which includes a format check unit. . An optical character reading device characterized in that when the recognition result does not satisfy a predetermined format, recognition processing is performed again on the image obtained from the image sensor. (2) The optical character reading device according to claim 1 further includes a new line detection circuit that detects whether or not a line that has entered the field of view of the image sensor has been updated, and the new line detection circuit detects a new line. An optical character reading device characterized in that after detection, outputs a first obtained recognition result that satisfies a predetermined format. (3) The optical character reading device according to claim 1, further comprising a timer for measuring a predetermined time and an output control section, wherein after a recognition result is not obtained for a predetermined period of time, the first predetermined predetermined reading result is obtained. What is claimed is: 1. An optical character reading device that outputs a recognition result that satisfies a format. (4) In the optical character reading device according to claim 3, the contents of the previous register and the contents of the next recognition result are further compared with the previous register that stores the recognition result until the next recognition result is obtained. It is characterized by having a comparator and outputting a recognition result that satisfies a predetermined format and has different content from the previous one, or a recognition result that is first obtained and satisfies a predetermined format after no recognition result has been obtained for a certain period of time. optical character reader. (5) The optical character reading device according to claim 1, further comprising a paper detector that detects proximity or contact between the image sensor housing and the paper, and first after the paper detector detects the paper. An optical character reading device that outputs a recognition result that satisfies a predetermined format. (6) In the optical character reading device according to claim 5, the content of the previous register and the content of the next recognition result are further compared with the previous register that stores the recognition result until the next recognition result is obtained. It is characterized by having a comparator and outputting a recognition result that satisfies a predetermined format and has different content from the previous one, or a recognition result that satisfies the predetermined format obtained first after the paper detector detects the paper. Optical character reader. (7) The content to be compared in the comparator in the optical character reading device according to claim 4 or 6 is a function code included in the recognition result, and the function codes are directly or indirectly compared in the comparator. An optical character reading device characterized in that it is compared to.