JPH0478981A - Optical character reader - Google Patents

Abstract

PURPOSE:To shorten time for reading by executing a character identification processing based on an inverted picture only in an optimum segmented state to execute character identification without fail and preventing the character identification processing based on the inverted picture from being executed in the other segmented state in which character identification is made defective. CONSTITUTION:A character identifying circuit 83 is operated both as an identifying means and as a detecting means, and when the picture is segmented in the optimum segmented state while a segmenting means 82 successively segments pictures stored in a storing means 81, this is detected by the detecting means. While receiving this detected result, the identifying means executes the identification processing based on the inverted picture prepared by inverting the picture in the optimum segmented state by a character inverting means 85. This inverted picture is in the optimum segmented state positioning a character at the center, and identification is satisfactorily executed. Therefore, the character identification processing is executed only based on the picture in the optimum segmented state concerning the inverted picture, and it is not executed concerning the inverted picture in the state that the character is not completely positioned within the segmented picture. Thus, time until obtaining an identified result can be shortened.

Description

DETAILED DESCRIPTION OF THE INVENTION <Industrial Application Field> The present invention relates to an optical character reading device that optically reads characters such as kanji, kana, alphabets, and numbers.

<Prior Art> Optical character reading devices that optically read character strings representing product prices, product codes, etc. and input the reading results into computers have been widely used in the distribution industry and other industries. The system is designed to improve the efficiency of data entry. Such an optical character reading device generally illuminates a reading surface on which a character such as "" is formed, such as the surface of a product's packaging, and uses a lens system to image the reflected light from the reading surface on the light-receiving surface of an image sensor. By doing so, the optical image of the characters on the reading surface is converted into an electrical signal.

The electrical signal output from this image sensor is binarized by level discrimination using an appropriate threshold level.
20 Character identification is performed based on binary signals.

To identify a character, first, an image of one character is extracted from the image read by the image sensor, and then the characteristic amount of the character in the cropped image is compared with the characteristic amount of the character pattern stored in advance. It is done in such a way that

Recently, by changing the order in which signals are output from the memory that stores images read by an image sensor, characters can be identified by reversing the top and bottom of the read image. Devices that can also read information are also coming into use.

The reading process for inverted characters in such a device is shown in FIGS. 7 and 8. As shown in Fig. 7, when a string of inverted characters enters the field of view 1 of the image sensor, the image read by the image sensor is stored in a memory (not shown) in the form shown in Fig. 7. Assume a case. As shown in FIG. 8fa), by successively producing memory signals for each predetermined area that can contain one character, the image of the predetermined area is sequentially cut out, and the cut-out image and the pre-stored character pattern are combined. A comparison is made. Since character patterns are prepared only for upright characters, in this case there are no corresponding characters, and therefore the character string cannot be recognized.

Next, by reading out the stored signals of the memory in the reverse order for each predetermined area, the stored image is turned upside down and cut out one by one as shown in FIG. 8(b). and,
Character identification processing is performed by comparing the cut out image and the character pattern.

In this case, the image to be identified will contain upright characters due to the inverted processing performed by reversing the order of successive output from memory, so the characters will be located approximately in the center of the cut out image. Characters can be identified in the states indicated by reference lines β1 to β15.

<Problem to be solved by the invention> However, in the above-mentioned reading device capable of reading upright and inverted characters, each process of character extraction and character identification in the upright state shown in FIG. Since the character extraction and character identification processes are performed even in the inverted state shown in Figure 8 (b+), it takes a long time to obtain the identification results, which reduces the time required for data input. There was a problem with it being too long.

SUMMARY OF THE INVENTION Accordingly, it is an object of the present invention to provide an optical character reading device that solves the above-mentioned technical problems, shortens the time required to obtain an identification result, and speeds up data input. With the goal.

Means for Solving the Problems> To achieve the above object, the optical character reading device of the present invention includes an image sensor that forms an optical image of a character formed on a reading surface on a light receiving surface, and an image sensor that forms an optical image of a character formed on a reading surface. A storage means for storing an image read by an image sensor, a cutting means for cutting out an image of the size of one character from the storage means, and an identification means for identifying a character from the image cut out by the cutting means. , an optical image reading device having an editing means for extracting and outputting a correct identification result from the identification result by the identification means, in which a character is located approximately in the center of the image cut out by the cutting means; The identification means includes a detection means for detecting an appropriate appearance state, and a character reversing means for inverting at least an image cut out in the optimum appearance state, and the identification means detects the above character when the detection means detects the optimum appearance state. The editing means identifies characters based on the image obtained by inverting the image in the optimal output state using the reversing means, and the editing means identifies the characters based on the identification result based on the image from the cutting means and the character inversion. The present invention is characterized in that it outputs the correct identification result among the identification results based on the inverted image from the means.

<Operation> According to the above configuration, the image read by the image sensor is stored in the storage means, and the stored image is cut out into images of the size of one character by the cutting means and identified by the identification means. . The image cut out by the cutting means is not necessarily in the optimal state where the characters are located approximately in the center; for example, there may be a state in which only about half of the characters are included. For such cut-out images, the identification means is unable to correctly identify the characters, or incorrect identification results are discarded by the editing means.
Only correct identification results can be obtained from this editing means.

While the images stored in the cutting means or the storage means are successively cut out, when the image is cut out in the above-mentioned optimum output state, this fact is detected by the detection means.

In response to this, the identification means performs identification processing on the inverted image created by inverting the image in the optimum appearance state using the character reversing means. This inverted image is, of course, an image in the optimal state where the characters are located in the center, and the identification by the identification means can be performed satisfactorily.

In this way, for inverted images, character identification processing is performed only on images in the optimal appearance state, and character identification processing is not performed for inverted images in which characters are not completely included in the cut-out image. do not have. As a result, processing by the character identification means can be reduced.

The editing means outputs either the correct identification result based on the image as it is cut out by the cutting means or the identification result based on the above-mentioned inverted image, and outputs the correct identification result based on the image cut out by the cutting means and the identification result based on the inverted image. It can be read whether the device is in an upright position or in an inverted position.

<Example> Embodiments will be described in detail below with reference to the accompanying drawings showing embodiments.

FIG. 1 is a block diagram showing the basic configuration of an optical character reading device according to an embodiment of the present invention. This device is housed in a handy housing and includes a light source 2 that illuminates a reading surface 1 on which characters to be read are formed, and an optical image formed on the light receiving surface 6a by reflected light from the reading surface 1, which is converted into an electrical signal. It is equipped with a scanner 4 that houses a two-dimensional image sensor 6 and the like.

When the scanner 4 is placed on the reading surface 1, the reflected light from the reading surface 1 is reflected by the reflecting mirror 3 and further guided to the light receiving surface 6a of the image sensor 6 via the lens system 5. An optical image of the characters formed on the reading surface 1 is formed thereon. The image sensor 6 converts this optical image into an electrical signal and supplies it to the binarization circuit 7. This binarization circuit 7 level-discriminates the output of the image sensor 6 using an appropriate threshold level and converts it into a binary signal of NJ or "0".

The image sensor 6 has a large number of photoelectric conversion pixels arranged two-dimensionally so that an optical image of one or more lines of character strings formed on the reading surface 1 can be formed on the light-receiving surface 6a at once. The electrical signal obtained for each pixel is binarized by the binarization circuit 7 described above.

The output signal of the binarization circuit 7 is given to the main body processing section 8 via lines 1-9. This main processing unit 8 temporarily stores the signal from the cable 9 in a storage circuit 81 . Since the signal from the binarization circuit 7 is a binary signal of NJ or "0", the storage area of the storage circuit 81 stores the binary image formed on the light receiving surface 6a of the image sensor 6. become.

The image stored in the memory circuit 81 is read out one by one by a character cutting circuit 82 in a predetermined area that can accommodate one character, and the image thus cut out is read out by a character identification circuit 83.
is given and an identification process is performed. This identification process is
This is performed by calculating the difference between the feature amount of the character pattern of the upright character stored in advance in the character identification circuit 83, and if the character is located approximately in the center of the image extracted by the single character extraction circuit 82, the character is recognized. or will be recognized. This recognition result is given to the editing section 84. This editing section 84 outputs the identified information if the identification result is correct,
If it is an error, the identification result is ignored and the start of the next reading is instructed.

The image read out from the memory circuit 81 by each predetermined area by the character cutout circuit 82 is also provided to the character reversal circuit 85 . This character reversing circuit 85 is supplied with a cutout signal or a line that is output when the character identification circuit 83 or the single character cutout circuit 82 detects the optimal appearance state in which a character is located approximately in the center of the cutout image. 86. In this manner, in this embodiment, the character identification circuit 83 functions not only as an identification means but also as a detection means.

The above cutting signal is also given to the single character cutting circuit 82, and this single character cutting circuit 82 responds to the above cutting signal to obtain the cutting position information of the cut out image in the optimum cutting state. (For example, the coordinates of the cut-out image within the storage area of the storage circuit 81)
is applied from line 87 to character reversal circuit 85. In response to the cutout signal, the character reversal circuit 85 inverts the cutout image in the optimum output state by reading out the signals constituting this image from the storage circuit 81 in the reverse order.

An image that has been subjected to this inversion process (an image of an area for one character)
is applied from line 88 to character identification circuit 83 to identify the character in the image. The identification results are from the editorial department 8.
given to 4.

FIG. 2 is a diagram for explaining the inversion process of the character reversal circuit 85. Coordinates (x+, y+1) to (X., y) within the storage area 81a of the storage circuit 81 for each pixel in a predetermined area S that is extracted from the single character extraction circuit 82 or the storage circuit 81.
. ), the one-character extraction circuit 82 will, for example, (Xy+), (X2.y+), . . . , (x., y
, )(x+, y2), . . . are read out from the storage circuit 81 in this order. For example, the coordinates (X+, y+) are given to the character reversing circuit 85 as cutting position information, and based on this, the character reversing circuit 85 calculates the coordinates (Xi, y.). , y□) in order from (x
e, yyu), (Xn-l+ y,), ... (x+,
memory circuit 8 in the order of yyo), (x,,y.-3),...
1, the signals of each pixel are read out to create an inverted image.

FIG. 3 is a flowchart for explaining the operation of the above reading device. In step nl, the character cutting unit 82 sequentially cuts out one character's worth of images in an erect state from the memory circuit 81 (and in step n2, the cut out erect images are sequentially identified by the character identification circuit 83). If the character identification circuit 83 does not derive a cutting signal in step n3, the process returns to step n1 and continues the cutting process of the erect image.

When the character identification circuit 83 detects the optimal output state and outputs a cutout signal in step n3, the identification result of the character identification circuit 83 based on the cutout image at this time is sent to the editing section 84 in step n4. Further, in step n5, the optimum output image is read out from the storage circuit 81 in the reverse order, and the character inversion circuit 85 performs an inversion process.

The inverted image thus obtained is identified by the character identification circuit 83 in step n6. The cutout image identified at this time is of course the image in the optimal output state, and the identification result at this time is given to the editing section 84 in step n7. The subsequent processing returns to step nl.

The editing unit 84 outputs the correct one of the identification results for the erect image and the inverted image. If any of the outputs is incorrect, the reading operation continues, and if all of the outputs are correct identification results, for example, if the character to be read is vertically symmetrical, such as "ro 690," For example, a message indicating that both an upright image and an inverted image have been correctly identified is displayed on a display device (not shown) to alert the operator.

FIG. 4 is an explanatory diagram showing the operation of reading an inverted character, centering on the processing of cutting out characters from the memory circuit 81. Since the single character cutting circuit 82 reads out the image stored in the memory circuit 81 for each predetermined area S where one character can fit, there is a situation where no character is included in the area S, or only about half of the character is included. There are states to enter. If the image is not located at the center of the cut out image, character identification processing is performed based only on the erect image. And for example the reference jl!
When the optimum appearance state is reached in which a character is located approximately in the center of the area S as shown by 21 to ff25, the character reversal circuit 85 performs the inversion process as described above, and the memory circuit 81 as shown by reference numerals f231 to f35 is reached. The image within the image is inverted and cut out, and the character recognition circuit 83 recognizes the characters that have been turned into a royal state through this inversion process.

As described above, in the optical character reading device of this embodiment, when the character identification circuit 83 detects the optimal output state and outputs a cutting signal, in response to this, the image in the optimal output state is turned upside down. The inverted image is supplied to a character identification circuit 83, and characters are identified based on this inverted image. That is, if the cutout state is other than the optimum appearance state and character identification is impossible, the identification process based on the inverted image is not performed in vain. As a result, the processing of the character identification circuit 83 is reduced and the character identification processing is speeded up, and identification of both upright and inverted characters can be completed in a short time. As a result, the efficiency of data input work using the optical character reading device can be significantly improved.

Furthermore, in this embodiment, as described above, if correct reading results are obtained for both the upright and inverted images, the operator's attention is called to the operator's attention, so that the computer or the like may have received incorrect information. Prevented from being entered.

FIG. 5 is a block diagram showing the basic configuration of another embodiment of the present invention. In FIG. 5, parts corresponding to those shown in FIG. 1 described above are given the same reference numerals. In this embodiment, the signal from the memory circuit 81 is delayed in the delay circuit 90 by the time required for cutting out one image in the one character cutting circuit 82, and then sent to the character reversing circuit 82.
1 is given. The image given to the character reversal circuit 91 is the same as the image that has been subjected to the above-mentioned delay processing or is cut out by the single character cutout circuit 82. That is, the character reversing circuit 91 always performs processing to invert the delayed image by cutting out the single character cutting circuit 82.

The cutting signal output from the character recognition section 83 is given to the single character cutting circuit 82 and the character reversing circuit 91, and when the single character cutting circuit 82 is given the cutting signal, the cutting signal required to cut out one image is given to the single character cutting circuit 82 and the character reversing circuit 91. Stop the cutting process after a certain amount of time, and then
1 provides the above-mentioned inverted image to the character recognition section 83 in response to the above-mentioned cutting signal.

In this way, when the cutout signal is derived, that is, when the character is in the optimal output state where it is located approximately in the center of the cutout image, character recognition processing is performed based on the above-mentioned inverted image. It turns out.

FIG. 6 is a flowchart for explaining the operation.

In step ml, characters are cut out one character at a time by the character cutting circuit 82, and this cut out image is processed by the delay circuit 90.
After this delay, in step m2, the character reversal circuit 91 performs an inversion process. On the other hand, the image cut out by the single character cutout circuit 82 is identified by the character identification circuit 83 in step m3. The above process is repeated until the cutout signal is derived in step m4, and when the cutout signal is derived, the character identification circuit 83 performs character recognition in step m5, and sends the recognition result to the editing section 84. give.

In step m6, the character identification circuit 83 is connected to the character reversal circuit 9.
Characters are identified based on the inverted image from 1. Then, in step m7, the identification result is provided to the editing section 84. The subsequent processing returns to step ml. delay circuit 90
As a result of the above delay processing, the inverted image derived to line 88 after the character identification circuit 83 derives the cutout signal becomes an inverted image in the optimal output state corresponding to the cutout signal.

With this configuration, as in the case of the first embodiment, it is possible to reduce wasteful character identification processing and speed up the processing.

Note that the present invention is not limited to the above embodiments. For example, in each of the above embodiments, reading is performed with the scanner 4 fixed relative to the reading surface l, or a one-dimensional image sensor is applied instead of the image sensor 6. 4, the reading surface 1 may be scanned to read characters. Further, in the embodiment shown in FIG. 5 above, the delay circuit 90 may be provided after the character reversing circuit 91. Various other design changes may be made without departing from the gist of the present invention.

<Effects of the Invention> As described above, according to the optical character reading device of the present invention, character identification processing is performed based on an inverted image only in the optimal appearance state in which character identification can be reliably performed. , in other cut-out states where character recognition is poor, character recognition processing is not performed based on the inverted image, so the time required for reading is shortened and characters can be read at high speed. This can contribute to speeding up data input.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the basic configuration of an optical character reading device according to an embodiment of the present invention, FIG. 2 is an explanatory diagram for explaining inverted processing, and FIG. 3 is an explanatory diagram for explaining reading operation. FIG. 4 is an explanatory diagram showing the reading operation; FIG. 5 is a block diagram showing the basic configuration of another embodiment of the present invention; FIG. 6 is a flowchart explaining the reading operation; 7 and 8 are explanatory diagrams for explaining the prior art. 6... Image sensor, 81... Memory circuit (memory means), 82... 1 character cutting circuit (cutting means), 83...
・Character identification circuit (identification means), 85.90 ・Character reversal circuit (reversal means)

Claims (1)

[Scope of Claims] 1. An image sensor in which an optical image of a character formed on a reading surface is formed on a light-receiving surface; a storage means for storing the image read by this image sensor; and 1 from this storage means. A cutting means for cutting out an image of the size of a character, an identification means for identifying characters from the image cut out by this cutting means, and an editing method for extracting and outputting a correct identification result from the identification result by this identification means. an optical image reading device comprising: a detection means for detecting an optimal appearance state in which characters are located approximately at the center of the image cut out by the cutting means; and character reversing means for inverting the image, and when the detecting means detects the optimal appearance state, the identification means is based on the image obtained by inverting the image of the optimal appearance state with the character reversing means. and the editing means outputs a correct identification result, either one based on the image from the cutting means or the identification result based on the inverted image from the character reversing means. An optical character reading device characterized by: