JP2001236463A - Optical character reader - Google Patents

Abstract

PROBLEM TO BE SOLVED: To automatically detect that a pencil has been used or a document has been corrected with correction liquid. SOLUTION: The difference between the output image of a sensor 9, obtained with the light irradiating the top surface of a document 3 from a reflection light source 1, and the transmission output image of a sensor 10 obtained with the light irradiating the reverse surface of the document from a light source 2 are measured and calculated. When the ratio of a black image occupying the difference image is found to be larger than a specific value, it is decided that correction liquid has been applied.

Description

DETAILED DESCRIPTION OF THE INVENTION

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optical character reader for optically reading characters written on a form, and more particularly to an optical character reader which is modified by applying a white paint correction liquid (hereinafter referred to as correction liquid). The present invention relates to an optical character reader capable of detecting an area.

2. Description of the Related Art In general, an optical character reader (hereinafter referred to as OCR)
) Is a device for optically reading and recognizing characters and symbols written using a writing implement specified in advance on a form, and the writing implement is generally a pencil, a ballpoint pen, a fountain pen, or the like. It is. In the case where the form reads characters written on the insurance contract which requires writing with a writing instrument such as an insurance contract which cannot be easily erased, for example, with a ballpoint pen, the OCR according to the prior art requires a writer. However, the correction contents are corrected using the correction liquid, and the correction characters on the correction liquid are read even if the form is inappropriate as a contract document.
References describing techniques related to the above-mentioned prior art include a technique for reading characters corrected by a correction liquid described in Japanese Patent Application Laid-Open No. 8-279911 without error,
Japanese Unexamined Patent Application Publication No. 334489 discloses a device capable of recognizing characters regardless of the background color.

The OCR according to the prior art described above does not have a function of detecting whether a correction liquid is applied to a form, and can recognize characters even if the form is inappropriate as a contract. There was a defect that there is. Further, the conventional OCR does not have a function of determining whether a writing implement entered in a form is any writing implement, and therefore, it is obligatory to fill in a writing implement that cannot be easily erased like the insurance contract. There was a problem that character recognition was performed even in a form. As described above, the OCR according to the conventional technology has a problem that an originally invalid form cannot be detected, and an inappropriate form must be manually removed. SUMMARY OF THE INVENTION An object of the present invention is to eliminate the above-described disadvantages of the related art, and to provide an optical character reading device that can detect whether a character written on a form has been corrected by a correction liquid. is there. It is another object of the present invention to provide an optical character reading device capable of detecting the type of writing implement written on a form and alerting an OCR user.

In order to achieve the above object, an optical character reading apparatus according to the present invention obtains a difference between a reflected image due to light reflected from a surface of a form and a transmitted image transmitted through the form. The first feature is that it is determined that the correction liquid is applied when the ratio of the black image in the difference image is larger than a predetermined value. Further, the optical character reading device according to the present invention detects the reflectance of an unwritten portion of a form and the reflectance of a written portion, and detects the reflectance of the written portion is higher than the reflectance of the unwritten portion. A second feature is that it is determined that the correction liquid is applied. Further, the optical character reading device according to the present invention irradiates a plurality of wavelengths of blue, green, and red to unfilled portions of the form, stores the reflectance of the reflected light at the white level of each wavelength, and stores Is irradiated with the light of the plurality of wavelengths to detect the reflectance at the black level of each wavelength, and the reflectance at the black level is 10% of the reflectance of the white level at each wavelength at the white level. In the case where the reflectance is within the predetermined range as described above, it is determined that entry with a pencil is determined in the third step.
The feature of.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an optical character reading device according to an embodiment of the present invention will be described in detail with reference to the drawings.
1 is a diagram showing a schematic configuration of an optical character reading device according to the present embodiment, FIG. 2 is a diagram showing an example of a form corrected by a correction liquid, and FIG. 3 is an optically scanned form shown in FIG. FIG. 4 is a view for explaining a correction fluid detecting procedure according to the present embodiment, and FIG. 5 is a view showing a schematic configuration of an optical character reading apparatus according to another embodiment of the present invention. 6 is a diagram showing another example of the form corrected by the correction liquid, FIG. 7 is a diagram showing a waveform when the form shown in FIG. 6 is optically scanned,
FIG. 8 is a diagram for explaining the reflectance of a pencil and a ballpoint pen. The optical character reading device according to the present embodiment shown in FIG. 1 detects a light source 1 that irradiates light for character detection onto the surface of a form 3 on which characters are written, and detects light reflected by the form 3 from the light source 1. Sensor 4 for converting and outputting an analog signal detected by the image sensor 4 to a digital signal, a binarizing circuit 8 for binarizing the converted digital signal, 3, a light source 2 for irradiating light to the back surface, an image sensor 5 for detecting light passing through the form 3 from the light source 2, and an A / D converter for converting an analog signal detected by the image sensor 5 into a digital signal. 7, a binarization circuit 9 for binarizing the converted digital signal, and a difference comparison circuit 10 for outputting the output difference between the binarization circuits 9 and 8. A character written on the form 3 is detected from the (image) signal thus obtained, and as described later, an output difference between the binarization circuits 8 and 9 is obtained by a difference comparison circuit 10, thereby detecting the presence or absence of the correction liquid and issuing a warning. It is configured to do so. Here, as shown in FIG. 2, for example, the form 3 corrected with the correction liquid is coated with a circular correction liquid 11 in an entry frame, and a character 12 (the character itself is “1”) is placed on the correction liquid 11. It shall be completed. By the way, the optical character reading device configured as described above converts the form 3 into the sensors 4 and 5.
, And light from each of the light sources 1 and 2 is detected by the image sensors 4 and 5. The light reflected from the form 3 detected by the image sensor 4 is, for example, as shown in FIG.
When optically scanning at the position -A ', FIG.
As shown in (a), the entry frame and the character portion are detected as a concave waveform 13 lower than the threshold. Similarly, the light transmitted through the form 3 detected by the image sensor 5 is shown in FIG.
As shown in (b), the portion of the entry frame and the portion coated with the correction liquid are detected as a concave waveform 14 lower than the threshold value. Next, the optical character reading apparatus converts the analog signal detected by each of the sensors 4 and 5 into a digital signal by A / D conversion circuits 6 and 7, and binarizes it as an image signal by binarization circuits 8 and 9. , The difference comparison circuit 10. In the binarized image signal, the reflection image by the image sensor 4 is an image in which the character “1” is filled in black as shown in an image 15 in FIG. The output of the correction liquid 11 is small because the portion of the correction liquid 11 is thicker than the form 3 and the transmittance is reduced.
As shown in an image 16 in FIG. 4B, a circle portion in which the correction liquid is applied in the frame becomes an image in which the correction liquid is painted black. Next, in the present apparatus, the comparison circuit 10 compares the image 15 of FIG.
The difference image 17 shown in FIG. 4C is obtained by calculating the difference between the images 16 shown in FIG. 4B. In the difference image 17, the characters and the character frame in both the reflected light image 15 and the transmitted light image 16 disappear, and only the correction liquid remains, so that the periphery of the character remains black and the character part is white. Become an image. The optical character reading device according to the present embodiment uses the difference image 17
Is calculated, and if the area of the black region is larger than a predetermined value, it is determined that the report is a form coated with the correction liquid, and a warning is issued to the user. As an example of a method of issuing this warning, when the correction liquid is detected in the OCR operation computer, a warning is immediately displayed on a display or the like, and the image of the corresponding portion and the position on the form are displayed, or
Adding and storing the correction liquid presence / absence information and the character peripheral image in association with the read character data or the like for later identification;
Alternatively, it is conceivable that a warning character or the like is printed on a form at the time of reading by a printing device so that sorting can be easily performed. Thus, the optical character reading device according to the present embodiment is
The difference between the image due to the reflected light from the surface of the form and the image transmitted through the form is calculated, and when the ratio of the black image in the difference image is larger than a predetermined value, it is determined that the correction liquid is applied. Thereby, a form inappropriate as a contract or the like can be easily detected. Next, an optical reader according to another embodiment of the present invention for detecting whether a writing implement described in a form is a pencil or a ballpoint pen will be described. The configuration of the device according to the present embodiment is the same as the configuration shown in FIG. 1, except that the light emission wavelength of the light source 1 is a visible light of about 400 nm to 660 nm, and the light emission wavelength of the light source 2 is a near infrared ray of about 740 nm. It is. The optical reading device according to the present embodiment utilizes the fact that a general ballpoint pen has an absorption rate of near-infrared ray of 20% or more and a pencil of about 70%. The output is binarized with an appropriate threshold, and the difference between the output images is obtained. This difference image has a characteristic that the image remains black only when using a pencil,
The optical reading apparatus determines that the image is a pencil when the image remains black, and otherwise determines that the image is a writing implement such as a ballpoint pen. The apparatus can also issue a warning if only one of them is permitted to be used. That is, the optical character reading device according to the present embodiment is
The reflectance of an unwritten portion of the form and the reflectance of a character written portion are detected, and if the reflectance of the written portion is higher than the reflectance of the unwritten portion, it is determined that the correction liquid has been applied.
Next, an optical character reading device according to another embodiment of the present invention will be described with reference to FIG. The optical character reading apparatus shown in FIG. 5 includes a light source 18 for irradiating a form 19 with light, rotating mirrors 20 and 21 for reflecting light reflected from the light source 18 on the surface of the form 19, And a one-dimensional color image sensor 23 that receives the reflected light of the light from the light source 18 via the lens 22, and scans the reflected light reflected by the form 19 from the light source 18 on the form by rotating the turning mirror 20. It is configured to form an image on a one-dimensional color image sensor 23 via a lens 22. Here, the form 19 is composed of a white paint correction liquid 24 as shown in FIG.
Are applied in a circular shape, and the reflectance of the form 19 is about 75%. When the optical character reading apparatus configured as described above scans the BB 'portion on the form 19 with the one-dimensional image sensor 23, the sensor output waveform becomes a waveform 25 as shown in FIG. This waveform 25 is an output waveform of the sensor, and the portion where the correction liquid is applied on the form 19 shows a reflectance of about 80%, and since the reflectance of the form is higher than 75%, the output of the sensor is white. It has a shape that protrudes in the direction. In the present embodiment, the presence or absence of the correction liquid is detected by comparing the difference between the reflectance of the form and the reflectance of the correction liquid. More specifically, the optical character reading apparatus first detects the leading end of the form 19, stores the white level (sensor output) of the reflectance in a memory or the like,
Thereafter, the form 19 is conveyed to the place where the correction liquid 24 is applied (character entry frame portion), and the difference between the increased sensor output and the previously stored form level is calculated. The waveform of this differential signal is a differential waveform 26 as shown in FIG.
Appears like When the difference is larger than a predetermined threshold value, the apparatus according to the present embodiment determines that the correction has been performed by the correction liquid, and displays a warning on, for example, a display of an operation computer so that the OCR user can understand. Thereafter, the form becomes an error form and is discharged for confirmation, so that the form can be visually confirmed. This device can use this low level as a reference if low-reflection paper is used for the ground color such as recycled paper with little bleaching. Can be detected. In the optical character reading apparatus according to the present invention, in the configuration shown in FIG. 5, the light source 18 emits only red light centered at a specific wavelength, for example, 660 nm, or the light selected by a filter using a white light source. The correction liquid looks white to the human eye by printing the character writing area of the paper so that the reflectance is high for that wavelength and the other wavelengths are low, that is, a dropout color. Since the surrounding character entry area looks red, the operator can easily understand that the correction liquid has been applied. Further, in the optical character reading device according to the present embodiment, since the reflectance of the output waveform of the sensor is substantially the same as that of the form, the image in the case of binarization is read because the color components around the character are removed. It is possible to obtain the best clean image for At this time, the light source was set to about 740 nm infrared rays and 400 n.
By emitting visible light of m to 600 nm, installing multiple image sensors, placing an infrared filter and a visible light filter in front of the sensors, and reading characters written with a ballpoint pen and pencil with this optical system,
It can detect whether it is a ballpoint pen or a pencil. In this optical character reading apparatus, the reflectance of a ball-point pen is as high as about 80% with respect to the infrared ray of the wavelength, and the form paper shows the same degree. In the case of a pencil, the reflectance of the infrared ray is also 20%.
The difference is compared, and a ball-point pen or a pencil is detected by comparing the difference. That is, when characters appear in the output image of the visible light sensor and no characters appear in the output image of the infrared light sensor, the writing implement is a ballpoint pen, and the output image of the visible light sensor and the output image of the infrared light sensor are displayed. If a character appears in both of them, it can be determined that it is a pencil. An optical character reading device according to still another embodiment will be described. First, as shown in FIG. 8, the reflectance of a pencil and a ball-point pen is almost constant regardless of the wavelength of the pencil, whereas the reflectance of a ball-point pen at a wavelength up to a little over 600 nm is 10% or less. It is known that the reflectance increases with respect to wavelengths longer than that. That is, it is known that the reflectivity of a pencil is higher than that of a ball-point pen and the reflectivity from a wavelength of 400 nm to 700 nm is substantially constant, and the reflectivity of a ball-point pen increases from around 600 nm. The optical character reading apparatus according to this embodiment uses a light source of three colors of red, green and blue and a color sensor, and first sets a white level to red (wavelength) for a part where no character is written (white) such as an end of a reference form. 660 nm),
Each color of green (wavelength 540 nm) and blue (wavelength 480 nm) is stored, and when a black level exceeding a threshold set in a preset character entry area is input, the ratio of the black level to the stored white level Is calculated for each of the red, green and blue colors, and the values are stored. Next, when the reading of the character entry area is completed, the apparatus takes a histogram of the reflectance of each color previously stored for each color, compares the highest value with each color, and compares each color. Is not much different for each color (within ± several% like pencil in Fig. 8)
In this case, it is determined to be a pencil, and otherwise, it is determined to be a ballpoint pen. That is, the optical character reading device according to the present embodiment can determine the pencil and the ballpoint pen by using the difference in the reflectance of the pencil and the ballpoint pen with respect to the wavelength. In the present embodiment, a light source of three colors is used. However, a configuration may be adopted in which a white light source emits multiple colors using filters of each color. That is, the optical character reading device according to the present invention irradiates a plurality of wavelengths of blue-green-red light to the blank portion of the form, stores the reflectance of the reflected light at the white level of each wavelength, and stores Is irradiated with the light of the plurality of wavelengths to detect the reflectance at the black level of each wavelength, and the reflectance at the black level is 10% of the reflectance of the white level at each wavelength at the white level. If the reflectance is within the predetermined range, it is determined that the entry is made by a pencil. As described above, the optical character reading device according to the present embodiment determines whether the character written on the form is a pencil or a ballpoint pen by observing the change in the reflectance of each wavelength as described above, and permits the form to be read. A warning can be issued to the user in the case of a writing instrument other than the writing instrument used.

As described above, the optical character reading apparatus according to the present invention obtains the difference between the reflected image due to the reflected light from the surface of the form and the transmitted image transmitted through the form, and calculates the black image in the difference image. If the ratio occupied by is larger than the predetermined value, it is determined that the correction liquid is applied.The reflectance of the unfilled portion of the form and the reflectance of the character written portion are detected, and the reflectance of the character written portion is determined. By determining that the correction liquid has been applied when the reflectance is higher than the unfilled portion, it is possible to easily find a form to which the correction liquid has been applied. Further, the optical character reading device according to the present invention irradiates a plurality of wavelengths of blue, green, and red to unfilled portions of the form, stores the reflectance of the reflected light at the white level of each wavelength, and stores Is irradiated with the light of the plurality of wavelengths to detect the reflectance at the black level of each wavelength, and the reflectance at the black level is 10% of the reflectance of the white level at each wavelength at the white level. In the case where the reflectance is within the predetermined range, it is possible to determine the writing implement described in the form by determining that the entry is made with a pencil. Therefore, the danger of inputting incorrect data can be prevented.

[Brief description of the drawings]

FIG. 1 is a diagram showing a schematic configuration of an optical character reading device according to an embodiment of the present invention.

FIG. 2 is a diagram showing an example of a form corrected by a correction liquid.

FIG. 3 is a diagram showing a waveform when the form shown in FIG. 2 is optically scanned.

FIG. 4 is a view for explaining a correction fluid detection procedure according to the embodiment;

FIG. 5 is a diagram showing a schematic configuration of an optical character reading device according to another embodiment of the present invention.

FIG. 6 is a view showing another example of a form corrected by the correction liquid.

FIG. 7 is a diagram showing a waveform when the form shown in FIG. 6 is optically scanned.

FIG. 8 is a diagram for explaining the reflectance of a pencil and a ballpoint pen.

[Explanation of symbols]

1 ... light source (for reflection), 2 ... light source (for transmission), 3 ... form,
4, 5 ... image sensor, 6, 7 ... A / D converter,
8, 9 ... binarization processing section, 10 ... difference comparison section, 11 ... applied correction liquid, 12 ... characters written on the correction liquid,
13 ... output waveform of sensor 9, 14 ... output waveform of sensor 10, 15 ... output image of sensor 9, 16 ... output image of sensor 10, 17 ... image 15 and image 17 Difference image, 18 light source, 19 report, 20, 21 mirror, 22 lens, 23 one-dimensional image sensor, 24 applied correction fluid, 25 .. The output waveform of the sensor 23, 26.

Claims (3)

[Claims] A difference between a reflected image due to light reflected from the surface of a form and a transmitted image transmitted through the form is calculated, and when a ratio of a black image in the difference image is larger than a predetermined value, the correction liquid is used. An optical character reading device, wherein it is determined that the liquid is applied. 2. The reflectance of an unwritten portion of a form and the reflectance of a character-written portion are detected, and when the reflectance of the character-written portion is higher than the reflectance of the unwritten portion, a correction liquid is applied. An optical character reading device, characterized in that it is determined that the character is present. 3. An unfilled portion of the form is irradiated with light of a plurality of wavelengths of blue, green and red, and the reflectance of reflected light at a white level of each wavelength is stored. And the reflectance at the black level of each wavelength is detected, and the reflectance at the black level is 10% or more of the reflectance of the white level at each wavelength at the white level and within a predetermined range. An optical character reading device, which determines that writing is performed with a pencil in the case of reflectance.