JPS5887667A - Cutout system for manuscript information - Google Patents

Abstract

PURPOSE:To eliminate the need for print of cutout mark, by providing two sensors having different response waveform, and using an ink to which only one sensor senses, for the part at the inside of a manuscript frame on a document, format information and manuscript information. CONSTITUTION:A reading signal of format information and manuscript information read at a sensor 20 is binary-coded 22 and given to an AND circuit. Further, a signal in the manuscript frame by a sensor 21 is binary-coded 23 and given to address detection circuits 28, 29 and then to the circuit 25 via a delay circuit 24, which adjusts the time for position shift of the sensors 20, 21 toward the sub-scanning direction. Then, only when the inside of the manuscript frame is scanned by the sensor 20, the output of the circuit 24 is 1 and in other cases, it is 0. Thus, the manuscript information within the manuscript frame only out of the information read by the sensor 20 passes through the circuit 25 and is stored in a memory 26. A character recognition circuit 27 reads out the manuscript information from the memory 26 and performs character recognition processing. The result of recognition is outputted from an output device 33.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention provides a system for optically reading information on a form, such as an optical character reader, which separates and cuts out information handwritten on a form from other format information. Regarding improvements.

For example, in a conventional optical character reading device, as shown in FIG. 1, by detecting a cutting mark printed near a handwriting frame 1 on a form, A method was used to recognize the width in the sub-scanning direction and extract handwritten information from the scanned information.

However, this method prints marks on the form that are not format information or handwritten information but only for the purpose of cutting out, so the appearance of the form tends to be unnatural, and improvements are desired.

Therefore, an object of the present invention is to provide a handwritten information extraction method that solves the above-mentioned problems.

According to the present invention, a device for optically reading forms is provided with first and second sensors (scanners) having different sensitive wavelengths. On the other hand, the inside of the handwritten frame on the form read by this device is printed with ink that only the second sensor is sensitive to, and the printing of format information outside the handwritten frame and the entry of handwritten information inside the handwritten frame are performed by the first sensor. This is done using ink that only the sensor is sensitive to. Then, when the reading device reads the form, handwritten information within the handwriting frame is extracted from the information read by the first sensor based on the information read by the second sensor.

Hereinafter, the present invention will be specifically explained with reference to the drawings.

An example of a form to which the present invention is applied is schematically shown in FIG. Format information outside the handwriting frame 11 on this form 10 (ABC...
) is printed with black ink, for example. The handwritten information (that person...) is written in the handwritten frame 11 using black ink, for example. Further, the inside of the handwritten frame 11 (the shaded area) is filled and printed with, for example, fluorescent yellow ink.

According to the present invention, on the side of the device that optically reads such forms, there is provided a first sensor (scanner) having sensitive wavelength characteristics that is sensitive to black ink but not substantially sensitive to fluorescent yellow ink. A second sensor (scanner) is provided which has a sensitive wavelength characteristic that is sensitive to fluorescent yellow ink but not substantially sensitive to black ink. And the ledger is
These two sensors read simultaneously. As the first sensor, for example, a known blue-filtered CCI (charge coupled device) scanner is used. Further, as the second sensor, for example, a known Yellow F/L/9H CCD
A scanner is used.

Therefore, when the form 10 shown in FIG. 1 is read by such a reading device, the image read by the first sensor will be as shown in FIG. 3, and the image read by the second sensor will be as shown in FIG. 4. . Therefore, the handwritten frame can be identified from the information read by the second sensor, and the handwritten information can be easily extracted from the information read by the first sensor.

In the example of FIG. 2, the handwritten frame is printed in one line, but it may be printed separately for each character as shown in FIG. This makes it easier to cut out individual handwritten characters.

FIG. 6 shows an example of a reading device for executing the above-mentioned handwritten information extraction and will be described.

The condyles 21 are first and second sensors that scan the form.

Both sensors are at the same position in the main scanning direction, but the second sensor 21 is located a certain distance further than the first sensor 20 in the sub-scanning direction.

The read information signal from the first sensor is binarized by a binarization circuit n and sent to an AND circuit 6. Note that depending on the nature of the character recognition algorithm of the character recognition circuit r, the read information signal from the first sensor may be multivalued quantized.

The inkstone removing information signal from the second sensor 21 is binarized by the binarization circuit 23 and sent to the address detection circuit 29, while being sent to the AND circuit δ via the delay circuit 24. The delay circuit 24 connects the first and second sensors.

21 is adjusted to account for the positional deviation in the sub-scanning direction, and the information of the same pixel point on the form read by the first and second sensors 20 and 21 is made to reach the AND circuit 5 at the same time.

Therefore, the output of the delay circuit U is "1" only during the period when the inside of the handwriting frame is scanned by the first sensor, and the output of the delay circuit 24 is 10 otherwise. Of the read information, only the handwritten information within the handwriting frame passes through the AND circuit 5 and is stored in the memory 26.
is stored in

The address detection circuit side 29 receives the read information signal of the second sensor 21 from the binarization circuit 23 and detects the main scanning direction address X and the scanning direction address Y of the four corners of the handwritten frame on the form. do. This address (X, Y) is stored in the latch (9) and read out to the system controller 32 as required.

The character recognition circuit γ reads handwritten information from the memory 26, and reads out the handwritten information! Executes identification processing. Recognition result F output device 3
3 (for example, a printer).

The system controller 32 not only performs integrated control of each part, but also refers to the latch recess as necessary to perform processing such as editing of recognition results.

As described above, the present invention requires printing a conventional cutout mark near the handwriting frame, so mistakes such as writing handwritten information in the wrong position may occur. It becomes difficult. Furthermore, when extracting handwritten information by recognizing the position of the handwritten frame from the cutout mark printed outside the handwritten frame, cutting errors are likely to occur due to effects such as skew of the form, but the present invention detects the handwritten frame itself. Therefore, such cutting errors can be reliably prevented.

[Brief explanation of the drawing]

Fig. 1 is a diagram for explaining the conventional handwritten information extraction method, Fig. 2 is a diagram showing an example of a form to which the present invention is applied, and Figs. 3 and 4 are diagrams showing two sensors with different responsive wavelengths. FIG. 1 is a diagram showing an image read from the form, FIG. 5 is a diagram for explaining a modified example of the form, and FIG. 6 is a block diagram showing an example of a reading device to which the present invention is applied. DESCRIPTION OF SYMBOLS 10... Form, 11... Handwriting frame, Addition... First sensor, 21... Second sensor, n, 23... Binarization circuit, 24... Delay circuit, 5...・Logic product circuit, 26・
...Memory, 27...Character recognition circuit, Path, 29...
Address detection circuit, 30...Latch, 32...System controller, 33...Output device. Agent Patent Attorney Makoto Suzuki・-１＼・１゜＼０）-１Figure 1Figure 2Figure 3nFigure 4Figure 5Figure 6Figure 11 1 page

Claims (1)

[Claims] 1. A device that optically reads information on a form with a handwritten portion is provided with first and second sensors with different responsive wavelengths,
The inside of the handwriting frame on the form is printed with ink that only the second sensor is sensitive to, and the information outside the handwriting frame, such as format information, is printed with the ink that only the first sensor is sensitive to, and the handwriting is The handwritten information in the frame is written with ink that only the first sensor is sensitive to, and when the device reads the form, the handwritten information is written from the information read by the first sensor based on the information read by the second sensor. A handwritten information extraction method characterized by cutting out handwritten information within a frame.