JPS6362025B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a numeric reading method for reading numeric information with guidelines written on an input form using a facsimile transmitter.

Conventionally, methods have been known to use facsimile transmitters to read numbers with guide lines written on input forms, but these methods read the numbers in each segment (detection area) of the ``Japanese'' character or similar number entry frame. The method involved counting the number of black pixels and comparing it with a certain threshold to determine the presence or absence of a written line segment, and comparing it with numerical writing rules to read numerical information. It has the drawback that incorrect connections are likely to occur due to black line noise caused by variations in the line width of each character, blurred or protruding lines, and instantaneous interruption of the line.

In order to eliminate these drawbacks, the present invention detects the length of the line segment in each segment of the character ``Japanese'' character or a numeric entry frame similar to it. , categorized into blurred line segments and correct reading line segments. Numerical information with unknown line segments or with two or more blurred line segments is rejected, and numerical information with one blurred line segment is read correctly as a blurred line segment. It is characterized by being treated as a line segment, and its purpose is to reduce misreading and reject characters caused by blurred input line segments, input line segments that protrude from the number entry frame, and black line noise due to momentary line interruptions, etc. .

FIG. 1 is a diagram showing the relationship between an example of an input form used in the present invention and scanning lines when the input form is input into a facsimile transmitter, in which numeral 1 indicates an input form for entering numerical information, 2 is a reference mark indicating the left edge of the input paper 1, and 3 and 4 are tilt marks for detecting the inclination of the input paper 1 with respect to the main scanning direction of the facsimile transmitter (the distance from the reference mark 2 is 2 mm and 152 mm, respectively). , 5 is a timing mark indicating the position of numerical information in the main scanning direction; 6, 7
are scanning designation marks indicating the location of numerical information in the sub-scanning direction (distance from reference mark 2 is 2).
mm, 152 mm), 8 is a number entry frame in the shape of a "day" for entering numerical information, l ₁ to _{l 5} are part of the scanning line when input form 1 is scanned with a facsimile transmitter. This is what is shown. Note that reference mark 2 (width 1 mm), tilt marks 3 and 4 (width 2 mm), timing mark 5 (width 2 mm, height 2 mm), and scan designation marks 6 and 7 (width 2 mm) are in colors that can be detected by the facsimile transmitter. In addition, the number entry frame 8 is written on the input paper 1 in a color that cannot be detected by the facsimile transmitter.

Figure 2 shows input paper 1 being sent to a facsimile transmitter (main scanning density: 8 dots/mm, sub-scanning density: 8 lines/mm).
1 is a diagram showing an image signal of one scanning line when scanning is performed by 10, 10 is an image signal corresponding to scanning line _l1 ,
1 is an image signal corresponding to the reference mark 2, and 12 is an image signal corresponding to the inclined mark 3. Next, a method for correcting the reference position will be described below with reference to FIG. First, the image signal 10 is examined sequentially from the left side, and the image signal 11, that is, the reference mark 2, is detected by detecting consecutive locations where the number of black pixels is 6 or more and 10 or less. Then, the reference position can be corrected by extracting an image signal of a certain length from the reference mark 2 onwards.

FIG. 3 is a diagram showing a tilt detection method when input paper 1 is input to a facsimile transmitter.
The case where the inclination of the input paper 1 is the same as that in FIG. 1 is shown. Reference numerals 20 to 24 are image signals for which reference positions of scanning lines _l1 to _l5 have been corrected. First, the image signal 20 to 2
16 dots (2mm x 8 dots/mm) from the left edge of 4.
In the vicinity of 1216 dots (152 mm x 8 dots/mm), check for the presence or absence of consecutive locations where the number of black pixels is 14 or more and 18 or less. Then, the upper end of the inclined mark 3 can be detected in the image signal 20, and the upper end of the inclined mark 4 can be detected in the image signal 24. As a result, it is found that the inclination of the input paper 1 with respect to the main scanning direction of the facsimile transmitter is four scanning lines downward to the right.

FIG. 4 is a diagram showing a method for correcting the inclination of numerical information written on the input paper 1, and shows a case where the inclination of the input paper 1 is the same as that in FIG. 30 to 34 are image signals subjected to reference position correction, and 35 to 39 are parts obtained by dividing each of the image signals 30 to 34 into five. Since the input paper 1 has four scanning lines descending to the right, each of the image signals 30 to 34 of the five scanning lines is first divided into five parts, and the image signals 35 to 39 are sequentially extracted.
By rearranging them on one scanning line, an image signal with tilt correction can be created. Thereafter, the scanning lines are sequentially updated and the same procedure is performed. As described above, when the interval between the tilt marks 3 and 4 is n scanning lines, an image signal subjected to tilt correction can be obtained by dividing (n+1) scanning lines into (n+1) and rearranging them.

FIG. 5 is a diagram showing a method of detecting the timing mark 5 from the image signal that has undergone the reference position correction and tilt correction by the method described above, and a shows the timing mark 5 that has undergone the reference position correction and tilt correction. 5 and tilt mark 3, b is timing mark 5
c is a diagram in which the number of black pixels of the scanning line where exists is added in the sub-scanning direction, c is a diagram in which black pixels other than timing mark 5 are removed, 40 is noise, and 41 is the threshold for the number of black pixels in the sub-scanning direction ( In this example, it is 14 dots). First, the tilt mark 3 is detected from the tilt-corrected image signal using the method explained in FIG. become. Then, as shown in FIG. b, the threshold value 41 in the sub-scanning direction is
By slicing (14 dots) and removing consecutive locations where the number of black pixels in the main scanning direction is 4 or less, the noise 40 can be removed and the correct position of the timing mark 5 can be detected as shown in Figure c.

FIG. 6 is a diagram illustrating a method for detecting the presence or absence of line segments written in seven segments of the number entry frame 8 that have undergone reference position correction and tilt correction, 51,
52 is a specific timing mark, 61 to 63 are specific scanning designation marks, and a to g are segments (hatched area: 2 mm x 2 mm).

First, start from the left end of the image signal that has undergone tilt correction.
By examining the vicinity of 1216 dots (152 mm x 8 dots/mm) and detecting consecutive locations where the number of black pixels is 14 or more and 20 or less, the existence of the scan designation mark 7, that is, the number entry frame 8 in the sub-scanning direction, is detected. Detect location. Next, by examining the vicinity of 16 dots (2 mm x 8 dots/mm) from the left end of the detected image signal of the scan designation mark 7, and detecting consecutive locations where black pixels are 14 or more and 20 or less, Scan designation marks 61 to 63 are detected. Among the detected image signals of the scan designation mark 7, the presence or absence of the scan designation marks 61 to 63 and the specific timing mark 5 detected in the explanation of FIG.
1 and 52, segments a to g shown in FIG. 6 are detected.

Figure 7 shows segments a to g detected in Figure 6.
Figure a is a diagram showing a method for detecting a horizontal line segment in segment a, Figure b is a diagram showing a method for detecting a vertical line segment in segment b, and 70 and 71 are The input line segment, l _h is the length of the input line segment 70 mapped in the main scanning line direction, l _v is the input line segment 71
This is the length mapped in the sub-scanning direction. First, figure a
The horizontal line segment detection method shown in FIG. The scan designation mark 61 is detected by the method explained in FIG. 6, and the logical sum of the image signals in which the scan designation mark 61 exists is taken in the sub-scanning direction. Then, by counting the black pixel rows existing between the specific timing marks 51 and 52 detected in FIG. 5, the length l _h of the drawn line segment is detected.

Next, a vertical line segment detection method shown in FIG. b will be explained. Scanning designation mark 7 is scanned by the method explained in Fig. 6.
, and an image signal between scanning designation marks 61 and 62 is detected. Then, among the detected image signals, the image signals at the position indicated by the specific timing mark 51 detected in FIG. By counting the line segment length l _v
Detect. The length of the drawn line segment in segments c to g is detected in the same manner as above.

FIG. 8 shows an embodiment of the present invention, in which 80 is a facsimile transmitter that scans the input paper 1, 81 is an interface circuit that extracts an image signal from the facsimile signal received from the facsimile transmitter 80, and 82
83 is a reference position correction circuit that detects the reference mark 2 in the image signal in each scanning line and corrects the reference position by extracting the image signals after the reference mark 2; The tilt marks 3 and 4 are detected from the signal, and the facsimile transmitter 8
An inclination detection circuit 84 detects the inclination of the input paper 1 with respect to the main scanning direction of 0, and 84 writes an image signal to a buffer memory or reads an image signal from the buffer memory according to the inclination value detected by the inclination detection circuit. 85 is a buffer memory (for the image signal of the maximum tilt scanning line) that temporarily stores image signals necessary for tilt correction; 86 is an image signal that has undergone reference position correction and tilt correction; 87 is a numeric area detection circuit that detects the scan designation mark 7 from the image signal that has been subjected to tilt correction; 88 is the scan designation circuit that detects the scan designation mark 7 from the image signal in which the scan designation mark 7 is present; Segment detection circuit for detecting mark 6, 89 is segment a, d, g
90 is a vertical line segment detection circuit that detects the length of the drawn line segments of segments b, c, e, and f; 91 is a line segment extending beyond the drawn line segment length, an unknown line; 92 is a numeric judgment circuit that judges numbers based on combinations of the lengths of written line segments classified by the written line segment identification circuit; 93; Reference numeral 94 denotes an interface circuit that encodes the numbers determined by the number determination circuit, and 94 an information processing device.

Next, the operation shown in FIG. 8 will be explained. However, in this description, it is assumed that the input paper 1 is inputted tilted downward to the right by four scanning lines with respect to the main scanning direction of the facsimile transmitter 80.

First, the facsimile transmitter 80 sends input paper 1
to send facsimile signals. The interface circuit 81 receives the facsimile signal,
Image signals are extracted in units of one scanning line. The reference position correction circuit 82 receives the image signal from the interface circuit 81, detects the reference mark 2 by the method explained in FIG. 2, and extracts the image signals after the reference mark 2. Through this operation, the left and right positional deviation of the input paper 1 with respect to the facsimile transmitter 80 is corrected.

Next, the tilt detection circuit 83 receives the position-corrected image signal and detects the tilt marks 3 and 4 by the method explained in FIG. Check the scan line difference between and 4. As a result, it is detected that the input paper 1 has an inclination of four scanning lines downward to the right with respect to the main scanning direction of the facsimile transmitter 80. Then, the slope correction circuit 84 is provided with the means necessary for slope correction (in this explanation, since the slope is 4 scanning lines downward to the right, each image signal of 5 scanning lines is divided into 5 and rearranged as shown in FIG. 4). Instruct.

The tilt correction circuit 84 loads the image signal for which the reference position has been corrected in accordance with the instruction from the tilt detection circuit 83 into the buffer memory 85, divides the image signal of the 5 scanning lines into 5 parts by the method explained in FIG. 4, and sequentially divides the image signal into 5 parts. Tilt correction was performed by taking out and rearranging.1
Create a scanning line image signal. This operation corrects the inclination of the input paper 1 with respect to the facsimile transmitter 80.

The timing mark detection circuit 86 takes in the image signal in which the tilt mark 3 is present among the image signals subjected to the tilt correction, and adds the number of black pixels in the sub-scanning direction by the method explained in FIG. 5 to determine the number of black pixels. 14 or more dots are considered to have black pixels. Then, by detecting a location where the number of black pixels is 14 or more consecutive in the main scanning direction, the position of the timing mark 5 in the main scanning direction is detected, and the horizontal line detection circuit 89
and notifies the vertical line segment detection circuit 90 of the position of the timing mark 5.

The numerical area detection circuit 87 checks the presence of the scan designation mark 7 on the image signal subjected to the tilt correction using the method explained in FIG. 6, and outputs the image signal in which the scan designation mark 7 exists to the segment detection circuit 88. do. When the segment detection circuit 88 receives an image signal in which the scan designation mark 7 is present, it checks whether the scan designation mark 6 is present using the method described in FIG. The image signal in which the scan designation mark 6 exists is output to the horizontal line segment detection circuit 89, and the image signal in which the scan designation mark 6 does not exist is output to the vertical line segment detection circuit 90.

The horizontal line segment detection circuit 89 receives the image signal in which the scanning designation mark 6 is present, and calculates the logical sum of the image signal 18 scanning line segments (2 mm x 8 lines/mm) in the sub-scanning direction using the method explained in FIG. 7a. Take it. The horizontal line lengths of segments a, d, and g are detected by counting black pixels at positions where timing mark 5 sent in advance from timing mark detection circuit 86 does not exist. The vertical line segment detection circuit 90 detects the image signal 18 scanning line segments (2 mm x 8
books/mm), and the logical sum of the sections in which the timing mark 5 previously sent from the timing mark detection circuit 86 is located using the method explained in FIG. 7 is calculated in the main scanning direction, and the black pixel at the position of the logical By counting the vertical line lengths of segments b, c, e, and f are detected.

The input line segment identification circuit 91 is a horizontal line segment detection circuit 89.
The entered line segments of each segment sent from the vertical line segment detection circuit 90 are classified into protruding line segments, unknown line segments, blurred line segments, and correct reading line segments according to the length. The numerical judgment circuit 92 examines the type of line segment sent from the input line segment identification circuit 91, and determines that there are no correct reading segments for protruding line segments, numerical information for unknown line segments, and two blurred line segments. Numerical information that exists above is determined to be an undefined character, and a faded line segment of numerical information that has one blurred line segment is considered to have a correctly read line segment. Numerical information that is not determined to be an undefined character is compared with the combinations of the presence and absence of normal reading line segments in the table shown in FIG. The interface circuit 93 outputs the encoded numerical information to the information processing device 94. Through the above operations, the numerical information written on the input form 1 can be input from the facsimile transmitter 80 to the information processing device 94.

Note that FIG. 7 is only one embodiment of the present invention; for example, a transmission path may be interposed between the facsimile transmitter 80 and the interface circuit 81, or a similar scanning mechanism may be used instead of the facsimile transmitter 80. You may install a device with Furthermore, the number entry rules shown in FIG. 9 are merely examples, and it is also possible to express and read simple symbols (for example, ", 〓,", etc.) in the number entry frame.

As explained above, this method uses input paper on which the reference mark, tilt mark, timing mark, and scan designation mark are written in a color that can be detected by the facsimile transmitter, and the number entry frame is written in a color that is not detectable by the facsimile transmitter. This has the advantage that the reference position and inclination of the input paper can be corrected, and the position of the numeral frame shaped like the Japanese character ``日'' or similar thereto can be accurately detected. In addition, the line segment lengths in the seven segments of the number entry frame are detected, and depending on the line segment lengths, line segments are extended, unknown line segments, and unknown line segments are detected.
In order to judge numbers by classifying them into blurred line segments and correct reading line segments, this method counts the black pixels in each segment, uses a threshold to determine the presence or absence of a line segment, and compares it with the number entry rules to read the number. This method has the advantage of reducing misreadings and rejects caused by the use of writing instruments with different line widths, variations in line width within one character, blurred line lines, and the like. In addition, because of the method of detecting the line segment length, segments b, c, e,
f also has the advantage of being less susceptible to black line noise (horizontal line segments) caused by instantaneous line interruptions.

[Brief explanation of drawings]

Fig. 1 is a diagram showing an example of the input form used in the present invention and the relationship between the scanning lines when the input form is input to a facsimile transmitter, and Fig. 2 is a view when the input form is scanned by the facsimile transmitter. Figure 3 is a diagram showing the tilt detection method when an input form is input to a facsimile transmitter, and Figure 4 is a diagram showing the tilt correction of numerical information written on the input form. Figure 5 is a diagram showing a method for detecting timing marks from an image signal that has undergone tilt correction. Figure 6 is a diagram that shows how to detect timing marks from an image signal that has undergone tilt correction. Figure 7 is a diagram showing a method for detecting line segments of the seven segments of the number entry frame.
The figure is a block diagram of one embodiment of the present invention, and FIG. 9 is an explanatory diagram showing an example of number entry rules expressed by combinations of the presence and absence of normal reading line segments. In the figure, 1 is input paper, 2 is reference mark, 3, 4
5 is a tilt mark, 5 is a timing mark, 6 and 7 are scanning designation marks, 8 is a number entry frame, and l ₁ to l ₅ are scanning lines.

Claims (1)

[Claims] 1. In the numeric reading method that uses a facsimile transmitter to read numeric information written on an input form,
a reference mark for detecting a positional shift of the input paper in the main scanning direction of the facsimile transmitter; a tilt mark for detecting the inclination of the input paper with respect to the main scanning direction of the facsimile transmitter; and a main character of the numerical information. A timing mark indicating the position of the numerical information in the scanning direction and a scanning designation mark indicating the position of the numerical information in the sub-scanning direction are written in a color that can be detected by the facsimile transmitter,
and the facsimile transmitter uses the input paper on which the character "day" or a numeral frame similar to it is written in a color that cannot be detected by the facsimile transmitter at the position specified by the timing mark and the scan designation mark. Detecting the left and right positional deviation and inclination of the input paper by detecting the reference mark and the tilt mark from a facsimile signal obtained by inputting the input paper to a transmitter;
and a means for correcting the reference position and inclination of the input paper according to the detected left and right positional deviation and inclination of the input paper, and detecting the timing mark and the scanning designation mark. ” or similar number box 7
means for detecting two segments and the length of a drawn line segment in each segment, and classifying the segment into protruding line segments, unknown line segments, blurred line segments, and correct reading line segments according to the length of the drawn line segments; A number reading method using a facsimile transmitter, characterized in that it combines means for reading numerical information written in the number entry frame by comparing combinations of line segments with predetermined number entry rules.