JPH087031A - Character frame detecting device - Google Patents

Abstract

PURPOSE:To improve the detection precision of a character frame even if the printing quality of the character frame is low and the size of the character frame is not standardized by resetting a threshold in accordance with the peak value of a historgram. CONSTITUTION:When the number of peaks exceeding the threshold in the historgram generated in a histogram generation part 3 does not become (the number of frames X2), a threshold setting part 8 lowers the setting value of the threshold if the peak value of the histogram is low, and raises the setting value of the threshold if the peak value is high. Thus, the number of the peaks exceeding the threshold becomes (the number of frames X2). It is desirable that a peak judgement part judges whether the width of the peak of the histogram is within a prescribed range or not, and judges whether the intervals of the peaks are within the prescribed range or not. Thus, the peak coordinate position can be judged to be the character frame or not.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention detects a postal code frame described in a postal matter and, according to the detected position of the postal code frame, characters used in a postal matter sorting / sorting machine or the like for sorting postal matter. The present invention relates to a frame detection device.

[0002]

2. Description of the Related Art Conventionally, as a character frame detecting device used in a mail sorting machine, a device shown in FIG. 9 is known.

FIG. 9 is a block diagram showing the structure of a conventional character box detecting device.

FIG. 10 shows the Y-axis direction (vertical axis direction) of a character frame.
11 is a diagram showing a histogram of FIG. 11, and FIG. 11 is a diagram showing a histogram of the character frame in the X-axis direction (horizontal axis direction).

In FIG. 9, a paper sheet L is an object to be processed by the present apparatus in which a postal code frame is described as a character frame.

The scanning unit 21 is a circuit that optically scans the paper sheet L and outputs image data obtained by photoelectric conversion. The image memory 22 is a memory that stores image data read by the scanning unit 21. The histogram generation unit 23 calculates the projection accumulation data in the Y-axis direction and the X-axis direction from the image data stored in the image memory 22, and obtains the histogram in each of the Y-axis direction and the X-axis direction.

The character frame detecting section 30 includes a frame Y coordinate position detecting section 24, a frame X coordinate position detecting section 25, and a character cutting section 26.
And the frame detection determination unit 27.

The frame Y-coordinate position detection unit 24 detects the Y-coordinate position of the character frame using the histogram in the Y-axis direction output from the histogram generation unit 23. Further, the frame X coordinate position detection unit 25 uses the histogram generation unit 2 within the Y coordinate position range of the character frame obtained from the frame Y coordinate position detection unit 24.
The X-coordinate position of the character frame is detected using the histogram in the X-axis direction obtained from No. 3. The character cutout unit 26 is
The coordinate position of the character frame obtained from the frame Y-coordinate position detection unit 24 and the frame X-coordinate position detection unit 25 is used as cut-out position data, and image data for each character is extracted from the image data stored in the image memory 22. Is cut out. The frame detection determination unit 27 uses the coordinate position of the character frame obtained from the frame Y coordinate position detection unit 24 and the frame X coordinate position detection unit 25 and the character cutout information obtained from the character cutout unit 26 to extract the character frame. The presence / absence of characters and the presence / absence of characters in the character frame are determined.

In the character-frame detecting apparatus configured as described above, the frame Y-coordinate position detecting unit 24, as shown in FIG. 10, has a histogram having a height corresponding to the width of the character frame from the histogram in the Y-axis direction. The peak values A3 and A4 are calculated, and the start coordinate position Y3 in the Y-axis direction of the character frame corresponding to A3 and the end coordinate position Y4 corresponding to A4 are calculated.

Next, the histogram generator 23 generates a histogram in the X-axis direction in the Y coordinate position range of Y3 to Y4. At this time, the histogram in the X-axis direction is shown in FIG.
A height peak corresponding to the vertical width of the character frame is obtained as shown in.

The frame X-coordinate position detection unit 25 sets a predetermined threshold value T on the histogram in order to select a peak based on the character frame, and obtains the number of peaks exceeding the threshold value T.

Then, when the number is (number of frames × 2), the frame detection determination unit 27 determines that a character frame has been detected.

[0013]

However, in the conventional character frame detecting device as described above, the print quality of the character frame is poor such that the frame is printed faintly, or the size of the frame is larger or smaller than the standard. If the size of the character box such as is not standardized, the threshold value T
Since the number of peaks that exceed is not (number of frames x 2),
There is a problem in that the character frame cannot be detected and the character frame detection accuracy decreases.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a character box detecting device which does not deteriorate the detection accuracy of the character box even if the description of the character box has a low print quality or the box size is nonstandard. I am trying.

[0015]

In order to achieve the above-mentioned object, a character box detecting device of the present invention is a projection accumulated data in the vertical and horizontal directions of the character frame from image data of a paper sheet in which the character frame is described. And a peak position is determined from the two histograms generated by the histogram generation unit, and a predetermined threshold value is set for one of the two histograms. In a character frame detection device having a character frame detection unit for detecting the position and presence / absence of the character frame by counting the number of peaks exceeding, the character frame detection unit is configured to detect peaks exceeding the threshold value. It is characterized by having a threshold value setting unit for adjusting the threshold value according to the number.

Further, it is determined whether the width of the peak of the histogram generated in the X-axis direction of the character frame is within the specified range, and whether the interval between the peaks is within the specified range. A peak determination unit may be provided.

Further, a plurality of the character frames may be arranged in a straight line, and the threshold value may be determined with respect to a histogram in the arrangement direction of the character frames. In this case, the paper sheets are postal items. A postal code frame may be described as the character frame.

[0018]

In the character frame detection device configured as described above,
In the histogram generated by the histogram generation unit, if the number of peaks exceeding the threshold does not reach (the number of frames × 2), the threshold setting unit determines the threshold value of the threshold if the peak value of the histogram is low. If the peak value is high, lower the setting value and increase the threshold setting value. As a result, the number of peaks exceeding the threshold value is set to (number of frames × 2).

The peak determining section determines whether or not the width of the peak of the histogram is within the specified range, and whether or not the interval between the peaks is within the specified range. It is possible to determine whether the coordinate position is in the character frame or not.

When a plurality of character frames are arranged on a straight line on a paper sheet, the peak for each character frame can be obtained at once by setting the threshold value for the histogram in the arrangement direction of the character frames. It is possible to detect all character frames at once.

In particular, when the paper sheet is a postal matter, the position of the postal code frame can be detected.

[0022]

Embodiments of the present invention will now be described with reference to the drawings.

(First Embodiment) FIG. 1 is a block diagram showing the configuration of the first embodiment of the character box detection device of the present invention.

In FIG. 1, the paper sheet L is a postal matter in which a postal code frame in which five character frames are arranged in a straight line is described, and is a processing target of the present apparatus.

The scanning unit 1 is a circuit that optically scans the paper sheet L and outputs image data obtained by photoelectric conversion. The image memory 2 stores the image data read by the scanning unit 1. Memory.

The histogram generator 3 calculates projection accumulation data in the Y-axis direction (vertical direction of the character frame) and X-axis direction (horizontal direction of the character frame) from the image data stored in the image memory 2, and the Y-axis is calculated. Generate a histogram in the direction and the X-axis direction.

The character frame detection unit 10 is composed of a frame Y coordinate position detection unit 4, a frame X coordinate position detection unit 5, a character cutout unit 6, a frame detection determination unit 7, and a threshold value setting unit 8. . The frame Y coordinate position detection unit 4 detects the Y coordinate position of the character frame using the histogram in the Y direction output from the histogram generation unit 3, and the frame X coordinate position detection unit 5
The X coordinate position of the character frame is detected using the Y coordinate of the frame obtained from the frame Y coordinate position detection unit 4 and the histogram in the X axis direction obtained from the histogram generation unit 3.

The character slicing unit 6 uses the coordinate position of the frame obtained from the frame Y coordinate position detecting unit 4 and the frame X coordinate position detecting unit 5 as the slicing position data and is stored in the image memory 2. Image data is cut out for each character from the image data.

The frame detection determination unit 7 includes a frame Y coordinate position detection unit 4
Using the frame position information obtained from the frame X-coordinate position detection unit 5 and the character cutout information obtained from the character cutout unit 6, it is determined whether or not there is a character frame and whether or not there is a character in the character frame. are doing.

The threshold value setting unit 8 sets the threshold value T, which is set for determining the presence or absence of a character frame on the histogram obtained by the histogram generation unit 3, to the number of peaks exceeding the threshold value T. The process of adjusting accordingly.

Next, the operation of this embodiment will be described with reference to FIGS.

FIG. 2 is a flow chart showing the procedure of character frame detection by the character frame detection device shown in FIG. 1, and FIG. 3 is a diagram showing a histogram of the character frame in the Y-axis direction,
4 is a diagram showing a histogram in the X-axis direction when the character frame has a standard size, FIG. 5 is a diagram showing a histogram in the X-axis direction when the character frame is faint, and FIG. 6 is a character frame. Is a diagram showing a histogram in the X-axis direction when is smaller than the standard, and FIG. 7 is a diagram showing a histogram in the X-axis direction when the character frame is larger than the standard.

First, the scanning unit 1 optically scans the paper sheet L which is the detection target of the character frame description, and stores the image data obtained by the photoelectric conversion process in the image memory 2.
Here, normally, in a postal code frame or the like, a frame is described in red which is a dropout color, but the light source of the scanning unit 1 uses, for example, a yellow-green fluorescent lamp that does not use red as a dropout color, Even a red postal code frame is stored in the image memory 2 as image data equivalent to the black postal code frame.

Next, the histogram generator 3 uses the image data of the image memory 2 to generate a histogram in the Y-axis direction in the region where the character frame may be described as shown in FIG. 3 (step S1).

Subsequently, the frame Y-coordinate position detection unit 4 finds the peak values A1 and A2 of the histogram, which is the height corresponding to the width of the character frame, from the histogram in the Y-axis direction, and calculates the start coordinates of the character frame in the Y-axis direction. Position Y1 and end coordinate position Y
2 is calculated (step S2). Next, the histogram generation unit 3 generates a histogram in the X-axis direction for the image area range Y1 to Y2 of the Y coordinate, as shown in FIG. 4 (step S3).

Then, the frame X-coordinate position detection unit 5 obtains the number of peaks in which the height of the histogram in the X-axis direction exceeds the threshold value T, and registers the X-coordinate position (step S).
4).

At this time, the registered X coordinate position X [i]
(I is an arbitrary number) is a vertical axis candidate of the character frame. Also,
The value of the threshold value T is set by the threshold value setting unit 8 as a value obtained assuming a standard size of the character frame.

Next, in step S5, the frame detection determination unit 7 checks the number of vertical axis candidates for the registered X coordinate position X [i].

At this time, if the number of vertical axis candidates is 10, it is determined that all the character frames have been detected, and the character frame detection processing ends (step S11).

On the other hand, if the number of vertical axis candidates is not 10 in step S5, the number of vertical axis candidates is checked in step S6. When the number of vertical axis candidates obtained by the threshold value T is less than 10, such as when the frame is faint as shown in FIG. 5 or when the frame is smaller than the standard as shown in FIG.
The threshold setting unit 8 uses the new threshold N as a threshold.
Set. The new threshold value N is a value that satisfies the following expression (step S7).

[0041]

[Formula 1] N = T × α (0 <α <1) When a new threshold value N is set, the frame X-coordinate position detection unit 5 determines that the height of the histogram in the X-axis direction exceeds the threshold value N. The number is calculated, and the X coordinate position X [i] is newly registered as a vertical axis candidate (step S9).

Then, the newly registered X coordinate position X
Check the number of vertical axis candidates [i] (step S1
0) and if the number of vertical axis candidates is 10, the frame detection determination unit 7 determines that all character frames have been detected (step S1).
1) The character frame detection process is ended.

On the other hand, when the number of vertical axis candidates is not 10, it is determined that the character frame has not been detected (step S12), and the character frame detection process is terminated.

Further, as shown in FIG. 7, when the character frame is larger than the standard, the number of vertical axis candidates is checked in step S6, and as a result, the number of vertical axis candidates obtained by the threshold value T is 1
When it is greater than 0, the threshold value setting unit 8 sets a new threshold value N ′ satisfying the following expression as a threshold value. The new threshold value N'is a value that satisfies the following expression (step S8).

[0045]

[Formula 2] N ′ = T × β (β> 1) When a new threshold value N ′ is set, the frame X coordinate position detection unit 5 determines that the height of the histogram in the X axis direction is the threshold value N ′.
The number of peaks exceeding is calculated, and the X coordinate position X [i] is newly registered as a vertical axis candidate (step S9).

Then, the newly registered X coordinate position X
Check the number of vertical axis candidates [i] (step S1
0), if the number of vertical axis candidates is 10, the frame detection determination unit 7 determines that all character frames have been detected (step S11),
The character frame detection process ends.

On the other hand, when the number of vertical axis candidates is not 10, it is determined that the character frame has not been detected (step S12), and the character frame detection process is terminated.

As described above, by providing the threshold setting unit 8 for setting the threshold according to the peak value of the histogram in the character frame detecting device, the character frame is printed faintly or is printed in the standard size. If the vertical axis candidate of the character frame is obtained from the histogram in the X-axis direction when the value is smaller, the threshold value is reset to N = T × α (0 <α <1) to determine the coordinate position of the character frame. Can be detected.

Further, when the character frame is printed larger than the standard size and the vertical axis candidate of the character frame is obtained from the histogram in the X-axis direction, the threshold value is N ′ = T × β (β
By resetting to> 1), the coordinate position of the character frame can be detected.

(Embodiment 2) FIG. 8 is a diagram showing the configuration of the second embodiment of the present invention, in which the width of the peak of the histogram generated in the X-axis direction in the character box detection unit 20 is within the specified range. It is different from the first embodiment in that a peak determining unit 19 is added to determine whether or not the peak interval is within a specified range. The other structure is similar to that of the first embodiment, and the description thereof is omitted.

In the present embodiment, the width of the peak on the histogram that exceeds the threshold T is measured at a point that intersects the threshold T, it is determined whether the width is within the specified range, and By determining whether or not the interval is within the specified range, even if the peak due to characters or the like is higher than the peak of the character frame, the peak can be ignored. Therefore, it is possible to prevent a portion other than the character frame from being detected as the character frame.

The peak width and the specified range of the interval between peaks here are set from a histogram obtained by assuming a standard character frame.

By the way, in the present embodiment, the case where the postal code has five digits has been described, but the present configuration can be applied even if the number of digits is different.

Further, in the present embodiment, the case where the character frames are arranged in the horizontal direction has been described by taking the postal code frame as an example. However, a threshold value is set on the histogram in the Y-axis direction and peaks exceeding it are set. The present configuration can be applied to the case where the character frames are arranged vertically by determining the number of.

[0055]

Since the present invention is constructed as described above, it has the following effects.

According to the first aspect of the present invention, by resetting the threshold value according to the peak value of the histogram,
Even if the print quality of the character frame is low or the size of the character frame is not standardized, the detection accuracy of the character frame can be improved.

According to the second aspect of the present invention, it is determined whether the width of the peak of the histogram exceeding the threshold value T is within the specified range, and whether the interval between the peaks is within the specified range. By determining whether or not it is possible to prevent a portion other than the character frame from being detected as a character frame.

In the third aspect, when there are a plurality of character frames, the position and presence / absence of all the character frames can be detected at one time.

According to the fourth aspect of the invention, the detection of the character frame is applied to the detection of the postal code frame of the mail,
The detection accuracy of the postal code frame is improved, and as a result, the reliability of the mail sorting operation can be improved.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a first exemplary embodiment of the present invention.

FIG. 2 is a flowchart showing a procedure for detecting a character frame by the character frame detection device.

FIG. 3 is a diagram showing a histogram of a character frame in the Y-axis direction.

FIG. 4 is a diagram showing a histogram in the X-axis direction when a character frame has a standard size.

FIG. 5 is a diagram showing a histogram in the X-axis direction when a character frame is blurred.

FIG. 6 is a diagram showing a histogram in the X-axis direction when the character frame is smaller than the standard size.

FIG. 7 is a diagram showing a histogram in the X-axis direction when the character frame has a size larger than the standard.

FIG. 8 is a block diagram showing a configuration of a second exemplary embodiment of the present invention.

FIG. 9 is a block diagram showing a configuration of a conventional character frame detection device.

FIG. 10 is a diagram showing a histogram of a character frame in the Y-axis direction.

FIG. 11 is a diagram showing a histogram of a character frame in the X-axis direction.

[Explanation of symbols]

 L paper sheets 1, 21 scanning unit 2, 22 image memory 3, 23 histogram generation unit 4, 24 frame Y coordinate position detection unit 5, 25 frame X coordinate position detection unit 6, 26 character cutting unit 7, 27 frame detection determination Part 8 Threshold judgment part 10, 20, 30 Character box detection part 19 Peak judgment part

Claims (4)

[Claims] 1. A histogram generation unit that generates two histograms composed of projection accumulation data in the vertical and horizontal directions of the character frame from image data of a paper sheet in which the character frame is described, and a histogram generation unit. Was done 2
While obtaining the position of the peak from the two histograms,
A character box detecting device having a character box detecting unit for detecting the position and presence / absence of the character box by counting the number of peaks exceeding a predetermined threshold value for one of the two histograms. 2. The character box detecting device according to claim 1, wherein the character box detecting section includes a threshold value setting section that adjusts the threshold value according to the number of peaks exceeding the threshold value. 2. The character frame detection device according to claim 1, wherein the character frame detection unit determines whether or not the width of the peak of the histogram is within a specified range, and the interval between the peaks is A character box detecting device comprising a peak determining unit for determining whether or not the character frame is within a prescribed range. 3. The character frame detection device according to claim 1, wherein a plurality of the character frames are arranged on a straight line, and the threshold value is with respect to a histogram in an arrangement direction of the character frames. A defined character box detection device. 4. The character frame detection device according to claim 3, wherein the paper sheet is a postal matter, and a zip code frame is described as the character frame.