JPS6122349B2 - - Google Patents

Description

[Detailed description of the invention]

The present invention relates to a postal code detection device that automatically detects postal codes written on postal items such as postcards and envelopes. In a postal code detection device, the usefulness of a hyphen (located between three digits of a parent number and two digits of a child number) as a means of detecting a postal code has been confirmed. Therefore, an object of the present invention is to provide a postal code detection device that can detect hyphen information in addition to character detection information, and can reliably detect a postal code by using the information in combination. It is. Next, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a drawing showing an example of a postal letter created with printed type. In the figure, the entire letter 10, postal code 11, addressee 12, postal mark 13, and stamp 14 are shown. This letter 1
The area surrounded by rectangles a, b, c, and d within 0 is scanned by the scanning means. It is assumed that scanning starts from point a and is performed in the direction of pq, that is, from bottom to top and from left to right. It is known that the mask detection method can be used to detect the position and size of letters and patterns on letters from scanned signals. and hyphen detection will be explained with reference to FIG. 2 and subsequent figures. FIG. 2 is a block diagram of an embodiment in which a postal letter 10 is being scanned by a scanner 20. In this example, Flying Spot Scanner (FSS)
Vertical direction (32
mm) and lateral (48 mm) area resulting in
Create a pattern of 64 meshes. this pattern
The signal is input to a pattern shift register 30 that shifts in synchronization with the scanning of the FSS. A detailed diagram of pattern shift register 30 is shown in FIG. This pattern shift register 30 has 64
It consists of 28 meshes, and the IN pattern shown in the figure is input and the pattern is output from OUT. The vertical positional relationship of the pattern is as shown in the figure, and the connection relationship of the shift registers in this two-dimensional pattern is such that, for example, the output of the first column from IN becomes the input of the next second column. The following connections are made in the same way. Further, 31 and 32 are detection points for mask detection, and their detailed positional relationship is shown in FIG. Next, a circuit 40 is shown which performs mask detection based on the information of the detection point and the shift timing of the pattern shift register 30. An outline of the mask is shown in FIG. AC1 for the mask
~AC8, AMA1~AMA8, AMB1~AMB8,
There are SMB3 to SMB5, which are logically summed with the contents of each position of the detection point. In other words, it becomes as follows. AMA1=B6+C6+D6+E6+F6+G6 AMA2=B7+C7+D7+E7+F7+G7 AMA3=B8+C8+D8+E8+F8+G8 : : AMB1=H6+I6+J6+K6+L6 AMB2=H7+I7+J7+K7+L7 AMB3=H8+I8+J8+K8+L8 : : SMB3 =F1+F2+F3+F4+F5+F6+F7 SMB4 =G1+G2+G3+G4+G5+G6+G7 SMB5 =H1+H2+H3+H4+H5+H6+H7 Detecting the bottom and top edges of characters using this mask is bottom and top detection. 6, the operation of which is shown in FIG. Figure 6 1 is an example of lower end detection, with AMA3, 4, 5,
2 and 3, that is, the pattern exists in AMA3, and the other AMA4, AMA5, AC2, AC3
It is assumed that the lower end is detected when there is no pattern in
It is assumed that the upper end is detected when the number is 4 or 5.
In other words, as you can see from the figure, there is a pattern in AMA3, and the other AMA2, AMA1, AC
4. The top edge is detected when there is no pattern in AC5. Further, in parallel with these end detection operations, the height is detected by the height detector 44 by counting shift pulses from the time when the lower end is detected until the upper end is detected. Also, the 8th column of detection points of the pattern shift register from when the bottom edge is detected to when the top edge is detected, that is, AC
3. A compressed pattern figure is created in the compressed pattern register (included in the pattern shift register 30) by taking the logical sum of the patterns corresponding to AMA3, AMB3, etc. but passing through the time. The compression pattern register and an example of the compression pattern are shown in FIG. 6. The length of consecutive "1"s from the 0th bit on the compressed pattern register to "0" (corresponding to a white pattern) until the next "0" becomes the width of the character. In other words, when the "0" bit is not "0", it is not detected as a character or a hyphen. Width detection of characters and hyphens that perform these operations is performed by a width detector 43. Characters and hyphens are detected based on the upper edge, lower edge, height, and width information. Also, although I did not explain this here, SMB and AMB masks are used for slightly larger characters, but I will not explain them here. Types generally used as postal codes in printed letters are mainly those output from line printers (LP) and types called T-3 fonts. Width (W) and height (H) of those numeric characters
is T-3 font, H = 3.5mm, W = 2.1mm or
For LP (OCR-B font) H=2.3mm, W=
It is 1.5mm and the width to height ratio (W/H) is 60
%, and it can be seen that the height is higher. Also, characters with a height of 6 mm or more are unthinkable. Here, the conditions for determining it as a character are when the height (H) is greater than or equal to the width (W), and when the height is
If it is less than 12 meshes, it is determined as a character and sent to the next position determination circuit 70 together with height information.
This condition also takes into consideration blurring of upper and lower printing. Also, as above, as a hyphen check, the height is 3 meshes and the width is 4 meshes ~
14 meshes, a hyphen is detected and sent to the next position determining circuit 70. Character hyphen detector 4
When a character or a hyphen is detected in step 5, the information is sent to the position determination circuit and also to the XY position detection circuit 60, which serves as a reference for shifting. By counting the shift pulses from the timing generation circuit 50, position information in the X-Y direction when a hyphen character is detected is sent to the position determining circuit 70. Mask detection circuit 40 and X using FIG. 1 as an example
An example of the output of the -Y position detection circuit 60 is shown in FIG.
FIG. 7 shows an example of the scanning area of the example letter in FIG. It is a hyphen mark.
The order of detection is the order of the numbers attached to M and H. Also, x written next to M and H,
y, z are the detected characters, the X coordinate of the hyphen, Y
This is coordinate and height information. These contents are input to the position determination circuit 70. A detailed block diagram of the position determining circuit 70 is shown in FIG. 8, and the position determining circuit 70 will be explained in detail using that diagram. The hyphen, character height and XY position are entered,
The results are stored in the first memory 71 in the following order. Mark X coordinate Y coordinate Height M ₁ x ₁ y ₁ h ₁₁ M ₂ x ₂ y ₂ h ₂ M ₃ x ₃ y ₃ h ₃ M ₄ x 4 y ₄ h ₄ H ₅ x ₅ y ₅ h ₅ M _{6 x 6} y ₆ h ₆ M ₁₂ x ₁₂ y ₁₂ h ₁₂ H ₁₃ x ₁₃ y ₁₃ h ₁₃ M ₁₄ x ₁₄ y ₁₄ h ₁₄ M ₁₉ x ₁₉ y ₁₉ h ₁₉ After the data is stored in the first memory 71, the control unit 76 For grouping in the Y direction, the Y-coordinate values are compared by the Y-direction grouping circuit 72 and stored in the second memory 74 by the rearrangement circuit 73 in descending order of the Y-direction values. An example of its storage is shown below.

[Table] In the above example arranged in descending order of Y-direction values, the y-coordinate values of the first group are y ₁ = y ₂ = y ₃ = y ₄
= y ₅ = y ⁶ = y ⁷ = y ₁₁ = y ₁₅ = y ₁₉ , and these characters and hyphen strings are set as group 1 and are similarly grouped. When the grouping in each Y direction is completed and the data is completely stored in the second memory 74 for each group, the postal code detection circuit 75 operates. First, group 1 information is input to a character spacing (pitch) check circuit 750, and the character spacing (pitch) is measured. FIG. 7 3 is a diagram showing an example of pitch measurement for group 1,
The pitch between M ₁ and M ₂ is usually determined by the difference between x ₂ and x ₁ , and the pitch is P ₁ . A block that looks like a postal code is checked from these _P1 to _P2 pitches. The method for detecting this block is as shown below: ``Method of writing postal codes based on the provisions of the Postal Act, etc.'' (Ministry of Posts and Telecommunications Notification No. 196 of 1952)
(hereinafter referred to as Document A). Block detection involves detecting three characters at equal pitches from the left of the X coordinate (in descending order of value) as a block that seems to be a postal code. In the example 3 in Figure 7, since P ₂ = P ₃ , M ₂ , M ₃ , M ₄
The three characters are detected as the highest priority block. Next, three characters M ₁₁ , M ₁₅ , and M ₁₉ are detected as the next block. Next, the hyphen check circuit 752
It is confirmed that the character examples M ₂ , M ₃ , and M ₄ are postal codes. This may be due to measurement errors, for example.
It is valid when P ₂ ≠ P ₃ , etc. Furthermore, when a hyphen is not detected in group 1, the block on the left side is preferentially output as a postal code and its position is output. The final decision for each group is
When a postal code block is detected by the postal code check circuit 751, the position address of that postal code is output. Also, when no block is detected in group 1, the postal code check circuit 75 in the second memory is used to check the next group.
A signal is input by 1, and the next second group 1
information is input to the postal code detection circuit 75. If no block is detected, the above operations are performed until a group exists. Note that the position determining circuit 70 (FIGS. 2 and 8)
It is clear that the same logic functions can be accomplished in a conventional digital computer as shown in logic circuits and other functions. It may be programmed according to the detailed logic functions shown here.

[Brief explanation of the drawing]

FIG. 1 shows the scanning area for postal code detection in an example postal letter containing a printed postal code. FIG. 2 is a block diagram of a circuit for detecting postal codes. Figure 3 shows the pattern
Detailed diagram of shift register. FIG. 4 is a positional relationship diagram of detection points for mask detection in the pattern shift register. FIG. 5 is an example of a mask for mask detection. FIG. 6 is an explanatory diagram of the operation of mask detection, where 1 is lower end detection and 2 is upper end detection. Further, 3 is a compression pattern register for width detection. FIG. 7 is an example of information output after mask detection in the example of scanning the letter in FIG. 1, and is an explanatory diagram of the operation of the position determining circuit. 1 is an example of scanning the letter in FIG. 1, 2 is an example of mask detection and output of XY position, and 3 is an example of pitch measurement after grouping. FIG. 8 is a detailed block diagram of the position determining circuit.

Claims (1)

[Claims] 1 Scanning means for scanning a predetermined area of mail, a pattern shift register that receives signals from said scanning means, and a pattern shift register that receives signals from said pattern shift register to determine the top, bottom, and height of each pattern included in said area. a mask detection means for detecting height and width; a character hyphen detection means for detecting a character and a hyphen based on the relationship between the height and width of a pattern based on the output from the mask detection means; and a hyphen is detected by the character hyphen detection means. What is claimed is: 1. A postal code detection device comprising means for determining a character in a predetermined positional relationship with the hyphen as a postal code.