JPH03182981A - Address reading and classifying machine - Google Patents

Abstract

PURPOSE:To eliminate a need to true up leading directions of mixed postal matters to input them by reading the address in accordance with the carrying direction of a postal matter discriminated based on detection results of the zip code area and the postage print area. CONSTITUTION:The image on the surface of a postal matter is picked up by a scanning part 11 and is converted to image data by a binarizing circuit 20 and is stored in an image buffer 23, and image data of rough resolution is stored in a feature pattern buffer 22 where the whole of the image of the postal matter is stored, and this image data is used to detect not only the zip code area and the postage print area but also the carrying direction of the postal matter. Image data where the zip code area and the postage print area on the postal matter are eliminated is extracted from the feature pattern buffer 22 in accordance with detection results, and the address area is detected from this image data to read the address, and the postal matter is classified in accordance with the address. Thus, it is unnecessary to true up leading directions of mixing postal matters to input them, and the operability is improved.

Description

[Detailed Description of the Invention] [Purpose of the Invention (Industrial Application Field) This invention reads the address or other address written on the mail being transported, and uses the read address. This article relates to an address reading and sorting machine that sorts addresses.

(Prior art) Postal sorting operations can be roughly divided into dispatch sorting operations, which are the sorting operations for sending the collected mail to the destination post office, and sorting operations, which are the sorting operations for delivering the arrived mail. There is a certain delivery classification business.

In the past, dispatch sorting operations have been largely automated using automatic postal code reading and sorting machines, but delivery sorting operations have lagged behind in automation.

However, in recent years, by reading the recipient's address,
There is a growing movement among postal ministries in various countries to automate the sorting of mail down to the delivery district level (automation of delivery sorting operations).

However, there are various kinds of delivery boxes that are sorted by postal code and delivered, such as those that pass through the automatic postal code reading and sorting machine mentioned above, those that are sorted manually, and those that are delivered at special mail receiving stations in the city. Mail is mixed.

In this case, if the leading edge directions of mixed mail pieces are not input into the address reading and sorting machine, the direction in which the addresses are written cannot be fixed, and the number of identification candidates is determined by correcting the reading order of characters within the area and the rotation of character recognition. There is a possibility of unreadability or misreading due to large size, and unstable reading performance. Therefore, when inputting addresses into the address reading and sorting machine, the leading edge of these mixed mail pieces must be aligned manually by the operator. It is man-powered.

For this reason, it has the disadvantage of poor operability.

(Problems to be Solved by the Invention) As mentioned above, this method eliminates the drawback that the leading directions of mixed mail must be aligned and input, resulting in poor operability. To provide an address reading and sorting machine that does not require manual labor to assemble and is very easy to operate.

[Structure of the Invention] (Means for Solving the Problems) The address reading and sorting machine of the present invention reads the address written on the mail to be transported, and sorts the mail using the read address. A scanning means for obtaining a pattern signal by scanning the surface of the above-mentioned mail piece, a converting means for converting the pattern signal obtained by the scanning means into image data, and a converting means for converting the postal item according to the image data from the converting means. a detection means for detecting a number area and a postal stamp face area; a determination means for determining the conveyance direction of the mail by determining the inside of the area based on the detection result of the detection means; The apparatus is comprised of a reading means for reading the address using the image data from the first converting means according to the transport direction of the mail, and a sorting means for sorting according to the address read by the reading means.

(Function) The present invention reads the address written on the mail being conveyed and sorts it using the read address, and by scanning the surface of the mail, a pattern signal is generated. The obtained pattern signal is converted into image data by a converting means, the postal code area and the postage stamp area are detected according to the converted image data, and the area within the above area is detected based on the detection result. The direction of conveyance of the mail is determined by this determination, and the address is read using the image data from the first converting means according to the determined direction of conveyance of the mail. The classification is done accordingly.

(Example) Hereinafter, an example of the present invention will be described with reference to the drawings.

FIGS. 2(a) and 2(b) show an example of an object to be read, such as a mail P, according to the present invention, on the surface of which an address area Pa in which an address and other address are written, and a postal code are written. It is composed of a postal code area Pb1, a postage stamp area Pc, etc., in which the postage stamp face of separate payment stamps, deferred payment stamps, stamps, etc. is written.

FIG. 1 schematically shows an address reading device used in an automatic address reading and sorting machine. That is, reference numeral 11 denotes a scanning section (photoelectric conversion section), which is provided along a transport path (not shown) for mail P that is transported in the direction of arrow a. This scanning unit 11 obtains a pattern signal by optically scanning the postal item P and photoelectrically converting it. It is composed of a self-scanning type CCDC eye sensor that converts into a signal. The scanning section 11 is configured to output electrical signals for each line in synchronization with a line synchronization signal from an oscillator (not shown).

The output of the scanning section 11 is supplied to an address area detection section 12. The address area detection unit 12 detects the postal code area pb and the postage stamp area Pc from the image data of the postal item P, and also detects only the address etc. excluding the postal code area pb and the postage stamp area Pc. This is to extract image data of the address area Pa.

For example, it includes a binarization circuit 20, a preprocessing section 21, a feature pattern buffer 22, an image buffer 23, and a control processor 24.

The binarization circuit 20 is a circuit that binarizes the electric signal for each line from the scanning section 11 in synchronization with a line synchronization signal from an oscillator (not shown), and this binarization signal is sent to the preprocessing section 21. and is output to the image buffer 23. Further, from the binarization circuit 20, a mask signal M corresponding to the length of the above-mentioned mail P is generated by the electric signal from the above-mentioned scanning section 11.
a is output to the preprocessing section 21, feature pattern buffer 22, image buffer 23, and control processor 24. The overall mask signal Ma is a mask signal corresponding to the length of the mail P, and is output in synchronization with the line synchronization signal.

The preprocessing unit 21 processes 8 lines/m from the binarization circuit 20.
Feature extraction is performed by two-dimensionally compressing image data with a resolution of m into image data with a resolution (coarse) of 2 lines/mm.

The feature pattern buffer 22 stores the two-dimensional feature pattern of the entire postal item by sequentially storing image data compressed by the preprocessing unit 21 for the entire postal item P. be.

The image buffers 23 each store image data of a resolution of 8 lines/mm from the binarization circuit 20 for the entire area of the mail P.

The control processor 24 detects the coordinates of the postal code area Pb1 and the postage stamp area Pc on the postal item P by analyzing the two-dimensional feature pattern supplied from the feature pattern buffer 22 using a built-in control program. It detects the transport direction of the object P and also detects the address area Pa.

As shown in FIG.
Shift registers 31a, 31b, 3 for line delay
1c, shift registers 32a, 132b, 32c, and 32d in which data for four pixels of each line is stored, a selector 33, and a RAM 34.

That is, the pattern data of the line currently being read is sequentially supplied to the shift register 32a pixel by pixel,
The shift register 31a stores pattern data for one line one line before the line currently being read, and the shift register 31b stores pattern data for one line two lines before the line currently being read. The shift register 31c stores pattern data for one line three lines before the line currently being read, and the shift register 32a stores pattern data for the currently read pixel and pattern data for three pixels before. , shift register 3
2b stores pattern data for four pixels one line before the same pixel as the shift register 32a, and pattern data for four pixels two lines before the same pixel as the shift register 32a is stored in the shift register 32c. , the shift register 32d has the same pixel as the shift register 32a.
Pattern data for four pixels before the line is stored. As a result, 4×4 pixel (16 bit) data including the pattern data supplied from the binarization circuit 20 is output to the selector 33. Further, the selector 33 is supplied with a recording address from the control processor 27. This selector 33 selectively outputs 4×4 pixel (16 bit) data or the recording address from the control processor 24 to the RAM 34. Data corresponding to 4 x 4 pixel (16 bit) data, that is, a characteristic pattern, is output from the data terminal to the RAM 34. In other words, data with a resolution of 8 lines/mm is converted to data with a resolution of 2 lines/mm. Conversion tables are now stored. The conversion table of the RAM 34 is set by storing the data supplied from the control processor 24 to the data terminal at the recording address from the control processor 24 supplied via the selector. ing.

The control processor 24 controls the feature pattern buffer 22.
can be accessed pixel by pixel, and by scanning predetermined rectangular areas, that is, areas A and C in Figure 4, projection information in the vertical and horizontal directions is obtained as shown in Figure 5. , detect the maximum value on each axis, and use the standard postal code area pb as a binary pattern with 1/2 value as the threshold value, absolute position information, pitch information, etc. as pattern features. This is for determining the entry frame pb. The detection of the postal code area pb is performed by first detecting the projection onto the vertical axis, and then by limiting the detection area in the Y direction to the above area and detecting the horizontal projection.

FIG. 6 is for explaining the detection of the separate payment stamp mark in the charge stamp face area Pc. That is, similar to the postal code area pb, a predetermined rectangular area, that is, areas B and D in FIG. 4, was scanned as shown in the figure, and the logical sum of black pixels on the Generate a projection. To detect the projection of a predetermined range (approximately 25 to 30 mm) from the left end of the postal item P, separate payment stamps, deferred payment stamps, etc. are standardized so that their outer circumference is within the range of 25 to 30 mm), Similarly, the Y projection within the detection area is calculated by the same means, and the projection block detected first from the upper end is defined as a circumscribing rectangle, and Xc
, the ratio with Yc is in the range of 0.9 to 1.1 (detection error), and black is detected first in R on the diagonal from the vertex of the detected rectangle toward the center (approximately 45 degrees). It is determined whether the number of pixels up to is within 110% of the value of "r (X c + Y c ) / 2 Xo, 2".

Then, it is determined whether the black pixel detection distance from the four vertices satisfies two or more vertices, thereby satisfying the condition of a circle. (There is no other charge stamp face area Pc that satisfies this shape). By these determinations, it is possible to detect the presence or absence of separate payment stamps and deferred payment stamps, which are especially large in quantity among the mail P, that is, the presence or absence of the postage stamp surface area Pc.

By detecting the postal code area pb, separate payment stamp, deferred payment stamp, fee amount stamp area Pc, the fifth
The area surrounded by the area XAc5YAo of the postal code area pb shown in the figure and the area X5 of the postal stamp face area Pc shown in FIG.
The area surrounded by SYs is excluded when detecting the address area Pa.

What has been explained in FIG. 6 is an example of detecting separate payment stamps and deferred payment stamps, but the characteristics can be similarly detected for marks such as stamps, payee stamps, and toll meters.

Therefore, the control processor 24 extracts the remaining image data excluding the detected postal code area pb and stamp face area Pc from the feature pattern buffer 22, and extracts vertical and horizontal projection information from the extracted image data. Then, the coordinates of the address area Pa are determined based on these projection information, and only the image data corresponding to the coordinates of the address area Pa is read out from the image buffer 23 and output to the address recognition unit 13 at the next stage, and the above] 3. The conveyance direction of the postal item P is output according to the detected position of the postal code area Pb1 postage stamp area Pc.

The address recognition unit 13 recognizes the address according to the transport direction of the mail P and the image data in the address area Pa from the address area detection unit 12, and uses classification designation data for the address, that is, the classification in the subsequent sorting unit. This outputs data indicating pockets. The addressee recognition unit 13 determines the character reading order within the area and corrects the rotation of character identification based on the transport direction of the mail P from the addressee area detection unit 12.

Next, the operation in such a configuration will be explained.

For example, now, the mail P shown in FIG. 2(a) or (b) is conveyed in the direction of arrow a and reaches the scanning section 11. Then, the scanning unit 11 optically scans the mail P and performs photoelectric conversion, and the photoelectric conversion data is supplied to the address area detection unit 12. Then, the address area detection unit 12 detects the coordinates of the postal code area pb and the postage stamp area Pc from the image data of the postal item P, and calculates the coordinates from the position and shape of the postal code area Pb and the postage stamp area Pc. The transport direction of the postal item P is determined, and image data of the address area Pa excluding the postal code area pb and the postage stamp area Pc is extracted.

That is, when the photoelectric conversion data from the scanning section 11 is supplied to the binarization circuit 20, the mask signal Ma is output from the binarization circuit 20, and the mask signal Ma is outputted from the binarization circuit 20, and the data is sent to the preprocessing section 21, the feature pattern buffer 22, the image buffer 23, and to control processor 24.

Further, the photoelectric conversion data from the scanning section 11 is converted into a binary signal, that is, image data with a resolution of 8 lines/mm by a binarization circuit 20. The image data from this binarization circuit 20 is transferred to an image buffer 25 (or 2
6) and supplied to the preprocessing section 21. The preprocessing unit 21 compresses the image data from the binarization circuit 20 into image data with a resolution (coarse) of 2 lines/mm, and stores it in the feature pattern buffer 22.

5. Then, when the storage process for the entire mail P in each buffer is completed, the mask signal Ma is stopped and an interrupt signal is supplied to the control processor 24. Then, the control processor 24 determines that the storage of the compressed image data for the entire postal item P into the feature pattern buffer 22 is completed, and starts the built-in program. Thereby, the control processor 24 scans the predetermined areas in the feature pattern buffer 23, that is, areas A and C in FIG. 4, extracts vertical and horizontal projection information, and extracts the maximum value in each axis. is detected, and the standard postal code area Pb is determined using a binary pattern using the 1/2 value as a threshold, and the postal code area pb is determined based on absolute position information, pitch information, etc. as pattern features.

The control processor 24 also controls the feature pattern buffer 23.
A predetermined range from the left end of the postal item P is created by raster scanning the predetermined areas within the area, that is, areas B and D in FIG. 6 Similarly, calculate the Y projection in the detection area using the same means, and define the projection block detected first from the upper end as a circumscribed rectangle, and the ratio of Xc and Yc is 0.9. ~1.
1 (detection error), and the number of pixels diagonally from the vertex of the detected rectangle toward the center (approximately 45 degrees) until black is first detected is r (Xc + Yc) / 2XO
, 2J is within ±10% of the value. Then, it is determined whether the black pixel detection distance from the four vertices satisfies two or more vertices, thereby satisfying the condition of a circle. Through these determinations, the presence or absence of separate delivery or deferred delivery, which has a particularly large amount of mail P, is detected, that is, the presence or absence of the postage stamp area Pc.

Incidentally, the detection operation of the postal code area Pb1 postage stamp area Pc is shown in the flowchart of FIG.

Next, the control processor 24 detects the detected postal code area Pb.
The remaining image data excluding the stamp area Pc is extracted from the feature pattern buffer 22, projection information in 7 vertical and horizontal directions is extracted from the extracted image data, and the coordinates of the address area Pa are determined using these projection information. is determined, and only the image data corresponding to the coordinates of this destination area Pa is stored in the image buffer 23.
The address recognition unit 13 of the next stage reads out the information and outputs it to the address recognition unit 13 at the next stage.

The address recognition unit 13 recognizes the address according to the transport direction of the mail P and the image data in the address area Pa from the address area detection unit 12, and uses classification designation data for the address, that is, the classification in the subsequent sorting unit. Outputs data indicating pockets. As a result, the subsequent sorting section sorts the mail P corresponding to the sorting pocket according to the sorting designation data.

As described above, in a system that reads the address written on the mail being transported and sorts it using the read address, an image of the surface of the mail is taken by a scanning unit, and 2 It is converted into image data by a value conversion circuit and stored in an image buffer, and the resolution of this converted image data is converted into image data with a coarse resolution. [8] Detecting the postal code area and postage stamp area and detecting the transport direction of the mail piece using the rough resolution image data stored in the feature pattern buffer, According to the detection results of the postal code area and postal stamp area, image data from which the postal code area and postal stamp area of the postal item have been removed is extracted from the feature pattern buffer, and the address area is detected from this extracted image data. Then, the image data from the image buffer corresponding to the detected address area is used to read the address, and classification is performed according to the read address.

Therefore, the conveying direction of the mail can be easily detected, and from this detection result, the address area and its writing direction can be fixed at a high angle. Thereby, it is possible to reduce unreadability and misreading due to the character reading order within the area that remains ambiguous, and the increase in identification candidates due to rotation correction of character identification. As a result, there is no need for manual labor to align the leading edge of mixed mail, that is, the top and bottom directions.
9 It becomes very easy to operate.

[Effects of the Invention] As described in detail above, according to the present invention, it is not necessary to align and input the leading directions of mixed mail pieces, and it is possible to provide an address reading and sorting machine with very good operability.

[Brief explanation of drawings]

The drawings show an embodiment of the present invention; FIG. 1 is a diagram schematically showing the overall configuration, FIG. 2 is a plan view showing an example of mail, and FIG. 3 is a diagram showing the configuration of the preprocessing section. FIG. 4 is a diagram showing the characteristic detection area, FIG. 5 is a diagram for explaining the detection of the postal code area, FIG. 6 is a diagram for explaining the detection of the postage stamp area, The figure is a flowchart for explaining the detection operation of the postal code area and the postage stamp area. P... Postal item, Pa... Address area, pb... Postal code area, Pc... Postage stamp area, 11... Scanning unit (scanning means), 12... Address area detection unit , 13...
Address recognition section (reading means) 20... Binarization circuit 0 (first conversion means) 21... Preprocessing section (second conversion means) 22... Characteristic pattern buffer, 23...
Image buffer, 24... Control processor (detection means, determination means), 31a s~32a1~...
Shift register, 33...Selector, 34...RA
M.

Claims (1)

[Claims] An address reading and sorting machine that reads the address written on the mail being transported and sorts it using the read address, by scanning the surface of the mail, scanning means for obtaining a pattern signal; conversion means for converting the pattern signal obtained by the scanning means into image data; and detection for detecting a postal code area and a postage stamp area according to the image data from the conversion means. means, determining means for determining the conveying direction of the mail by determining the inside of the area based on the detection result of the detecting means, and according to the conveying direction of the mail determined by the determining means, An address reading and sorting machine comprising: a reading means for reading addresses using the image data from the first converting means; and a sorting means for sorting addresses according to the addresses read by the reading means.