Image analysis – Applications – Mail processing
Reexamination Certificate
1999-07-16
2003-12-16
Mehta, Bhavesh (Department: 2625)
Image analysis
Applications
Mail processing
C209S003100, C209S584000, C209S900000, C382S321000
Reexamination Certificate
active
06665422
ABSTRACT:
BACKGROUND OF THE INVENTION
Field of the Invention
The invention relates to the recognition of distribution information on mail items. In the broadest sense, the invention relates to the field of automatic mail-item processing, and particularly to systems in which an automatic OCR address-reading method is supplemented and improved by video coding.
Technical Background
Optical character recognition (OCR) is an established image-processing method in which images are recorded with an appropriate resolution, and machine-readable characters (letters, numbers) are ascertained from the images. This method is widely used in, for example, text processing, form reading and automatic sorting of mail items.
Systems for automatically reading addresses (OCR) are well-known in the field of letter processing, and are described in, for example, DE 195 31 392. Modern OCR letter-sorting systems can attain processing rates of ten letters per second, or 36,000 letters per hour, and higher. However, the recognition reliability varies greatly with the writing style and the overall quality of the address information located on the letter surface. In the event of successful recognition, the relevant letter can be provided with a machine-readable barcode. This barcode permits further mechanical processing up to an arbitrary sorting order. In particular, the use of barcodes permits a sorting of mail items, such as long letters or small parcels, up to the sorting level of the mail delivery, at which the delivery person sorts the items according to their distribution sequences.
The need for address reading also exists for larger parcels, newspapers and periodicals.
Although the above-described address reading is a tried-and-true technique, its use for the sizeable number of periodicals, newspapers, small packages and parcels that are mailed is limited by difficulties in automatically and reliably locating and reading the address region, also referred to as the ROI (Region of Interest). The problems associated with automatic ROI determination through image-processing methods occur particularly with flat mail items such as periodicals, newspapers, books and parcels because of background interferences or interfering surfaces. This is particularly due to the fact that different address labels are applied to a background that contains text (newspapers) or graphics (periodicals, magazines, books). Errors in ROI determination are critical, because they render the remaining automatic process impossible.
Although complex image-processing algorithms for ROI determination exist, various manual actions are necessary for performing the ROI search according to the current state of the technology.
In accordance with DE 394 30 73, a letter is placed on a digitization tray and fixed against stops. A marking pen is used to enter scanning coordinates for the address block of interest; the coordinates of the corner points of the address block are transferred to the system's control computer by way of an interface. This method can be used to process mass mailings, because the coordinates need only be entered once. This method is not effective for individual mail items, and cannot be performed for parcels.
The use of computer displays to support the automatic ROI search is described in Pattern Recognition, Volume 23, 1990, pp. 347-362, “Preprocessing and Presorting of Items for Automatic Sorting Using OCR,” and EP 0 589 119.
The following disadvantages are associated with this practice:
special input means that do not interrupt keyboard entry are necessary;
precise location information about the ROI is necessary;
large images of periodicals or parcels cannot be processed on a standard computer display;
there is no integration of the ROI determination into the complete address-reading solution, including OCR and video coding.
There is therefore a need for a solution for fast and reliable determination of regions of interest, particularly address regions, with the aid of video coding in large video images of mail-item surfaces having different backgrounds, the solution permitting an online processing of the items. No input means should be necessary in addition to the keyboard for assuring fast video coding.
SUMMARY OF THE INVENTION
In many case, for distribution information that is located in large images with different backgrounds, it is not possible to automatically find and exclusively evaluate the information without the intervention of a video coder. The disclosed embodiment of the invention has, for the first time, combined video coding with the automatic OCR evaluation process, including the search for the regions (ROI) containing the desired information, especially with the zoom function initiated by the operator and the processing and display of the zoomed image section, all with little input effort, making the processing extremely fast. The multi-stage video-coding and OCR process has proven particularly effective, thus permitting an online processing of the mail items.
Moreover, this solution requires no special, costly monitors or workstations.
The invention is described in detail below by way of an embodiment.
REFERENCES:
patent: 4992649 (1991-02-01), Mampe et al.
patent: 5031223 (1991-07-01), Rosenbaum et al.
patent: 5311999 (1994-05-01), Malow et al.
patent: 5754671 (1998-05-01), Higgins et al.
patent: 6239397 (2001-05-01), Rosenbaum
patent: 39 43 073 (1989-12-01), None
patent: 195 31 392 (1995-08-01), None
patent: 0076332 (1981-09-01), None
patent: 9589119 (1992-09-01), None
patent: 08103733 (1996-04-01), None
Keil Göran
Seidel Gert
Desire Gregory
Kinberg Robert
Mehta Bhavesh
Siemens Aktiengesellchaft
Venable LLP
LandOfFree
Method and device for recognizing distribution data on... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and device for recognizing distribution data on..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and device for recognizing distribution data on... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3123260