At the turn of the 20th century, mail volume burgeoned dramatically while mail handling became an extremely labor-intensive and time-consuming process. The answer to these problems was cutting-edge mechanization for its time. First, sorting machines were proposed by inventors already in the early 1900s. However, broad development of mechanization in postal operations was not until the mid-1950s. During the 1960s and 70s, mechanization turned into the backbone of letter-sorting operations bringing increased productivity. By the mid-1970s, escalating mail volume required even more efficient methods and equipment.
The transition from mechanization to automation for the U.S. Postal Service started in 1982, when the first optical character reader was installed in
In the 80s, the first OCRs were confined to reading the ZIP Code. In the 90s, they expanded their capabilities to reading the entire address, and in 1996, the Remote Computer Reader (RCR) for the USPS could recognize about 35% of machine printed and two percent of handwritten letter mailpieces. Another 10 years brought about an unparalleled breakthrough in reading quality.
Today, modern systems can recognize 93% of machine-printed and about 88% of handwritten letter mail more than 90% cumulatively. Due to this progress in recognition technology the most important factor in the efficiency of mail sorting equipment the cost of processing the mail dropped from an estimated $55 per 1,000 letters manually to less than $5 per 1,000 letters with automated sorting. Another significant achievement of the last decade is that technology has enabled the automation of mail types that were difficult or impossible to automate 10 years ago, for example, magazines and parcels.
Thus, a decade of intensive investment in automated sorting technology, based on OCR, resulted in high recognition rates of machine-printed and handwritten addresses delivered by state-of-the-art systems. Analysis of the entire mail stream, and the understanding of the current state of the technology and its potential, may give us answers about the future of mail processing.
Is there potential for further improvement in address recognition rates?
Today it is more difficult to benefit from further automation of mail sorting. The reason is the low volume of non-automated mail. With the high read rates already achieved by the industry, the volume of residual items that has yet to be processed automatically is comparatively low. To further automate the sorting of these leftover mailpieces would provide lower additional savings, when compared to the savings achieved in the past. Postal operators in countries with lower mail volumes face a similar problem even though their current level of automation may not be high further automation would be too costly for the low mail input.
One way to reduce the cost of automation is to rely on smaller suppliers that can often come up with more efficient solutions than larger vendors. Actually, many state-of-the-art solutions offered by large suppliers are based on or include the software provided by small subcontractors.
Until recently, this was not easy, as suppliers provided integrated solutions that included both OCR and transport components. This resulted in the tight integration of sorting equipment and OCR module(s), hampering the replacement or updating of one of the components independently.
This is no longer the case with the introduction of the OCR/Video Coding Systems (OCR/VCS) open interface standard by the European Committee for Standardization. The OCR/VCS standard enables postal operators to work with different suppliers on needed replacements or expansions of sub-systems without incurring significant engineering costs.
Another opportunity to raise the efficiency of mail sorting is to rely on universal OCR technology. Universal OCR technology requires comparatively low expenses to customize and can be more easily applied to different address formats and coding rules.
Universal technology also helps postal operators to unify efforts across different types of mail streams letters, flats and parcels. In this case, the savings received from the reduction of labor on a combined mail volume helps justify the OCR improvement project.
Each OCR implementation is customer- and task-specific; therefore, the right system has to be selected in order to achieve the highest possible efficiency for a particular case. For example, if the reject mail stream contains up to 50%-60% of handwritten addresses, using an OCR engine that has a high performance level on both machine-printed and handwritten addresses will make automation investments more viable.
Although automated address reading has achieved a very high level of recognition and accuracy in recent years, there is still a portion of addresses on mailpieces that are not sufficiently recognized and therefore need to be entered manually. Since the automated mail-sorting environment is driven by speed, recognition rates and accuracy, the remaining manual entry operations have to be in step with the standards of the automated processes. The only way to make the manual stage as efficient as the automated sorting process is to equip operators with advanced keying tools that minimize the amount of keyboard interaction and reduce potential typing errors.
Modern video coding systems are robust, easy-to-use and can be readily configured to meet specific application requirements. As with the OCR component, it is important for the video coding system to be compliant with OCR/VCS Open Interface Standard.
What other automatic solutions can benefit postal services?
A significant percent of mail is incorrectly addressed and has to be forwarded or returned. This results in additional handling and processing costs. OCR technology can also assist in improving automated mail forwarding processes to solve this problem.
This modern technology has the capability to increase recognition accuracy while reducing errors to levels unachievable by human operators, an important factor to consider when implementing a mail processing solution.
All pieces of mail delivered by post must have stamps or "postal indicia" indicating that the proper postage and/or a postal permit exist to allow the items to enter the mail delivery network. Indicia are used by the Postal Service to verify that postage was paid, and at what rate. Today, postage authenticity and value detection rely on manual labor and are error-prone processes. New systems automatically locate and detect different types of indicia on envelope images and reliably read the postage amounts printed on them, avoiding revenue losses.
There are many technologies available on the market that deserve a closer look: the detection of overlapping mailpieces (so called "double feed"), the verification of the format and layout of a mailpiece to detect items that require specific processing (like foreign mail), the verification of handwriting against a blacklist of samples to help with security problems, datamining to better define where and which services are needed, to name just a few.
Postal services are going to remain an integral part of the infrastructure for any economy. For example, recent growth in e-commerce has caused a rise in international and domestic postal parcel traffic. To sustain the role of mail as one of the most efficient means of business communication, postal services have to permanently improve their organizational and technological infrastructure for mail processing and delivery, ensuring that mail is relevant, universal and non-intrusive. New technology advances are one of the driving forces that will enable postal operators to make innovative changes.
As Vice President for Business Development at Parascript, John Buck is responsible for augmenting business growth and leading new partner development efforts. He can be reached at john.buck@parascript.com.
