| 摘要 |
<P>PROBLEM TO BE SOLVED: To provide a letter classification system for large-lot letters, which improves automatic treatment efficiency by reducing a load on a manual work. <P>SOLUTION: The letter classification system for classifying destinations of a series of large-lot letters to be sent from a big sender includes: an image scanner 2 for scanning surfaces with the destinations of the respective letters written thereon and generating letter images; a reading part 3 for extracting destination area images with the destinations written thereon from the letter images, and reading destination information from the destination area images; a destination database 4 for associating the destination area images with the destination information read by the reading part 3 from the destination area images so as to store them by each large-lot letter; and a destination estimation part 5 for comparing reading impossible images from which the reading part 3 cannot read the destination information with destination area images of past large-lot letters when information of the past large-lot letters by the same big sender is stored in the destination database 4, and estimating the destination information of the reading impossible images. <P>COPYRIGHT: (C)2012,JPO&INPIT |
