Non-digitizing Data Restoration with Using Indirect Data Processing

摘要

In this study, we proposed the method of automatics searching predefined events location in digital images of old paper-tape data recording, which in essence is indirect processing. The main idea of proposed algorithm is isomorphic transformation of paperȁ3;tape digital images to the time-serial data. The time-serial data obtained by this transformation is clustered and classified to obtain the positions of predefined events in digital paper-tape-images. The regular approach to data processing (pattern recognition for image or advanced image processing) is very laborious. In most cases only part of information contained in paper tape is interesting. The method can be used as preliminary estimation amount information and position in a paper-tape of this information before more sophisticated data processing. Proposed algorithm is fast and has sufficient good quality of data restoration for the positioning of predefined events.