Bank bill information identification method

Abstract

The invention discloses a bank bill information identification method, which comprises the following steps: S1, de-noising an image of a bill; S2, performing tilt correction on the de-noised image; S3, extracting a continuous digital region in the corrected image; S4, positioning the boundary of the extracted continuous digital region; S5, segmenting the continuous digital region after boundary positioning to obtain a plurality of digital region blocks; S6, performing single character segmentation on the digital region blocks; S7, finding out characters with the highest similarity with to-be-identified characters through template matching, and classifying the to-be-identified characters according to the template similarity; and S8, for the classified to-be-identified characters, setting acorresponding algorithm according to the difference between the templates of all types of to-be-identified characters, and finally determining identification results of the to-be-identified characters. The system can overcome the defects of a traditional working mode of manual bill processing and improve the working efficiency.

Description

technical field [0001] The invention relates to image processing technology, in particular to a bank note information identification method. Background technique [0002] OCR technology can be divided into text-type OCR and professional-type OCR according to the object of use. Text-based systems often include two major functions: layout processing and text recognition. The layout processing includes layout tilt correction, layout analysis, and marking the sequence of text to be recognized; while text recognition is the recognition of text sequence blocks. , you can convert the image of the text into the internal code of the text machine. At present, Tsinghua Wentong (TH-OCR), Shenyang Institute of Automation (SY-OCR), Beixin (BI-OCR) and so on are more common. Most of the professional systems are for specific industries, such as the financial system, customs, taxation and other information formats are relatively fixed, and the amount of character recognition is relatively ...

AI Technical Summary

Problems solved by technology

However, the division and identification management of bills is not very good. Therefore, various enterprises and units are very inefficient in the process of processing bills, and there are also some problems: (1) The backup of paper bills is small. It is easy to cause information loss when biting, which may lead to irreparable problems; (2) the authenticity of paper bills can only be judged by manpower, but it is easy to interfere and misjudge the judgment results on subjective factors; (3) Paper bills need to be manually filed and manually searched, which will waste manpower and material resources, and will also consume a lot of financial resources of the enterprise, resulting in low service quality; (4) Paper bills cannot be processed in the information age. Online transmission makes some financial requirements impossible to meet; (5) bill information is diverse, and the digital information is not as relevant as Chinese characters. When reading manually, the entire number string will be wrong due to misjudgment, making the message delivery error

Images