OCR bill problem identification method and system based on feature detection

Abstract

The invention discloses an OCR bill problem identification method and system based on feature detection. The method comprises the steps of performing background elimination and denoising processing ona bill image to obtain a first image; performing projection transformation on the first image, removing a separator, and identifying characters to obtain a bill text; extracting a keyword combinationof the bill text; matching the keyword combinations in a knowledge base to obtain a rule set with the highest matching degree as a reference rule set; and if the audit content information of the billtext is greater than a preset risk threshold of the reference rule set, sending alarm information. According to the method, the problem that original file handwriting definition, cleanliness and thelike are difficult to recognize is effectively solved, bills can be imported on a large scale, similarity recognition can be rapidly carried out on the bills, OCR recognition error correction is achieved through intelligent verification of intelligent recognition based on text character features, and the accuracy is further improved. According to the positioned risk points, the auditing workload is rapidly reduced, and the user experience of the informatization system is improved.

Description

technical field [0001] The present disclosure relates to the fields of OCR image recognition and pattern matching, in particular to a method and system for OCR recognition of bills based on feature detection. Background technique [0002] A bill is a written certificate that specifies the actual situation of transactions and matters related to economics, transportation and other businesses, and is the paper or electronic basis for the payment of goods and rights certificates. Bills can be used as the basis for extraction, consumption, and authority. Due to the variety of types, pattern matching, positioning and identification of bills are very difficult; due to the variety of bills, such as invoices, air tickets, train tickets, receipts, receipts, Stubs, shopping coupons, etc., these different bills have different positions of feature words, expressions of sensitive words are also different, and the capitalization of numbers is not standardized. Therefore, if the number of b...

AI Technical Summary

Problems solved by technology

Bills can be used as the basis for extraction, consumption, and authority. Due to the variety of types, pattern matching, positioning and identification of bills are very difficult; due to the variety of bills, such as invoices, air tickets, train tickets, receipts, receipts, Stubs, shopping coupons, etc., these different bills have different positions of feature words, expressions of sensitive words are also different, and the capitalization of numbers is not standardized. Therefore, if the number of bills is large, in the audit work, the audit rules Content collection, including: a collection of audit rules for invoice type, unit name, and amount. There are various categories, so it is very difficult to identify and judge manually, and it is extremely error-prone

Images