Invoice classification method based on character recognition and semantic analysis

Abstract

The invention discloses an invoice classification method based on character recognition and semantic analysis. The invention aims to detect the types of invoices. The method includes the following steps that: an image acquisition device is adopted to acquire an invoice image; the length and width of an invoice in the image are detected through Hough transform; the length and width are compared with preset thresholds, so that the preliminary classification of the invoice is realized; a corresponding invoice template is called according to a preliminary classification result, and a region wherecharacter recognition is needed is cut out; and a third-party character recognition system such as a Baidu recognition system is called to recognize characters in the above region; and after recognition is completed, semantic analysis is performed on a recognition result, and the accurate result of invoice classification is obtained. With the method of the invention adopted, a problem that accurate recognition cannot be realized in invoice recognition due to failure to classify an invoice type can be solved. The method is applicable to a plurality of different types of invoices, and has high application value in invoice recognition and smart financial reimbursement.

Description

technical field [0001] The invention relates to the technical fields of image processing and artificial intelligence, in particular to an invoice classification method based on character recognition and semantic analysis. Background technique [0002] In recent years, with the rapid development of my country's social economy, economic activities have become increasingly frequent, and both ordinary consumers and various types of enterprises have become more and more aware of the fact that consumption must be invoiced and reimbursed by the ticket. At present, hundreds of millions of invoices are used for reimbursement in my country every year, and the types of invoices are also increasing. If these invoices are manually processed, on the one hand, it will consume a lot of manpower and material resources, increase the workload of the financial personnel responsible for reimbursement, and take up a lot of extra energy for the reimbursement person. On the other hand, manual proc...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the deficiencies in the prior art, the present invention provides an invoice classification method based on text recognition and semantic analysis, which solves the problems of slow classification of invoice types and low efficiency of invoice reimbursement

Images