A banknote image recognition method

Abstract

The present invention proposes a banknote image recognition method, which uses the banknote images in the training set to train a classifier, and uses the trained classifier to identify the currency, denomination, orientation and version of the acquired banknote images to be tested, specifically as follows Steps: 1) Preprocess the banknote image; 2) Take several image blocks in the banknote image; 3) Calculate the image features of each image block to form the feature vector of the banknote image, as the image in the training set or the image to be tested eigenvectors of each image block; 4) using the eigenvectors of each image block of the image in the training set to train a classifier; 5) inputting the eigenvectors of each image block of the image to be tested into the classifier to identify the banknote image. The method of the present invention does not need to search for local differences between different banknotes, and only needs to extract corresponding features on the banknote images to perform subsequent recognition.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to a banknote image recognition method. Background technique [0002] With the development of economic globalization and trade integration, my country's opening to the outside world has been deepening, and trade exchanges with various countries have become increasingly frequent. At the same time, the increasingly popular outbound travel market has also promoted the rapid development of related businesses such as foreign currency payment and exchange. Therefore, effective management of various foreign currencies is very important. On the other hand, foreign markets also require banknote counters and sorting machines to support automatic multi-currency functions. [0003] For the sorting and counterfeit identification of banknote images, it is first necessary to identify the face value, orientation, and version of banknotes, and secondly, it is also necessary to ident...

AI Technical Summary

Problems solved by technology

However, in multi-currency recognition, or when there are many versions of the Hong Kong dollar currency, the number of templates will be large, so template matching will be time-consuming and affect processing performance; on the other hand, when banknotes have folded or missing edges, or For example, when the length and width of the U.S. dollar are the same but the white borders around it will be offset by printing, the template matching method will easily lead to recognition errors

Images