Intelligent card surface character recognition method based on adaptive parameter adjustment

Abstract

The invention discloses an intelligent card surface character recognition method based on adaptive parameter adjustment. The method comprises the following steps: S1, creating a template: creating template information required for recognizing the surface characters of the intelligent card; s2, parameter search: using a learning model to obtain an image preprocessing parameter satisfying an image identification precision requirement; and S3, image identification: according to the template information created in the S1 and the image preprocessing parameters obtained in the S2, identifying characters on the surface of the intelligent card by adopting a template matching method, and obtaining a final identification result.

Description

technical field [0001] The invention relates to the field of smart card image recognition, in particular to a method for recognizing characters on the surface of a smart card based on adaptive parameter adjustment. Background technique [0002] Smart cards are IC chip cards including bank cards and social security cards. Different layouts of smart cards have different card backgrounds, colors, and patterns. Smart card characters are different from handwritten characters. Smart card characters have the characteristics of standardized fonts, uniform sizes, and background interference (personalized background of smart cards). In addition, smart card character recognition needs to illuminate the card with a light source such as ring light or bar light to reduce the influence of natural light. However, changes in external light will still cause subtle changes in the surface illumination of the card. To sum up, the smart card character recognition method needs to adapt to differe...

AI Technical Summary

Problems solved by technology

[0004] In contrast, the character recognition method based on machine learning has the advantage that it can automatically adapt to various lighting and background changes after the model training is over; the disadvantages are: (1) It requires a large number of character training samples under different lighting and backgrounds; (2) ) The trained model cannot recognize the new card background, and the entire model needs to be retrained; (3) The model is prone to over-fitting and under-fitting states

[0005] The advantage of the character recognition method based on template matching is that character recognition can be realized without a large number of templates (single or a small number of samples). Influence, and the parameters are coupled with each other and there is no standard; (2) There are many character templates, and for each card background, a character template needs to be created to correspond to it; (3) The test is difficult, and it needs to be tested on a card first After the character recognition is successful, the batch cards are recognized and the accuracy is counted. If the recognition accuracy requirements cannot be met, the parameters need to be re-adjusted and recognized again

Therefore, when the illumination changes, the smart card character recognition method based on template matching needs to adjust the image preprocessing parameters of the corresponding character template for each background character template, and the process involves many parameters (m templates, n parameters, a total of m×n parameters) and the parameter adjustment results are uncertain, which leads to time-consuming and labor-intensive smart card character recognition process

Images