机电工程技术2025，Vol.54Issue(6)：135-139,5.DOI:10.3969/j.issn.1009-9492.2025.06.023
基于旋转圆形特征提取的硬币识别方法
Coin Identification Method Based on a New Spinning Circle Area Feature Extraction
摘要
Abstract
The features extracted based on a new spinning circle feature extraction method of selected ring area in the ROI region of front and back images of the coin obtained by CMOS image sensor,are calculated Pearson correlation with the template features,which can accurately identify the coin information and judge the coin authenticity.Using the mean grayscale to calculate the features,and using the multi-angle image features as templates,it can effectively solve the problem of recognition error caused by local texture differences when the coin enters the imaging equipment in different way.What is more,the credibility of the recognition results is increased by fusing the front and back coin image recognition result.For the coin image at any angle,the circular area rotation is converted into the translation of the rectangular area,which greatly reduces the amount of computation and meets the need of quickly identifying coin information in low-memory embedded system.Finally,using the method above all in practical application,the accuracy of genuine coins recognition can reach 98.875%,and the detection rate of game coins is 100%.关键词
硬币识别鉴伪/旋转圆形特征提取/任意角度图像/平均灰度特征/识别结果融合Key words
coin identification and counterfeiting/spinning circle feature extraction/arbitrary angle image/average grayscale feature/fusing recognition result分类
信息技术与安全科学引用本文复制引用
孙燕,余元超..基于旋转圆形特征提取的硬币识别方法[J].机电工程技术,2025,54(6):135-139,5.基金项目
广东省重点领域研发计划项目(2018B010109004) （2018B010109004）
