CAJ | 학술논문

目前我国流通的人民币都是彩色图像，且携带不同程度的噪声，本文提出用K-means聚类算法对人民币编号进行分割提取，为改进算法的抗噪能力，考虑了图像当前像素点与相邻像素间的影响，实验结果表明，本算法对彩色人民币编号的分割具有自适应性，分割效果较好，鲁棒性强。
목전아국류통적인민폐도시채색도상，차휴대불동정도적조성，본문제출용K-means취류산법대인민폐편호진행분할제취，위개진산법적항조능력，고필료도상당전상소점여상린상소간적영향，실험결과표명，본산법대채색인민폐편호적분할구유자괄응성，분할효과교호，로봉성강。
At present , the circulation of RMB in China are color images , and carry different levels of noise ,using K-means clustering algorithm to extract and split RMB No .is proposed in this paper ,in order to improve the anti-noise of the algorithm ,consider the image pixel point impact between adjacent pixels , the experimental results show that the algorithm has adaptivity characteristic for color RMB serial number segmentation , which has better segmentation and robustness as well .