【作者】 张颖； 陈雪波； 吴庆洪； 梁英顺；

【Author】 Zhang Ying~(1,2) Chen Xuebo~2 Wu Qinghong~2 Liang Yingshun~2 ~1(School of Electronic and Information Engineering, Dalian University of Technology, Dalian 116024, China) ~2(School of Electronic and Information Engineering, Anshan University of Technology, Anshan 114044, China)

【机构】 大连理工大学电信学院； 鞍山科技大学电信学院；

【摘要】 设计研究人民币纸币清分系统中的主要模块,如图像采集、处理和识别。在硬件部分,选用CM326A4作为系统的图像传感器,选用DSP芯片TMS320C6711进行纸币图像的处理和识别;在软件部分,选取预处理之后的纸币图像的尺寸特征用模糊逻辑推理方法识别图像的面值,提取矩作为特征,采用三层BP神经网络来识别纸币的正反面和正反向,并通过实验证明该系统的有效性和可行性。更多还原

【Abstract】 The main modules, such as image sampling, processing and identification, in identification system of RMB paper currency are designed and studied. In hardwares, CM326A4 is served as the image sensor and DSP chip TMS320C6711 is used to process and identify RMB images. In softwares, size feature is chosen to identify the value of RMB paper currency images by the method of fuzzy reasoning which have been preprocessed. Moment invariants are selected as features to recognize the orientation of RMB paper currency by a three-layer BP neural network. The validity and feasibility of the system are proved by experiments.更多还原