UC知道翻译!!求救!在线上等!
来源：百度知道 编辑：UC知道 时间：2024/06/06 10:02:42
摘要：随着社会的发展,字符由于其简洁方便而使用越来越广泛，如IC芯片、身份证、汽车牌照、信件、银行支票等上都有重要的字符信息，这些字符以大小写英文和数字字符为主，对这些字符的自动识别已成为模式识别的一项重要内容。IC芯片表面字符主要包括厂商名称和序列号，自动快速准确识别这些字符对于IC芯片制造及其应用具有重要意义。目前，字符识别主要有人工目测、模板匹配和神经网络等方法，人工目测方法由于人为因素的影响容易发生误测并易使人疲劳，模板匹配要求字符比较完好，容错泛化能力不强。人工神经网络模式识别方法是近些年提出的新方法，它具有良好的容错能力、分类能力、并行处理能力和自学习能力等优点。
本文主要分析芯片表面字符的数字图像处理，对芯片字符的处理包括字符的采集、256级灰度化、滤波、二值化、去离散噪声、芯片字符区域定位、倾斜调整、字符分割、归一化、重排等。通过对芯片表面字符的数字图像处理，取得了阶段性成果。
Abstract: With the development of society, because of its simple characters and facilitate the use of more and more extensive, such as IC chips, identification cards, vehicle licence, letters, bank cheques, and so on are important characters, these characters to the case in English and figures The main characters, the automatic recognition of these characters have become an important part of pattern recognition. IC chip surface, including the main character's name and serial number, automatic rapid and accurate identification of these characters for the IC chip manufacturer and its application is of great significance. At present, the main character recognition of visual, template matching and neural networks, and other methods of artificial visual approach to human factors as the impact of measurement error-prone and vulnerable Shirenpilao, template matching requirements characters relatively intact, fault-tolerant generalization is not strong. Artificial neural network pattern