UC知道毕业论文翻译求助，急！行的进！加分
来源：百度知道 编辑：UC知道 时间：2024/06/23 14:28:43
摘要：索书号识别对图书馆藏书管理具有重要的意义。而索书号字符识别作为其中关键的环节，它综合了图像处理，文字识别以及机器学习等多种技术。本文根据自然条件下索书号字符模糊，褪色，破损，有污渍，光照不均匀等特点，提出了多种图像分割方法，其中包括：并行边界分割，串行边界分割，并行区域分割，串行区域分割。经过综合比价，最后为了更好的描述字符特征，提出了一种切实可行的分割方法：利用边缘点数量统计的文字图像分割方法，来进行文字图像的分割与提取。此算法能有效地应对自然环境下图书索书号的破损，褪色，光照不均匀等情况。
实验结果表明，本文提出的图像处理和分类器设计方法有着良好的效果，在识别准确率和识别速度方面都满足了实际应用的需要。
Call number to identify the management of the library collection is of great significance. Call Number and character recognition as one of the key elements, which combines image processing, character recognition, as well as a variety of technologies such as machine learning. In this paper, under the call number under the natural conditions of fuzzy characters, faded, damaged, there are stains, uneven illumination, etc., put forward a variety of image segmentation methods, including: parallel boundary partition, serial boundary segmentation, region segmentation parallel, serial region partition. After comprehensive comparison, the last character in order to better describe the characteristics of a practical segmentation method: the number of edge points using statistical methods of image segmentation of the text, text to image segmentation and extraction. This algorithm can effectively deal with the natural environment under the call number of books damaged, faded, uneven lighting an