本文本主要对图片文字提取展开研究,首先读取图片进行预处理;然后针对网格特征和方向特征对图片文字进行特征提取;为了提高识别系统的可靠性,采用多分类器集成方法,即通过多个互补的分类器来改善单个分类器的性能.%On this paper,the research on the main of text extraction,first reads the image preprocessing; Then in view of the grid features and direction of image text feature extraction; In order to improve the reliability of the recog-nition system,using the integration method of multiple classifiers by multiple complementary classifier to improve the performance of a single classifier.
展开▼