光学字符识别（OCR）一般是指通过将文本区域的像素模式与熟悉的学习模式进行比较，在图像采集中自动识别字母数字字符，类似于一般的物体识别（如匹配）。
深度光学字符识别的视觉功能是利用 "深度学习"（点阵字体、半字体、工业字体、手写体等），根据一些适合各种应用的预训练字体来读取和解释文本。因此，不需要额外的训练就能达到很高的识别率。预先训练好的深度学习网络存储在相机中。
视觉功能 深度 OCR 设计用于单个单词。
信息：
为了获得高质量的待识别字符，事实证明在图像采集时字符宽度为 8 像素是非常有用的。此外，ROI 周围的最小静区也是必要的，并且必须保持空闲。
有关深度 OCR 寄存器的概述，请参阅寄存器概述部分。
信息：
有关视觉功能用例，请参阅深度 OCR - 读取长字符串。
本节主题
Optical character recognition (OCR) generally refers to automated alphanumeric character recognition within image acquisitions by comparing the pixel patterns of the text areas with familiar learned patterns, analogous to general object recognition (such as in Matching).
Vision function Deep OCR reads and interprets texts based on a number of pre-trained fonts suitable for a wide variety of applications using "deep learning" (dot-matrix fonts, semi-fonts, industrial fonts, handwriting, etc.). It is therefore possible to achieve very high recognition rates without additional training. The pre-trained deep learning network is stored on the camera.
Vision function Deep OCR is designed to be applied to a single word.
Information:
For good quality of the characters to be recognized, a character width of 8 pixels in image acquisition has proven useful. A minimum quiet zone around the ROI is also necessary and must be kept free.
For an overview of Deep OCR registers, see section Overview of registers.
Information:
For vision function use cases, see Deep OCR - Reading long character strings.