跳到主要内容Skip to main content

OCR(光学字符识别)OCR (Optical Character Recognition)

OCR (Optical Character Recognition)

  • 图像语音视频Image, speech & video
  • 第三站 · 用深Station 3 · Going Deeper
  • 引用核验 2026-11-17refs checked 2026-11-17
解决什么问题What problem it solves

把图片或扫描件里的文字识别成可编辑、可检索的文本。Turn text in images and scans into editable, searchable text.

这是什么What it is

OCR 通常包括版面分析、文字检测、字符识别和后处理。

现代系统还要处理表格、多栏、手写体和旋转,识别结果常需置信度与人工复核。

OCR usually includes layout analysis, text detection, character recognition and post-processing.

Modern systems also handle tables, multiple columns, handwriting and rotation; results often need confidence scores and human review.

OCR is a pipeline: layout analysis finds regions, detection locates text, recognition reads characters, post-processing fixes common errors.
OCR 不是一步:版面分析定位区域,检测框出文字,识别成字符,后处理修正常见错误。OCR is a pipeline: layout analysis finds regions, detection locates text, recognition reads characters, post-processing fixes common errors.
为什么需要知道Why it matters

RAG 处理纸质档案时,OCR 是第一步;识别错会直接变成检索错和答案错。When RAG ingests paper archives, OCR is step one. Recognition errors become retrieval errors and then answer errors.

别名Chinese aliases

光学字符识别 · 文字识别 · OCR 识别

出处Sources
  • Tesseract · 官方文档Tesseract documentation ↗
  • Tesseract · 官方仓库Tesseract repository ↗