Optical Character Recognition Techniques in Documents

摘要

Vehicles and other items often have corresponding documents, such as registration cards, that contain a significant amount of informative text information that can be used to identify the item. Typical OCR may not be successful when processing non-cooperative images. Thus, functions such as dewarping, text alignment, and line identification and removal can help OCR of non-cooperative images. Dewarping involves processing the image to determine the curvature of the document depicted in the image and to dewarp the image of the document to more precisely conform to the ideal of a cooperative image. Text alignment involves determining the actual alignment of the depicted text even if the depicted text is not aligned with the depicted visual cue. Line identification and removal involves identifying the portion of the image depicting lines and removing those lines prior to OCR processing of the image.