What is OCR technology?
Optical Character Recognition (English: Optical Character Recognition, OCR) refers to the process of analyzing and recognizing image files of text materials to obtain text and layout information.
Similar to image recognition and machine vision technology, the processing process of OCR technology is also divided into input, pre-processing, mid-term processing, post-processing and output process.
enter
For different image formats, there are different storage formats and different compression methods. Currently, there are OpenCV, CxImage, etc.
Pre-processing – binarization
Most of the pictures taken by digital cameras today are color images, which contain a huge amount of information and are not suitable for OCR technology.
For the content of the picture, we can simply divide it into foreground and background. In order to make the computer faster and better perform OCR related calculations, we need to process the color image first, so that only the foreground information and background information remain in the picture. Binarization can also be simply understood as “black and white”.
image noise reduction
For different images, the definition of noise may be different, and the process of denoising according to the characteristics of noise is called noise reduction.
tilt correction
Because ordinary users, when taking pictures of documents, it is difficult to shoot completely in line with horizontal and vertical alignment, so the pictures taken will inevitably be skewed, which requires image processing software to correct.
Mid-term processing – layout analysis
The process of dividing document pictures into paragraphs and branches is called layout analysis. Due to the diversity and complexity of actual documents, this step still needs to be optimized.
character cutting
Due to the limitations of photographing and writing conditions, characters are often stuck and pens are broken. Directly using such images for OCR analysis will greatly limit OCR performance. Therefore, character segmentation is required, that is, to separate different characters.
Character recognition
In the early stage, template matching was mainly used, and in the later stage, feature extraction was mainly used. Due to the influence of factors such as text displacement, stroke thickness, broken pen, adhesion, rotation, etc., the difficulty of feature extraction is greatly affected.
Layout restoration
People hope that the recognized text is still arranged like the original document picture, and the paragraphs, positions, and order are output to Word documents, PDF documents, etc., and this process is called layout restoration.
post processing
According to the relationship of specific language context, the recognition result is corrected.
output
Output the recognized characters as text in a certain format.
What are the applications of handheld terminals based on OCR technology?
Through the handheld terminal PDA loaded with OCR character recognition software, many scene applications can be realized, such as: car license plate recognition, container number recognition, imported beef and mutton weight label recognition, passport machine-readable area recognition, electric meter reading recognition, steel coil Recognition of sprayed characters.
Post time: Nov-16-2022