The Research And Implementation Of ID Card Image Information Extraction Based On OPENCV

Identification has come true with the development of the technology of Computer-Image and word-processing.The ID card is a symbol of citizen hood.It is presented to the citizen by the government.It is also regarded as the only thing that certificates the citizenship.It plays an important role in many occasions in our daily lives.Identification system has been widely used in each field,such as the public security,the banks,and the transport.Compared to manual work,this system is more accurate and more efficient.But in some public institutions which haven't been authorized by the public security organs,ID numbers have to be typed into the computers by manpower.The work efficiency is so low and it is easier to make mistakes.Therefore,there will be extensive application value to carry out ID information extraction that is based on image processing and analysis.The second-generation ID cards are made up of specific combinational codes,with 17 digital ontology codes and a check code.It separately represents a native place,the date of birth,3 numbers and a check code.This thesis will mainly use Open CV to do some ID information processing and analysis research.It includes three aspects: image preprocessing,image segmentation,character segmentation and character recognition.And then based on the researches,it will form ID information extraction system prototype.1.To solve the brightness problem: too dark or too light,the system will process the images with gray scaling,smoothing and gray stretching to ensure the improvement of the subsequent recognition accuracy.2.To follow the special layout structure,the system will use morphology methods to search the connected areas.Then it will divide the connected areas by using contour extraction and mark symbols to finish the regionalism of the ID cards.And then we will preprocess the regionalism of the ID cards with gray scaling,binary and so on to separate the characters.But the separated characters have different sizes.We have to normalize them to form appointed templates so that they can improve the stability of the system and decrease the noise of the image.3.To get the ID numbers and the information on the image.The system will study the identification of the ID numbers with the template matching.First,we will form a character warehouse according to the features of the characters and then by the template matching,we will normalize the characters and the templates to identify the characters.There are only 11 numbers on the ID and the size is steady,so it is very practical to use the template matching.The study has shown that it is right to use this way.4.As we mentioned,The ID card information extraction system based on OPENCV mainly include the Image preprocessing,Character Segmentation and Template matching.