Research And Imple Mentation Of ID Identification System Based On Smartphone

With the rapid development of smart phones and mobile Internet technologies,smart phones have been widely applied to people's lives and work.It is not only small,easy to carry,has high-definition camera functions,but also has a strong data processing capabilities,especially suitable for mobile office environment,which provides a solid foundation for mobile application development based on smart phones.Daily policing requires a large number of inspections of second-generation citizen ID cards to go out to work.Howeve r,traditional manual ID verification is mainly done manually by registering them on record books,and then returning to the office to enter policing through PC-side programs.In the information management system,if it is necessary to immediately check the record of the party's records,such as whether there are criminal records,etc.,it may be necessary to call the relevant business department to verify.This method is not only inefficient,but also easily leads to information loss or error.inconvenient.If you make full use of the characteristics of smart phones,develop a resident ID card information identification system,take photos and identify ID cards anytime,anywhere,and connect to the police information management system in real time to comple te the remote verification and input of ID information.It will bring great convenience to policing office work,but it will also increase the accuracy of verification or entry and improve work efficiency.Entrusted by a company,we developed a third-generation police information management system for a public security bureau in Guilin,Guangxi.The use of smart phones to identify identity card information is an important subsystem in the entire system.This article focuses on smart phones.The key technologies and algorithms involved in the development of ID information identification systems such as image preprocessing,ID card character recognition,and system scheme design have been studied and implemented.The main work is divided into the following three areas:1.Image preprocessing.In daily policing on the go out of office,using a mobile phone to take pictures of ID cards is often affected by factors such as insufficient illumination,tilted shooting angles,and arm shaking.This results in unclear photo shooting and results in poor recognition rates.Firstly,this paper uses Bemsen binarization method and mean-filtering noise reduction method to better reflect some necessary characteristics of ID image,which is conducive to the extractio n of text information and improves the image pre-speed;Then,the Canny edge detection algorithm is used to extract the ID card contours.Finally,the tilt correction of the ID photo is performed using a hybrid algorithm based on Hough straight line detect ion and perspective transformation,so that the subsequent character recognition accuracy rate is well guaranteed;A guide box is also set through the camera preview interface of the camera to guide the user to photograph the ID card,which can effectively assist the user in photographing the ID card.2.ID card character recognition.The accuracy of character recognition for the original Tesseract-OCR engine is poor,and it can not meet the needs of ID identification in practical applications.Based on the Tesseract-OCR and combined with the characteristics of ID cards,we have adopted various methods to optimize it:(1)Improve the accuracy of C hinese character recognition of address information by adding commonly used character dictionary files such as province name,city,county,village,group,and number;(2)Using the coordinate point location algorithm to cut the block where the ID text information is located and block recognition,effectively reducing the recognition range and increasing the recognit ion speed;(3)Training ID cards containing only digits 0-9 and the letter X The number-specific character library saves the storage space of the mobile phone and improves the identification accuracy of the ID number;(4)The identity of the entire identification number is verified by using the identity of the last digit of the ID number as the verification code.The method effectively improves the identification accuracy of the ID number;(5)By setting the Tesseract-OCR engine PSM parameters,the ID character recognition range is limited,which reduces the problem of not reading the graph;the experimental results show that through the above comprehensive optimization method,the character recognition accuracy of the ID card has been improved.Greater improvement.3.Design and implementation of identification information identification system based on smart phones.Considering the operation and processing capabilities of Android smart phones and field application scenarios,the above image preprocessing algorithms and ID card character information identification algorithms are applied and implemented in the system.The system mainly implements functions such as ID card character recognition and storage query.After testing,it can be concluded that the reco gnition rate for ID numbers reaches 98%,the recognition rate in C hinese reaches more than 92%.Basically achieved practical needs.