Design And Implementation Of ID Card Identification System Based On Android Platform

With the development of the times,identity card information is widely used in life.Therefore,how to efficiently enter ID information to ensure personal information security has become a research focus.In response to this problem,this article will design and implement an ID identification system based on the Android platform,mainly to locate and extract the ID card area,information area,character recognition,face recognition and other algorithms.First of all,for the ID card area with background,tilt or reversal and other issues,a method of localization and identification of ID cards is proposed.Using the Android system,an image acquisition interface was designed to add horizontal lines,touch screen focus,screens hots and other functions.The use of image processing knowledge,combined with the layout of the ID card layout,edge detection of the image,to find the maximum contour,external minimum enclosing rectangle processing,etc.,detect the tilt angle,and correct the ID card area.The method can accurately locate and extract the ID card area,and is suitable for locating and extracting the ID card information area.Secondly,for the ID address area,column cut can not get the complete single character,and the method of square word segmentation based on pixel jump is studied.Using image processing knowledge,the ID card information area is denoised,binarized,etc.,and the initial segmentation of the character is performed according to the transition of black and white pixels,and the character is again split or merged according to the character's inherent characteristics.In order to improve the character recognition rate,use the character's inherent characteristics and statistical characteristics to achieve pixel matching,mesh feature recognition characters,and use jTessBoxEditor to train different fonts for different information areas of ID card.Finally,the face recognition method based on block LBP-CNN is designed for the problem that the identity card and the licensee are the same person.The face detector is used to locate and extract the face area,and the geometric normalization and other processing are performed.The facet LBP feature and the entire CNN feature cascade areused as face identification to extract the identity card and the license respectively.Face features,the final use of Euclidean distance to measure their similarity,to determine whether the identity card and the holder is the same person.Experiments show that this method is better for recognizing human faces.