Research And Implementation Of Automatic Identification Of Id Card On The Android Platform Based On C/S Mode

As the development of artificial intelligence technology,automatic and accurate access to information on ID cards has become increasingly demanding in the public security,railway,banking industries and so on.With the advent of the mobile Internet era and the realization of high-performance mobile devices,smart terminals have become a platform for comprehensive transaction processing,making identification of mobile ID cards possible,which greatly improves the convenience of identification.Based on this,this paper realizes the automatic identification of ID cards on the Android platform through image processing and pattern recognition.The main work of this article is divided into the following four parts:(1)Character image segmentation.First of all,the image which obtained by the Android platform and meets the requirements will be processed by gray processing,scaling,adaptive threshold Canny edge detection,probability Hough transform to find lines and perspective transformation segmentation for a distortionless,no background ID image.Secondly,the rectangular area image and the character image containing the required information are separated using the proportional coordinates and the projection method respectively.(2)Chinese character recognition based on hand-designed features.Firstly,this essay proposes the following improved feature combination method for the case of huge categories of Chinese characters: extraction of peripheral features of Chinese characters,number of stroke crossings,and characteristics of Chinese character grids.The characteristic matching is based on Euclidean distance to complete the recognition of Chinese character printing respectively.As for information areas where the training set is small,such as the gender and ID number in the ID card,only the grid features are extracted.Then the feature matching is done according to the Euclidean distance.The recognition algorithm is used on a serverless,independent Android platform.(3)Chinese character recognition based on convolutional neural network.The technical improvement of this essay is: first for the defects of the feature matching method and the situation encountered by the Chinese character image of the ID card,the Python PIL library was used to obtain FZHei image data sets containing corrosion,expansion,rotation,and noise.Then a convolutional neural network was used to design the network.Finally a model with a test accuracy of 99.8% was trained on the TensorFlow platform,which could be used on the server side.(4)system implementation.First,completing the development of the three major algorithms of character area positioning,character segmentation,and character recognition.Then,by usingthe JNI mechanism in android,the program is ported to Android,and the image acquisition and ID card information display are completed.Finally an Android platform application is completed based on C/S mode.