基于神经网络的数字识别系统的研究与设计-计算机技术专业论文.docxVIP

AbstractAbstract Abstract Abstract The digital identification algorithm has been paid much attention in recent years， and it develops very quickly,the research direction has very high application value．In life，the digitai identification technology has a wide range of applications，such as： vehicle license plate recognition，notes and other related numbers call be realized． The artificial neural network is all important branch of artificial intelligence research，the artificial neural network can approximate any continuous nonlinear characteristics of unknown function and powerful parallel computing and generalization ability,to achieve digital identification problem solving，you need a valid pattem recognition tools．Neural network technology is all important way to simulate human intelligence and simulate human thinking from the model of human brain system． In this paper,the identification of the application process to collect the corresponding digital information，after the system identification into the computer Can identify the number,SO that the completion of the system output．After a large amount of sample training，the recognition of the digital object is extracted as the input of the system，SO as to realize the recognition of the digital object． Firstly,this paper introduces the research background of digital recognition，and the research status at home and abroad，and describes the application prospects of digital identification，and points out the difficulties of digital identification；In addition，we analyzed the pretreatment method of digital identification，including binarization processing and image de-noising and normalization，etc；Finally,the structure of BP neural network is analyzed，and the digital recognition method of BP neural network is used to solve the constraint and complexity of the previous digital identification method．In order to determine the weight initialization function and the optimal number of hidden nodes，using feature extraction technology,S