Research On The Image Restoration On BP Neural Network

With the development of science technology,image is one of the usual multimedia data which is enlarged at an explosive speed,It's important in traffic management and public safety,images making vehicle information on traffic flow and forensics fields,and which is one of important way and evidence on forensics.Image information is very important in judicial field.But for more reasons on internal and external factors in the process of image acquisition in video surveillance which to cause image blurring and reduction of information,and which greatly frustrate for its use in the process of law enforcement of public security.With safety needs promote on people in national and social development,fuzzy image restoration has become hot issues of social concern.With the progress of science technology,its study on image restoration has entered a stage of rapid development;traditional single degradation model on image restoration which do not meet the needs of social value on the field,and with progress of machine language and artificial intelligence algorithm,combination of image restoration and artificial intelligence has become trend of development in the field.BP neural network can be organized oneself of study in the input data under ideal state,which have great maneuverability since the simplicity of the BP neural network,and which can find inherent laws in the sample data entered into the summary.The advantage of BP neural network for arbitrary nonlinear fitting of input and output,and the variety of degenerate functions encountered in image restoration,this paper is based on the introduction and implementation of the traditional image restoration algorithm,a research has been given on BP neural network algorithm and image restoration technology,and promotion has been achieved by BP neural network algorithm,which has been verified by experiments.