Research On Shoeprint Matching Algorithm Based On Overall Feature

With the rapid development of science and technology,machine vision has been developed rapidly in various fields,especially in the criminal investigation,which has also been fully developed.However,the crime means of criminal and anti-detection ability is also growing,which the scene of evidence collection and testing work has a higher demand for the scene of evidence collection and testing work.The shoe image automatic identification technology is the use of computers on the crime scene left behind the shoe of processing,analysis and matching,to help to detect the case.In this paper,machine vision is applied to footprint recognition and in-depth research,the main research work is as follows:Firstly,according to different angles of shoeprints shooted,a method of secondarily rotary positioning is put forward,to place shoes in a unified position,and to extract outer contour features for subsequent shoe.Secondly,on the basis of extract the feature extracted,this paper presents a quadrilateral principle algorithm to simplify the external contour lines of the shoes,which puts up the differences between this algorithm.Thirdly,the paper uses local two value model(Local Binary Patterns,LBP),and extended algorithm model,to match internal texture.The improved LBP operator has rotational invariance,and its dimension is low and the matching speed is fastFourthly,combined with the shoeprints,based on the contour and texture combination of signs,the paper proposes a matching algorithm on basis of the characteristics of the whole shoe.The extracted features make the matching owe better foundation,and the result more reliable.The final experimental results shows that the overall features of the Shoeprint matching is well robus,simple and quick,and has high recognition rate.The validity is verified.