Arch Image Analysis And Its Application Research

The arch of foot is the key to support the weight of the whole human body and maintain the balance of the body,which has a great impact on sports function and physical health.The normal arch of foot can effectively cushion the impact of human body by using its physiological structure when people are engaged in various sports.For a long time,foot arch disease has been neglected because of its hidden location.However,research shows that the incidence of foot arch disease is still high among adolescents.Serious cases may affect the sports ability of adolescents,and even affect future work and life trends.However,there are many drawbacks in using traditional methods to identify the arch of foot.Therefore,it is necessary to detect the type of foot arch accurately in time so as to take early treatment measures.This paper compares the advantages and disadvantages of the commonly used detection methods and evaluation systems of arch types,and proposes a method of using arch coefficients combined with computer digital image processing to identify and judge arch types conveniently and quickly.The main contents of this paper are as follows:Firstly,it introduces the physiological structure and function of the arch of foot,compares several common identification methods of arch of foot and evaluation system of arch of foot,analyses their advantages and disadvantages,and finally chooses arch coefficient as evaluation criterion.After comparing and analyzing several common foot image acquisition methods,the image acquisition methods studied in this paper are introduced.Secondly,the image acquisition mode is analyzed and preprocessed,and the whole plantar image is extracted according to the characteristics of the collected image.Aiming at the problem of initial centroid sensitivity in K-means algorithm,the characteristics of image are analyzed,and the K value and initial centroid are determined.Using K-means algorithm,the whole foot image can be extracted by three iterations at most.Thirdly,the Mean Shift algorithm is studied in detail.Firstly,the RGB image is mapped to Luv feature space,and then the Luv feature space of the image is clustered by Mean Shift algorithm to complete the clustering of the full RGB image,and the plantar weighted region image is successfully extracted.Fourthly,the binary image of plantar pressure area is filled,filtered and smoothed by the principle of mathematical morphology,and the target area is extracted.Then,several groups of images are processed by the method of identifying arch coefficient.Finally,the test results are analyzed to verify the accuracy of the research data.Compared with traditional detection methods,the method studied in this paper has high accuracy,low time cost,low cost,strong generalization,and has certain improvement and promotion space.In the follow-up study,this method can be used in practical application after a little improvement.