Font Size: a A A

Archery Action Stage Division Based On Data Mining Technology

Posted on:2020-06-17Degree:MasterType:Thesis
Country:ChinaCandidate:S H FuFull Text:PDF
GTID:2417330572499232Subject:Sports training
Abstract/Summary:PDF Full Text Request
Objective:To analyze the kinematics of archery movement from two dimensions: the bow(vertical plane)and the pull string(horizontal plane).A low-precision kinematics acquisition and analysis program is proposed to assist archery practice.On the basis of perfecting the existing theory of division of action phases,the hypothesis that there are two different methods of pulling is proposed.The classification model was established to verify the hypothesis,and the differences of different pulling techniques in the aiming process were further compared,and the technical characteristics of different pulling movements were explored.Methods : Using literature method,expert interview method,experimental method,mathematical statistics and data mining method,10 national first-class archery athletes of Shanxi archery team were studied(4 male and 6 female),using VICON,Basler and other equipment to collect Kinematic data on complete archery movements.The top three performances were selected,and the horizontal and vertical data mining of archery movements was studied by the method of equalization discretization,cluster analysis and SMO classification analysis.Results:(1)The athlete's own movements show good consistency.The pull distance was significantly correlated with height(P=.915*),and the height of the bow was less correlated with height(P=.624).(2)The 10-segment equidistant segmentation enables the naked eye to recognize the archery action segmentation without substantially losing the action details.(3)The DBSCAN clustering method is more accurate than the existing calculation method for the recognition of the bowing motion.(4)The pulling bow technique can be divided into two types: the stable pulling method(SP method)and the phased pulling method(PP method).The 10-fold cross-check results of the classification model show that TP=6,TN=3,FP=1,and FN=0.Kappa statistic=0.7826 is between 0.61 and 0.80,indicating that the model has high universality.(5)The arrow line in the pulling phase is close to the guide line in a clockwise direction,and the aiming process of the two open bow methods is different.Conclusion:(1)The low-precision data acquisition scheme can theoretically realize the classification of the archery action stage and the classification of the bowing action.The practical application effect in the exercise training needs to be further verified.(2)There are two different pull techniques,the SP method and the PP method.The two methods have their own characteristics in terms of the bowing rhythm and the aiming process.(3)The clustering algorithm is more accurate than the existing calculation method of the bowing stage in recognizing the starting and ending position of the bowing.(4)The classification hypothesis of the bowing motion is verified by a good mathematical model.In the analysis of the bowing motion,the pull technique used by the subject should be considered.(5)The aiming process of the two methods of pulling the bow is different.The aiming stability of the SP method is better,and the aiming adjustment of the PP method is faster.(6)SP pulling method requires higher athletes' upper limb stability and endurance,but the action completion time is shorter.The PP open bow method requires higher explosive force and quick aiming ability of the upper limbs of the athletes.It is recommended to speed up the bowing speed and increase the bow movement range to reduce the aiming burden.
Keywords/Search Tags:data mining, action phase division, archery, pulling
PDF Full Text Request
Related items