The Design And Implementation Of Speed Skating Data Measure And Analysing System

This paper introduces a software from which we can extract motionparameter,and it can contribute to speed skating program.this software consists ofseveral parts as below. Image registration sub-system, Camera measure sub-system.Motion tracking sub-system,Motion data analysing sub-system.We providerequirement analysis,overall designing,Detail designing.of above parts.and show theway and the reason we choose this way to solve to every crusial point in the system.In image registration part,the procedure to handle this tough problem is asfollows:firstly,find corners point from image.Secondly, Try to match similar cornerpoints in two images.this step we can call it similarity-match. The last step is toevaluate projecting matrix between two images.Here,We use Ransac to get this matrix.In order to minimize the error brought by image registeration.We insert sets oftemplate.Meanwhile, standard frame whose function is to hold all frames in the videolike a global scene brought in.After we set the standard frame,we can project allframes to this standard frame.According to the camera calibration theory,Lots of objects which used tocalculated the relationship between image and ice-ground are settled in theice-ground.Then Tracking result will be projected into ice-ground space.Therefore,wecan use athlete’s location of tracking result to get motion data.In athlete tracking sub-system,sparse representation method is used to recognizeand track the target in every image. sparse representation used in this system is basedon paticle filter which is used to solve non-linear problem, framework.lots of particlesaround target needed to sampled used Gaussian model.Then this paper contains a setof templates.a number of templates to represent occulaton.Reconstruction errorneeded to get from using1-regularized least squares problem.Then we chooseminimum reconstruction error to be the tracking result.Next,tracking will continueaccording to the status-transfer model of particle filter.the weight of every particleneed to be reevaluated to renew the set of particles used in the trackingproblem.Lastly,the set of templates need to be updated to satisfy that they canrepresent the tracking result.every template has its weight in order to measure itsimportance.the template which has minimum weight needed to be discarded.Lots of race and athlete or other kinds of information can stored in this system which Above information maintenance is also been done in.Meanwhile,we cansearch motion data after given items we cared about.video can be recapped after wefind spedified race.Also,motion data can be extracted to show.