Application And Research Of Motion Capture Method Of Nonrigid Object

With the development of film,game,animation and other industries,motion capture technology has become a popular research field in computer vision in recent years,and promoted the development of computer vision from 2D to 3D.However,most of the current motion capture technologies rely on expensive sensing devices,which fails to achieve the motion capture of non-rigid object,which greatly limits the application and popularization of the technology.In order to realize the non-rigid object motion capture,two calibrated high-speed cameras are used to obtain 3D moving trajectory.Then the motion state of the object is reconstructed.And the reconstruction result is used to calculate the important parameters of the non-rigid objects,such as velocity,acceleration and the change of curvature.According to the different characteristics of the non-rigid object motion,different 3D motion capture methods are proposed.A practical 3D motion capture method of non-rigid object with point feature is proposed for the non-rigid object that cannot extract feature point.The method includes image spatial reconstruction,temporal series reconstruction and other steps.The spatial reconstruction and temporal series reconstruction are the core of 3D motion recovery.In spatial reconstruction,the coordinates of the points on the image are obtained by using ellipse fitting method,and the matching points are searched according to the Mahalanobis distance.Then,triangulation is used to calculate the 3D points.In the aspect of temporal series reconstruction,the searching method is used to match the coordinates of the points between sequential images.So,the 3D motion recovery of the movement process is realized.For the non-rigid object with obvious feature points,a motion capture method for non-rigid object based on optical flow is proposed.First of all,the image is processed by blur,then the corner points are detected for the left and right images.The detected corners are tracked by using optical flow and matching from right to left images.Finally,the motion recovery is realized and the motion capture of the non-rigid object is completed.The experimental result shows that the method improves the accuracy of the reconstruction.What's more,the time of the whole motion capture process is reduced.Through the reconstruction of the target,the 3D motion data of the object is obtained accurately.