3D动画图像中非标记运动运动目标捕获方法研究  

Research on non-marking moving motion target acquisition method in 3D animated image

作　　者：姬景聚[1] JI Jing ju(Xi’an University of Posts Telecommunications,Xi’an 710061,China)

机构地区：[1]西安邮电大学,陕西西安710061

出　　处：《电子设计工程》2020年第7期127-130,136,共5页Electronic Design Engineering

基　　金：西安邮电大学教改项目(301/312011906)。

摘　　要：传统的运动目标捕获方法在捕获运动目标时,未考虑背景景物的位置差异对3D动画图像中运动目标的影响,导致捕获结果准确性低。为了解决这一问题,研究了一种新的3D动画图像中非标记运动运动目标捕获方法。分割3D动画图像中非标记运动运动目标,将序列分成目标层、过渡层和背景层,对图像运动场的各个运动参数进行估算,得到层次运动模型,从而实现有效分割。建立层次化运动模型分割像素点,设定虚拟人,利用序列法使虚拟人完成3D动画图像中非标记运动运动目标的捕获。实验结果表明,提出方法能够很好地解决具有多组帧间全局运动参数的3D动画图像运动目标捕获的问题,较大程度提高了捕获结果的准确性,并加强了方法的鲁棒性。Traditional methods of moving object acquisition do not consider the influence of the position difference of background scenes on moving objects in 3 D animation images,which results in low accuracy of acquisition results.In order to solve this problem,a new method of capturing unmarked moving objects in 3 D animation images is studied.Segmenting the unmarked moving objects in 3 D animation images,dividing the sequence into target layer,transition layer and background layer,estimating the motion parameters of the image motion field,and obtaining the hierarchical motion model,so as to achieve effective segmentation.A hierarchical motion model is established to segment the pixels,and a virtual human is set up.The sequence method is used to capture the unmarked moving objects in 3 D animation images.The experimental results show that the proposed method can solve the problem of capturing moving objects in 3 D animation images with multi-group global motion parameters,greatly improve the accuracy of the captured results and enhance the robustness of the method.

关 键 词：3D动画图像 非标记运动 运动目标 目标捕获方法 

分 类 号：TN911.73[电子电信—通信与信息系统]