引入harris角点的体操动作动态三维图像构建  

Dynamic 3D image construction of gymnastic movement based on Harris corner

作　　者：简妮娜 Jian Nina(Zhujiang College,South China Agricultural University GuangdongGuangzhou 510900)

机构地区：[1]华南农业大学珠江学院,广东广州510900

出　　处：《现代科学仪器》2021年第5期220-223,共4页Modern Scientific Instruments

摘　　要：通过对真实人物动作特征的抓取和分析,可以提高三维虚拟人物动作的仿真度和识别度,为此提出了一种基于三维虚拟动态成像的动作捕捉技术。通过立体空间立体扫描识别技术对真实人物的动作进行动态图像的特征抓取,基于AR虚拟现实技术对所获取动态图像的轮廓进行分割,将所有动态特征体现在直方图中,同时引入harris角点检测扫描技术捕捉真实人物的体操动作,以此来构建动态三维虚拟动作图像。实验结果显示,所提出的方法能够在捕捉图像中实现动态特征的高度匹配,具有较强的识别能力和虚拟图像构建能力。By capturing and analyzing the motion features of real characters,the simulation and recognition of 3D virtual characters can be improved.Therefore,a motion capture technology based on 3D virtual dynamic imaging is proposed.Through the stereo scanning recognition technology to capture the dynamic image features of the real person's action,the contour of the dynamic image is segmented based on AR virtual reality technology,and all the dynamic features are reflected in the histogram.At the same time,Harris corner detection scanning technology is introduced to capture the real person's action,so as to construct the dynamic three-dimensional virtual action image.The experimental results show that the proposed method can achieve high matching of dynamic features in captured images,and has strong recognition ability and virtual image construction ability.

关 键 词：动态成像 动作捕捉 角点扫描 轮廓分割 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]