基于图像序列的三维人体建模方法研究  

Modeling of 3-D Mannequin Based on Image Sequence

作　　者：陈利珍[1] 邓中民[1] 

机构地区：[1]武汉纺织大学纺织学院,湖北武汉430200

出　　处：《针织工业》2013年第1期54-56,共3页Knitting Industries

基　　金：湖北省重大项目(C2010036)

摘　　要：通过立体视觉测量法对人台模型上的特征点进行二维测量,再根据双目视觉原理对测得的二维数据进行三维转换,然后通过插值算法获得更多的数据,使人体表面的数据信息相对来说更加光滑。然后结合个性化人体的正、侧面图像,对比转换成个性化人体的三维数据信息,从而建立个性化三维人体模型。结果表明:这种方法能够方便、快速建立所需要的人体模型,并且所建立的模型也比较接近于真实人体的图像。In this article, the method measured the model set feature points for the measuring method of stereoscopic vision, then convert the 2-D data to 3-D data according to the binocular principle of vision measurement. Then through the interpolation algorithm for more data information, to make the surface data information of model smoother. Then combined them with the front and side images of personalized body, compared to 3-D data infor- mation of personalized body, and to establish the personalized 3-D model of the human body. The results show that this method can easily and quickly build up the necessary human model, and that the model is relatively close to the real human body image.

关 键 词：立体视觉测量法 网格小平面法 三维人体建模 图像 

分 类 号：TS941.26[轻工技术与工程—服装设计与工程]