基于散斑结构光编码的三维人脸重建方法研究  

3D FACE RECONSTRUCTION METHOD BASED ONSPECKLE STRUCTURED LIGHT CODING

作　　者：黄丰智 范鹏 范智瀚 段智涓 Huang Fengzhi;Fan Peng;Fan Zhihan;Duan Zhijuan(National Key Laboratory of Fundamental Science on Synthetic Vision,Sichuan University,Chengdu 610065,Sichuan,China)

机构地区：[1]四川大学视觉合成图形图像技术国防重点学科实验室,四川成都610065

出　　处：《计算机应用与软件》2023年第9期218-221,228,共5页Computer Applications and Software

基　　金：四川省科技创新苗子工程项目(20-YCG045)。

摘　　要：三维人脸信息获取通过三维相机和三维人脸重建算法进行三维人脸建模。真实环境中三维重建方法往往存在匹配误差较高、表面粗糙等问题。针对三维人脸重建中的问题设计一套三维人脸重建算法用以提高三维重建效果,通过向人脸投射散斑图案增强人脸纹理,改进ZNCC算法生成平滑的三维人脸模型。通过仿真实验验证了该算法可以得到较高精度的平滑三维人脸模型。Three-dimensional face information acquisition used a three-dimensional camera and a three-dimensional face reconstruction algorithm for three-dimensional face modeling.Three-dimensional reconstruction methods in real environments often have problems such as high matching errors and rough surfaces.Aimed at the problems in 3D face reconstruction,a set of 3D face reconstruction algorithms is designed to improve the effect of 3D reconstruction.The face texture was enhanced by projecting speckle patterns on the face,and the ZNCC algorithm was improved to generate a smooth 3D face model.Simulation experiments verify that the algorithm can obtain a smooth three-dimensional face model with higher accuracy.

关 键 词：散斑编码 三维重建 立体视觉 三维人脸相机 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]