|
|
- 2016
陀螺仪在基于Hash结构的三维重建中的应用
|
Abstract:
针对3D重建过程中出现深度相机突然运动, 导致相机无法正确跟踪重建场景, 提出了一种结合陀螺仪与基于Hash结构的KinectFusion 3D重建的方法. 将陀螺仪得到的位姿与ICP算法得到的位姿通过扩展卡尔曼滤波器(EKF)数据融合, 获得更加精确的位姿, 改进了原始算法的3D重建效果. 实验测试结果表明, 该算法可以得到更加精确的系统位姿预测模型, 在相机突然运动时仍然能够实现有效跟踪.
In order to track camera and reconstruct scene correctly when depth camera moves suddenly leading to a failure of tracking reconstructed scene，a new 3D reconstruction method，which combines the gyroscope with the KinectFusion algorithm based on Hash structure，is proposed. A more precise pose can be obtained by a data fusion of the extended Kalman filter which integrates the pose data from gyroscope and that from the ICP algorithm，hence improving the effect of 3D reconstruction. Experimental tests show that this algorithm can obtain more accurate predictive model of the system pose and track camera effectively when depth camera moves suddenly
| [1] | Chen Y, Medioni G. Object modelling by registration of multiple range images[J]. <i>Image Vision Computing</i>, 1992, 10(3):145-155. |
| [2] | Nguyen T T, Mentré F. Evaluation of the Fisher information matrix in nonlinear mixed effect models using adaptive Gaussian quadrature[J]. <i>Computational Statistics & Data Analysis</i>, 2014, 80(4):57-69. |
| [3] | Ljung L. Asymptotic behavior of the extended Kalman filter as a parameter estimator for linear systems[J]. <i>IEEE Transactions on Automatic Control</i>, 1979, 24(1):36-50. |
| [4] | Bajaj C L, Bernardini F, Xu G. Automatic reconstruction of surfaces and scalar fields from 3D scans[C]//<i>Conference of Computer Graphics and Interactive Techniques.</i> Los Angeles, USA:ACM, 1995: |
| [5] | Caron F, Duflos E, Pomorski D, et al. GPS/IMU data fusion using multisensor Kalman filtering:Introduction of contextual aspects[J]. <i>Information Fusion</i>, 2006, 7(2):221-230. |
| [6] | Curless B, Levoy M. A volumetric method for building complex models from range images[C]//<i>Conference of Computer Graphics and Interactive Techniques.</i> New York, USA:ACM, 1996:303-312. |
| [7] | Izadi S, Kim D, Hilliges O, et al. Kinectfusion:Real-time 3D reconstruction and interaction using a moving depth camera[C]// <i>User Interface Software and Technology.</i> USA:IEEE, 2011:559-568. |
| [8] | Nie?ner M, Zollh?fer M, Izadi S, et al. Real-time 3d reconstruction at scale using voxel hashing[J]. <i>ACM Transactions on Graphics</i>(<i>TOG</i>), 2013, 32(6):169. |
| [9] | Censi A. An accurate closed-form estimate of ICP’s covariance[C]// <i>IEEE International Conference on Robotics and Automation</i>. Italy, 2007:3167-3172. |
| [10] | Freedman B, Shpunt A, Machline M, et al. Depth Mapping Using Projected Patterns:US 8150142[P]. 2012-04-03. |
| [11] | Newcombe R A, Davison A J, Izadi S, et al. KinectFusion:Real-time dense surface mapping and tracking [C]//<i>International Symposium on Mixed and Augmented Reality.</i> USA:IEEE, 2011:127-136. |
| [12] | Low K. Linear least-squares optimization for point-to-plane ICP surface registration[R]. Charlotte:University of North Carolina, 2004. |
| [13] | Zeng M, Zhao F, Zheng J, et al. Octree-based fusion for real-time 3D reconstruction[J]. <i>Graphical Models</i>, 2013, 75(3):126-136. |
| [14] | Hoppe H, Derose T, Duchamp T, et al. Surface reconstruction from unorganized points[J]. <i>Computer Graphics</i>, 1992, 26(2):71-78. |
| [15] | 109-118. |
