Dissertation > Excellent graduate degree dissertation topics show

Based on the three-dimensional image reconstruction algorithm sequence

Author: YangJiaFei
Tutor: SuiLianSheng
School: Xi'an University of Technology
Course: Applied Computer Technology
Keywords: Computer Vision Three-dimensional reconstruction Image sequence
CLC: TP391.41
Type: Master's thesis
Year: 2009
Downloads: 133
Quote: 0
Read: Download Dissertation

Abstract


Three-dimensional reconstruction is based on the calculation of the camera image information acquired objects in three-dimensional space geometry , thereby identifying the object and reconstruction process . It is the inverse problem of the imaging process , the input image is a three -dimensional geometric objects , light , surface material properties , the object color , camera parameters of the nonlinear function of other factors , to solve the three-dimensional structure model of the object . The solution of the problem is not unique, but the noise and discretization errors caused more sensitive, so quite difficult to solve . Three-dimensional reconstruction , including feature extraction , projective reconstruction and rebuilding three -part study scales . First, this thesis three-dimensional reconstruction of image sequences projective reconstruction algorithm, iterative factorization method and the Bundle Adjustment algorithm ( bundle adjustment method ) , found by experiment projective reconstruction algorithm uses five iterations factorization method and 10 Bundle Adjustment can achieve reconstruction requirements. Then , the paper studied the scale reconstruction algorithms algorithms implemented based on the absolute conic dual scale reconstruction algorithm is proposed based on the scale of reconstruction than the average distance assessment method, based on a large number of experiments found that the dual absolute conic Lu stick better. As for the camera calibration scale reconstruction of the dual absolute conic constraints , this project also studied camera calibration algorithm , a new sphere image sequence based camera calibration method . The calibration method compared to the traditional sphere calibration algorithm has a very distinct advantage . In the calibration results , the traditional sphere calibration algorithm for solving the three scaling factor, n calibration images produced 3n-1 , p.315 ambiguity , need to manually filter out the calibration results ; while the new calibration method uses a continuous sequence of images feature eliminates scaling factor of multiple solutions for solving the unique calibration results , without human intervention. The number of operations , due to the ambiguity of scaling factors , conventional calibration algorithm ball zoom factors of all possible combinations of calibration attempt , n calibration images attempt to 3n-1 , practical difference ; the new calibration method MOVES scaling factor ambiguity , one-time calibration results obtained , which greatly improved the sphere calibration algorithm performance.

Related Dissertations

  1. Three-Dimensional Numerical Modelling for the Head and the Electric Field Analysis of DBS,R742.5
  2. The Three-dimensional Structure Research of Facial Modiolus,R782.2
  3. Three-dimensional reconstruction and cochlear basement modal Numerical Simulation,R764
  4. Three Dimensional Surface Reconstruction Based on Image Sequence,TP391.41
  5. Design of Medical Image Three-Dimensional Visualization System Based on Windows,TP391.41
  6. The Study of Bronchial Anatomy of Right Lung with 16-slice Spiral CT,R734.2
  7. The Study on the Vestibular 3d Reconstruction of Guinea Pig and the Numerical Simulation of Temperature Field,R764
  8. Multi-view 3D Reconstruction and Application Based on Depth Fusion and Surface Evolution,TP391.41
  9. Research and Implementation of the 3D Reconstruction Based on Surface Structured Light Method,TP391.41
  10. Research on the Computer-Aided Surgery Planning for Lacrimal Passage,TP391.7
  11. An Iterative Method of Modeling Coronary Artery and Its Applications,R312
  12. Based on photometric stereo fine surface characteristics of three-dimensional reconstruction,TP391.41
  13. Multi- dimensional image based aviation city buildings Recognition Technology,TP391.41
  14. Based on the visual system of unmanned autonomous helicopter landing,V279
  15. Construction of three-dimensional graphics based on two-dimensional architectural design and implementation of virtual scene,TP391.41
  16. Stereoscopic three-dimensional reconstruction -related technology research and implementation,TP391.41
  17. Real-time modeling spatial augmented reality technology research related,TP391.41
  18. Researches on Key Technologies about Three-dimensional Reconstruction of Buildings Based on Omni-directional Image and Remote-sensing Image,TP391.41
  19. Study of Normal Diaphragm and Diaphragmatic Hernia on CT MPR,R816.4
  20. Three-dimensional computer reconstruction in narrow cranial disease diagnosis and surgery,R726.2
  21. Computerized 3D Reconstruction of Cat Optic Nerve Using Serial Sections,R338

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device
© 2012 www.DissertationTopic.Net  Mobile