Dissertation > Excellent graduate degree dissertation topics show

Spatial Temporal Refinement on Depth Video Processing

Author: ChenGuoZuo
Tutor: DuZuo
School: Zhejiang University
Course: Information and Communication Engineering
Keywords: 3D Stereo matching Time Consistency Affine transformation
CLC: TP391.41
Type: Master's thesis
Year: 2012
Downloads: 28
Quote: 0
Read: Download Dissertation

Abstract


In recent years, with the development of technology , the traditional two-dimensional display technology is more and more difficult to meet the people's daily life , the needs of the entertainment , 3D stereoscopic display to become the next generation of display technology hotspot . The binocular parallax - based three-dimensional display technology is the most widely used in many 3D technology , the fastest growing , and there are a variety of types of products into the market . This paper first introduces the principle of a general three-dimensional display system geometry , and analysis of the different camera configurations , the parallax distribution law . Field , viewpoint synthesis of stereoscopic video encoding and 3DTV applications require accurate and smooth depth video processing, the depth video growing demand increasingly high demand in the three-dimensional display . Although binocular video stereo matching current research is still relatively small, but for a single frame image matching has been a hot research field of computer vision and difficult . Reference research scholars at home and abroad , this paper proposes a stereo matching algorithm based color segmentation of local optimization . By matching cost integrated histogram analysis , expansion and plane fitting step seed points , with the mean-shift segmentation results to ensure the accuracy and smoothness of the matching results . Middlebury standard image test , the algorithm obtained better results of the test data . Simple binocular stereo matching algorithm is applied to video processing , due to the lack of time consistency constraints , and split block plane assumption will fail in some cases , resulting parallax video results for frame-by-frame match may exist the parallax fluctuations and matching errors . Herein by the affine transformation of the video between successive frames to establish contact of the time-domain , binding of overlap between the front and rear frame divided block parallax variation smoothness assumptions, the use of the correlation between the left and right view , parallax results unreliable region selectively optimized most matching error correction , easing parallax fluctuations between successive frames image , to obtain a more accurate , smooth , stable the parallax video results effectively enhance the video quality of the parallax .

Related Dissertations

  1. Research of Visualization Technology in the Virtual Test of Missile,TP391.9
  2. Research on the Classification Based on the Reconstruction of Solder Joint,TP391.41
  3. 3D Visualization Research of Medicial Ultrasound Image,TP391.41
  4. The Research on Localization Algorithm of Wireless Sensor Networks,TN929.5
  5. Study on the Road Condition Monitoring Based on Vehicular 3D Acceleration Sensor,TP274
  6. The Application Study of Delay-enhanced MRI in Normal Cartilage and Phase Ⅰ and Ⅱ of Cartilage Lesions of the Knee,R445.2
  7. Face floor hydrous mine water inrush prediction techniques to detect and study,TD745
  8. Research on CNC Milling System,TG547
  9. 3D Battlefield Posture Analysis and Demonstration System,P208
  10. Research on 3D Visualization Algorithm of Medical Image,TP391.41
  11. Synthesis of SnO2 by Flame Spray Pyrolysis: Structures Design and the Gas Sensing Properties,TB383.1
  12. Computer-assisted hip replacement surgery systems research and development,TP391.41
  13. Design and Research of the Inter-Chip Inductive Coupling Wireless Transceiver for 3D Stacking Package,TN859
  14. Study of Binocular Stereo Matching Algorithm and Algorithm Implementation Based on Multi-core.,TP391.41
  15. The Research on Electrode 3D Model Classification and Retrieval Based on SVM and Shape Features,TP391.41
  16. Research of Reconstruction and Camera Pose Estimation on Mobile Phone,TP391.9
  17. Design and Implementation of a Three-dimensional Widget Set,TP391.41
  18. Underwater 3D Reconstruction Based on Binocular Stereo Vision,TP391.41
  19. The Application of VANTAGE PDMS (3D Integrated Design System) in Electrical Engineering in Plant Design,TM621
  20. Three-dimensional Temperature Field Modeling and Visualization for Crystal Furnace,TP391.41

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