基于纹理和轮廓信息的图像序列三维重建算法研究
VOLUMETRIC STEREO AND SILHOUETTEFUSION FOR IMAGE-BASED MODELING
摘 要
ABSTRACT
ACKNOWLEDGEMENTS
CONTENTS
LIST OF TABLES
LIST OF FIGURES
CHAPTER 1INTRODUCTION
1.1 Background
1.2 Related works
1.3 Algorithm overview
1.4 Thesis overview
CHAPTER 2SHAPE FROM SILHOUETTE
2.1 Introduction to visual hull
2.2 Volumetric visual hull computation approach
2. 2. 1 Octree and marching cubes algorithm
2. 2. 2 Some existing projection test methods
2.3 The proposed visual hull computation approach
2. 3. 1 3D bounding box estimation
2. 3. 2 Visual hull octree construction
2. 3. 3 Visual hull mesh extraction
2.4 Summary
CHAPTER 3SHAPE FROM MULTI-VIEW STEREO
3.1 Depth map estimation from multi-view stereo
3. 1. 1 Geometric relation between multiple images
3. 1. 2 Greedy depth map estimation
3. 1. 3 Expansion-based depth map estimation
3.2 Point cloud cleaning and downsampling
3.3 Surface normal estimation
3.4 Summary
CHAPTER 4VOLUMETRIC STEREO AND SILHOUETTE FUSION
4.1 Introduction
4.2 Oriented point cloud from silhouette
4.3 Surface reconstruction with PSR method
4.4 Summary
CHAPTER 5EXPERIMETNAL RESULTS
5.1 Camera calibration
5. 1. 1 Camera model
5. 1. 2 Circular motion parameterization
5. 1. 3 Camera calibration
5.2 Three Terra Cotta Warrior sequences
5. 2. 1 General sequence
5. 2. 2 Captain sequence
5.3 Three other sequences
5. 3. 1 Skull sequence
5. 3. 2 BigHead sequence
5. 3. 3 Twins sequence
5.4 Two benchmark sequences
5. 4. 1 Dino sequence
5. 4. 2 Temple sequence
5.5 Summary
CHAPTER 6CONCLUSION AND FUTURE WORK
6.1 Conclusion
6.2 Future work
REFERENCES
PUBLICATION
哈尔滨工业大学硕士学位论文原创性声明
哈尔滨工业大学硕士学位论文使用授权书