Dissertation > Excellent graduate degree dissertation topics show

Image-based 3-D Reconstruction of Tree-like Objects

Author: LiuYanHong
Tutor: LiHua
School: Institute of Computing Technology
Course: Applied Computer Technology
Keywords: Image-based modeling Tree Skeleton Segmentation Track
CLC: TP391.9
Type: Master's thesis
Year: 2002
Downloads: 238
Quote: 0
Read: Download Dissertation

Abstract

The development of virtual reality technology to obtain a three-dimensional model of the entities in the virtual environment. Image-based modeling techniques is an object of concern in recent years, the rise of 3D modeling methods, trying to starting a three-dimensional model of data for specific objects from the two-dimensional image data. Tree-like object as a universal natural entities, the establishment of the three-dimensional model has important significance in natural scenery and medical biology. This paper discusses the problem of reconstruction based on the image of the tree-like object, double viewpoint image reconstruction model of tree branches. Involving image segmentation and recognition techniques, player segmentation and tracking algorithm in soccer video sequence. The main content of this paper include: 1. Branches model Stella class-D object from a dual viewpoint image reconstruction program. Using the point reconstruction method based on stereo vision, the reconstruction of trees through the trees trunk skeleton points corresponding two-dimensional to three-dimensional skeleton, and then added in the three-dimensional skeleton width information to generate a three-dimensional model of the tree branches. Proposed a method to automatically obtain the two-dimensional trunk skeleton trees data, two-dimensional reconstruction of the skeleton point correspondence between solving, as well as three-dimensional skeleton points, simplify and stems surface mesh generation. The experiments show that the proposed scheme to ensure the quality of the reconstruction in the case of a very small amount of interaction is indeed feasible and effective tree reconstruction methods. 2, one pair of a player in a football game video sequence, the real-time segmentation and tracking algorithm, to achieve a segmentation of the players, the team is to identify and track. (1) According to the characteristics of the color of the football field, the use of the statistics of the difference of the color components to achieve automatic segmentation of the players in the video sequence; (2) make full use of the color information of the image, the players and the two teams template of each color component. statistical histogram normalized correlation comparison, identification of the players belong to the team; (3) the use of the contextual information of the players, combined with player tracking based on template matching method. The success of the algorithm to achieve a real-time tracking of the players in the football scene, a good deal of the different team members blocked and can handle more than occlusion. Tree-like object-based image reconstruction techniques in the development stage, there are many problems to be solved. Which, due to the complex structure of the tree branches, also need further study to better accomplish the branches image segmentation and recognition. Nevertheless, the tree-like object-based image reconstruction technique still bears a huge potential for development, the extensive attention by people.

Related Dissertations

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Computer simulation
© 2012 www.DissertationTopic.Net Mobile