Dissertation > Excellent graduate degree dissertation topics show
Research on a Sparse Stereo Matching Algorithm Based on Belief Propagation
Author: ChenJiaKun
Tutor: YinZhiBen
School: Southwest Jiaotong University
Course: Applied Computer Technology
Keywords: Stereo vision Stereo matching Belief Propagation Delaunay triangulation Image restoration
CLC: TP391.41
Type: Master's thesis
Year: 2011
Downloads: 76
Quote: 0
Read: Download Dissertation
Abstract
|
Stereo vision is an important branch of computer vision technology has been widely applied to three-dimensional measurement , robot vision , virtual reality and other fields . The stereo matching is calculated based on the selected feature , so as to establish the corresponding relationship between the characteristics , is an extremely critical and difficult step in the stereoscopic vision , and its goal is to establish the disparity map . Stereo matching has a broad application prospects , is a hot research field of computer vision . Global constraint - based stereo matching problem is generally modeled as an energy minimization problem , using a variety of optimization methods to find the minimum energy function is defined . These dynamic programming optimization method , graph cuts , belief propagation (Belief Propagation, BP) . And BP algorithms for graph cut is the better of the two types of global stereo matching algorithm . However , these algorithms computing a lot more time than the algorithm based on local constraints . Proposed stereo matching algorithm based on BP algorithm , and is committed to reducing the computation time of the algorithm . Delaunay triangulation (Delaunay Triangulation, DT) grid is a form of image description . DT algorithm to the image feature points as vertices of a triangle so as to generate an image-based content of the DT grid . Using the value of the grid nodes pixel interpolation can effectively restore the image . This article first systematic introduction to the basic principles and steps of the stereo vision and stereo matching , and then analyze the DT grid nature and algorithms . This article take full advantage of the superiority of the the DT grid in the field of image coding , sparse stereo matching algorithm based on BP's . Algorithm as follows: extracting the characteristics of the left image , obtained sparse the left image , and to use these features to obtain a content-adaptive DT grid ; using stereo matching algorithm based on the BP , the left and right images of the match sparse sparse disparity map matching accuracy while reducing the amount of time of the algorithm ; using the the DT grid - based image recovery technology , the sparse disparity map triangle interpolation dense disparity map . The experimental results show that the method while reducing the matching time and obtain a higher accuracy disparity map .
|
Related Dissertations
- Research on Restoration of Images Collected by Reconnaissance System in Near-Space,TP391.41
- Research on Learning-Based Low-Level Vision Problem,TP391.41
- Image Restoration Method Research and Its Application Based on Probability PCA,TP391.41
- Research on the Vision Control and Trajectory Planning of Fruit Picking Robot,TP242.6
- Real-time Three-dimensional Profile Measurement and Software Developing,TP274
- Study of Binocular Stereo Matching Algorithm and Algorithm Implementation Based on Multi-core.,TP391.41
- Spatial Uncertainty in Stereo Visual Odometry,TP391.41
- Spatial Temporal Refinement on Depth Video Processing,TP391.41
- Underwater 3D Reconstruction Based on Binocular Stereo Vision,TP391.41
- The Research of Image Restoration Based on Phase Diversity Method,TP391.41
- OSG-based three-dimensional irregular terrain matching military standard applied research,E919
- Multi- dimensional image based aviation city buildings Recognition Technology,TP391.41
- Based on the geometry of multiple images and textures automatically rebuild,TP391.41
- Based on public safety characteristics of millimeter-wave radiation image,TP391.41
- Exercise and defocus blurred image restoration,TP391.41
- Based on single- camera stereo vision system full visual odometry research,TP391.41
- Stereoscopic three-dimensional reconstruction -related technology research and implementation,TP391.41
- Image restoration of key technologies,TP391.41
- Research on Binocular Stereo Vision System for Indoor Mobile Robot on Complex Environment,TP242
- Research on Recognizing and Locating Tomatoes Base on Binocular Stereo Vision Technology,TP391.41
- Research on Repeat Accumulate Decoding Algorithms,TN911.22
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
|