Dissertation > Excellent graduate degree dissertation topics show

Applications and Realization of Motion Estimation in Image Stabilization and Matching and Tracking

Author: LuoJun
Tutor: ShenZhenKang
School: National University of Defense Science and Technology
Course: Information and Communication Engineering
Keywords: Electronic Image Stabilization Motion Estimation Motion compensation Image matching The pseudo polar coordinates Fourier transform The Fourier - Merrill transform Digital Signal Processor Field programmable logic devices
CLC: TP391.41
Type: PhD thesis
Year: 2007
Downloads: 764
Quote: 6
Read: Download Dissertation

Abstract

Developed according to the actual topic needs paper for motion estimation techniques in electronic image stabilization and image matching tracking depth research and design to achieve a high-performance processor seeker. The thesis research contents are as follows: 1. Introduced the motion estimation technology content, summed up the common, the main method for electronic image stabilization, motion estimation and motion compensation methods. Mechanical image stabilization, the advantages and disadvantages of optical image stabilization and electronic image stabilization, summed up the domestic and international electronic image stabilization technology research and development status on the electronic image stabilization technology basic principle and general system structure. Study the basis of two-dimensional electronic image stabilization problem, an algorithm based on local demand fine bit plane matching motion estimation and Kalman filtering movement compensation diamond search electronic image stabilization, and in the study of one-way and two-way gray projection algorithm on a bi-gray projection motion estimation and Kalman filtering movement compensation electronic image stabilization algorithm. In addition, the problems caused by the lack of image content due to the motion compensation, a full-frame video stabilization algorithm based on image stitching technology. This algorithm to take the center of the statistical significance of the estimate as a sequence of video images true stable center, as a reference of motion compensation, guarantee the stability of the full frame video image sequence can be performed to a higher degree of confidence \Image matching tracking problem in the case of translation, rotation and scaling were studied based on the frequency and spatial image registration techniques. Pseudo-Polar Fourier transform, and discuss their application in image registration of rapid polar coordinates based on Pseudo-Polar Fourier Transform Fourier transform, and take advantage of this rapid polar coordinate Fourier transform in the frequency domain registration Fourier-Mellin transform-based image registration algorithm has been improved, and an improved image registration algorithm based on Fourier-Mellin transform; registration technology research in the airspace, positive, positive synthesis of major research anti-additive anti-synthesized several nonlinear optimization iterative algorithm. Then in the study on the basis of the frequency and spatial image registration techniques iteration algorithm for optimization problems, combined with the rapid polar coordinate Fourier transform Fourier-Mellin image registration algorithm and Levenberg-Marquardt algorithm against synthetic proposed A joint frequency domain and airspace image registration algorithm to accurately estimate the translation, rotation, scaling, movement between the image transformation parameters. Summarizes several image processing hardware system architecture, analyze their strengths and weaknesses, and then using field programmable logic devices (FPGA) and general-purpose high-speed digital signal processor (DSP)-based hardware systems constitute program, designed to achieve a kinds of high-performance processor seeker. FPGA design to achieve the the dedicated image pre-processor, the image pre-processor to complete the electronic image stabilization, global image preprocessing in image acquisition to improve the processing speed of the information processor system. Using high-performance DSP (TMS320C6203) as the core, the design of a high-speed processor and system controller, high-speed processor and system controller to the completion of a large amount of data efficiently, high-speed image-guided information processing.

Related Dissertations

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