Date of Award
8-2010
Document Type
Thesis
Degree Name
Master of Science (MS)
Legacy Department
Electrical Engineering
Committee Chair/Advisor
Birchfield, Stanley
Committee Member
Gowdy , John
Committee Member
Woodard , Damon
Abstract
This work proposes a piecewise linear network model to approximate structures observed in an image. An energy function is used to capture the characteristics of the structure. The energy function consists of two parts: the prior energy term and the data energy term. The prior energy term is calculated using prior information about the structures of interest. The data energy term is calculated using observations made from the image. The energy function is minimized using Reversible Jump Markov Chain Monte Carlo (RJMCMC) to get the approximate centerline of the structure. The algorithm was tested on a database of $150$ images containing underground roots taken by a minirhizotron camera. The results show the importance of a novel non-Gaussian term introduced to handle roots with low intensity near the centerline. It is possible to use the proposed model to detect other structures such as roads as shown by the preliminary results.
Recommended Citation
Apte, Akshay, "Detecting Piecewise Linear Networks Using Reversible Jump Markov Chain Monte Carlo" (2010). All Theses. 927.
https://open.clemson.edu/all_theses/927