Forward-Backward Time Stepping with Automated Edge-preserving Regularization Technique for Wood Defects Detection

Yong, Guang (2019) Forward-Backward Time Stepping with Automated Edge-preserving Regularization Technique for Wood Defects Detection. PhD thesis, Universiti Malaysia Sarawak (UNIMAS).

[img] PDF (Please get the password by email to repository@unimas.my , or call ext: 3914 / 3942 / 3933)
Forward-Backward Time Stepping with Automated Edge-preserving.pdf
Restricted to Registered users only

Download (4MB) | Request a copy

Abstract

Wood is used as a major structural material in a great variety of building and civil engineering application. The quality of the wood in the wood industry is graded in terms of strength, appearance, and durability. Difference grades of wood will serve to their own purposes and cost-effective way. Hole and knots are the main defects in wood which will weaken the wood structural strength and affect the grades of wood. Many non-destructive evaluation (NDE) systems have been developed to detect defects and voids in the wood. However, these NDE systems have their own limitations. Active infrared thermography is expensive and only can detect the thermal resistances defects. Ultrasonic testing has difficulty in the inspection of rough, irregular shape materials. X-rays computed tomography is expensive and radioactive to the environment. Besides these NDE systems, microwave non-destructive evaluation is a potential approach that is non-contact, non-radioactive, cheaper cost and able to investigate materials electromagnetic. Therefore, the microwave imaging technique is proposed to detect the wood defects. In this thesis, Forward-Backward Time-Stepping (FBTS) method is applied to microwave imaging wood defect detection application. This method is an active microwave imaging technique which is formulated in the time-domain utilizing Finite-Difference Time-Domain (FDTD) method. Since the microwave imaging is non-linear and ill-posed nature, FBTS method is integrated with the edge-persevering regularization technique to avoid the solution trapped in the local minima. However, it is not an easy job to find the optimum parameters for edge-preserving regularization technique. Therefore, two automated procedures are developed to determine these parameters iteratively. The FBTS integrated with automated edge-preserving regularization algorithm is implemented in C++ programming language executed in parallel computing. The FBTS integrated with automated edge-preserving regularization technique is first assessed with object detection application. The reconstructed images show that the algorithm was able to determine the dielectric properties within the region of interest (ROI) and estimate the shape, location of the embedded object. After that, the research is extended to investigate on realistic numerical wood model. The reconstructed images show that the FBTS integrated with automated edge-preserving regularization technique able to estimate the shape, size and the location of void in 50 cm diameter wood model and this technique able to detect as small as 25 mm diameter void in the wood model.

Item Type: Thesis (PhD)
Additional Information: Thesis (PhD.) - Universiti Malaysia Sarawak , 2019.
Uncontrolled Keywords: Microwave imaging, time-domain inverse scattering technique, ForwardBackward Time-Stepping (FBTS), edge-preserving regularization technique, object detection, wood defect detection, unimas, university, universiti, Borneo, Malaysia, Sarawak, Kuching, Samarahan, ipta, education, Postgraduate, research, Universiti Malaysia Sarawak.
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
T Technology > TK Electrical engineering. Electronics Nuclear engineering
Divisions: Academic Faculties, Institutes and Centres > Faculty of Engineering
Faculties, Institutes, Centres > Faculty of Engineering
Depositing User: YONG GUANG
Date Deposited: 11 Sep 2019 07:26
Last Modified: 25 May 2023 08:19
URI: http://ir.unimas.my/id/eprint/26811

Actions (For repository members only: login required)

View Item View Item