Location

Seattle, WA

Start Date

1-1-1996 12:00 AM

Description

This work has been prepared within the Brite-Euram project “DUALETO”, whose purpose is the development of a high-resolution Dual-Energy computerized tomography system (DE-CT). It is dedicated to non-destructive inspection of fiber-reinforced ceramic parts. The DE-measurement will be acquired simultaneously by an energy- and position-sensitive line detector. Using an appropriate calibration function, these DE-acquisitions are used to calculate the distributions of mass density of the ceramic’s two components and with them, the volumic fraction (VF) of fibers in the matrix material. Due to several reasons, these density data are very noisy. This paper presents a method for obtaining mass density data with a lower level of noise. For this, structural information about object, defects and bundles of fibers are derived from the low-energy measurement and stored in label images. Structure information is then used to perform a non-linear filtering on mass density data, in order to estimate the volumic fraction.

Volume

15A

Chapter

Chapter 1: Standard Techniques

Section

Radiography and Computed Tomography

Pages

473-480

DOI

10.1007/978-1-4613-0383-1_60

Language

en

File Format

application/pdf

Share

COinS
 
Jan 1st, 12:00 AM

Segmentation and Density-Evaluation of Fiber-Reinforced Materials by Dual-Energy Computerized Tomography

Seattle, WA

This work has been prepared within the Brite-Euram project “DUALETO”, whose purpose is the development of a high-resolution Dual-Energy computerized tomography system (DE-CT). It is dedicated to non-destructive inspection of fiber-reinforced ceramic parts. The DE-measurement will be acquired simultaneously by an energy- and position-sensitive line detector. Using an appropriate calibration function, these DE-acquisitions are used to calculate the distributions of mass density of the ceramic’s two components and with them, the volumic fraction (VF) of fibers in the matrix material. Due to several reasons, these density data are very noisy. This paper presents a method for obtaining mass density data with a lower level of noise. For this, structural information about object, defects and bundles of fibers are derived from the low-energy measurement and stored in label images. Structure information is then used to perform a non-linear filtering on mass density data, in order to estimate the volumic fraction.