Monday, June 24, 2024

10:40 AM - 11:50 AM

408-409

---

Overview

Dr. Toby Case | Aerospace Corporation, USA

---

Details

Most of the world performs measurements uniformly; however, some important applications for NDE require nonuniform measurement. Examples of uniform measurement are raster scans performed with uniform stepping and scanning, measurements across a bandwidth sampled at uniformly spaced frequencies, and, for computed tomography, x-ray images taken in a full circle at uniformly spaced angles. Signal processing, image processing, and reconstruction techniques are well-known and highly optimized for uniform measurements. For those well-versed in uniform measurements, there is the constant struggle of sampling sufficiently but not too much since one must reduce or eliminate aliasing artifacts yet not require so many samples as to exhaust memory resources or require prohibitively long data acquisition times. The alternative is nonuniform measurement. It is a technique that eliminates aliasing, reduces the required number of measurements, and may increase data acquisition throughput. Nonuniform measurement has at its heart nonuniform sampling theory, which is used extensively for compressive sensing. A brief, high-level introduction will be given for nonuniform signal processing, image processing, and reconstruction techniques where specific steps are taken to reconstruct uniform samples from nonuniform samples so that highly optimized techniques developed for uniform measurements may be applied. The cost of nonuniform measurement is increased computational complexity. However, this tutorial describes practical examples of NDE applications that are only made possible with nonuniform measurement. These examples specifically demonstrate how nonuniform measurements facilitate difficult inspections and can be used to close high impact inspection gaps.

---

Speaker