This paper describes a novel methodology to reduce the scanning time and to extract bar dimensional information from images generated by an inductive sensor. Using a sparsely populated data set obtained from a reduced number of scan lines, faster high resolution images are generated using image interpolation techniques. Having generated these images, image filtering, peak picking and neural networks methods are applied to extract bar dimensional information and accurate 3 D visualization.
|Title of host publication||Review of Progress in Quantitative Nondestructive Evaluation Volume 24|
|Number of pages||8|
|Publication status||Published - 9 Apr 2005|
|Event||Review of Progress in Quantitative Nondestructive Evaluation - Golden, CO, United States|
Duration: 25 Jul 2004 → 30 Jul 2004
|Conference||Review of Progress in Quantitative Nondestructive Evaluation|
|Period||25/07/04 → 30/07/04|