Title :
Statistical approaches to image restoration
Author :
Burrows, J.D. ; Little, J.A. ; Mardia, M.V. ; Taylor, C.C. ; Walder, A.N.
Author_Institution :
DRA, UK
Abstract :
Our work has been based on X-ray images of suitcases typical of those which can be seen at an airport. Specifically we have been trying to locate components, such as a detonator, wires and (less successfully) plastic explosive, which would constitute a bomb. Standard thresholding techniques to extract lead solder, for example, do not use any spatial information. By using prior knowledge, a number of contextual techniques have been explored. Our first step is to use a deblurring algorithm on the image; it is convolved with a blurred point spread function and then deconvolved with a prior encouraging neighbouring pixels to be alike. The results obtained always seem to improve the visual quality of images
Keywords :
X-ray imaging; image enhancement; image restoration; optical transfer function; security; statistical analysis; airport; blurred point spread function; deblurring algorithm; image quality; image restoration; neighbouring pixels; spatial information; statistical approaches; thresholding techniques; visual quality; x-ray images;
Conference_Titel :
Security and Detection, 1995., European Convention on
Conference_Location :
Brighton
Print_ISBN :
0-85296-640-7
DOI :
10.1049/cp:19950470