McFarlane, Nigel James Bruce; (2002) Foreign body detection in food materials using compton scattered x-rays. Doctoral thesis (Ph.D), UCL (University College London).

Text Foreign_body_detection_in_food.pdf Download (21MB)

Abstract

This thesis investigated the application of X-ray Compton scattering to the problem of foreign body detection in food. The methods used were analytical modelling, simulation and experiment. A criterion was defined for detectability, and a model was developed for predicting the minimum time required for detection. The model was used to predict the smallest detectable cubes of air, glass, plastic and steel. Simulations and experiments were performed on voids and glass in polystyrene phantoms, water, coffee and muesli. Backscatter was used to detect bones in chicken meat. The effects of geometry and multiple scatter on contrast, signal-to-noise, and detection time were simulated. Compton scatter was compared with transmission, and the effect of inhomogeneity was modelled. Spectral shape was investigated as a means of foreign body detection. A signal-to-noise ratio of 7.4 was required for foreign body detection in food. A 0.46 cm cube of glass or a 1.19 cm cube of polystyrene were detectable in a 10 cm cube of water in one second. The minimum time to scan a whole sample varied as the 7th power of the foreign body size, and the 5th power of the sample size. Compton scatter inspection produced higher contrasts than transmission, but required longer measurement times because of the low number of photon counts. Compton scatter inspection of whole samples was very slow compared to production line speeds in the food industry. There was potential for Compton scatter in applications which did not require whole-sample scanning, such as surface inspection. There was also potential in the inspection of inhomogeneous samples. The multiple scatter fraction varied from 25% to 55% for 2 to 10 cm cubes of water, but did not have a large effect on the detection time. The spectral shape gave good contrasts and signal-to-noise ratios in the detection of chicken bones.

Download activity - last month

Export as CSVXMLJSON

Download activity - last 12 months

Export as CSVXMLJSON

Downloads by country - last 12 months

Export as CSVJSONXML

Archive Staff Only

View Item