Rapid machine vision method for the detection of insects and other particulate bio-contaminants of bulk grain in transit

A machine vision method has been developed for the detection of adult beetles, rodent droppings and ergot in. harvested wheat. Samples were presented as a monolayer of around 25 generally non-touching but randomly oriented objects. Detection rates for artificially prepared samples ranged from 87% for ergot to 100% for adult Oryzaephilus surinamensis (saw-toothed grain beetle). When applied to two commercial sa less containing several insect species, detection rates were 89 and 96%. The rate of false positives arising from samples of wheat containing permitted admixture-only was low. The speed of the integrated algorithm package for, detection of all three contaminant types approaches that required to scan 3 kg grain in just 3 min (the throughput required by the UK trade). (C) 2002 Silsoe Research Institute. Published by Elsevier Science Ltd. All rights reserved.