Use of X-ray to identify contaminants in pelleted seed lots for biosecurity : a thesis presented in partial fulfilment of the requirements for the degree of Master of Science (Agricultural Science) at Massey University, Manawatū, New Zealand
Thousands of tonnes of seed, of which around 10% is pelleted, comes into New Zealand
through international trade every year. However, this trade also brings potential risks to
New Zealand biosecurity. Pelleted seeds can contain contaminants, including seeds
other than the crop species in the seed lot and inert matter; both may cause negative
effects on crop growth or bring pests and diseases. A reliable method is necessary to
inspect seed lots for the contaminants.
The conventional way to inspect for contaminants in pelleted seeds is to separate the
seeds from pellets and inspect visually. However, this is a time consuming and
potentially health damaging procedure. A faster and safer non-invasive inspection
method is needed urgently. X-ray imaging systems have the potential to non-invasively
identify contaminants in seed lots.
2-D X-ray was firstly applied in this research to determine if the system could separate
non-target seeds such as weed seed from naked crop “target” seeds, since if 2-D X-ray
cannot separate non-target seeds from naked target seeds, there is little chance to
separate seeds that are pelleted. In this research, three target species were used. These
were beet (Rapistrum, Ranunculus and spinach as contaminants), carrot (Polygonum,
Chenopodium and Solanum as contaminants) and lettuce (Sonchus and Lapsana as
contaminants), because of their high contamination rates in imported seed lots. Seed
shape parameters: dimensions, form, circularity, roughness and intensity, were used to
characterize seeds for further comparison. The results showed Ranunulus can be separated from beet by dimensions and intensity; Rapistrum can be separated by
elongation, circularity and intensity; spinach was hard to separate from beet. In the
carrot group, Chenopodium and Solanum can be separated from carrot by either
dimensions, elongation or circularity, while Polygonum cannot be separated from
Carrot. For contaminants in lettuce, Sonchus can be separated from lettuce by
dimensions and intensity; Lapsana can be separated by elongation and circularity.
However, all the separation above was based on mean values, seeds with extreme sizes
would limit the effects of shape parameters in seed separation.
Determining if pelleting seeds can also be separated using the same parameters was the
next important step for determining if 2-D X-ray can be used for pelleted seed
inspection. However, little literature can be found regarding specific pelleting materials
and pelleting procedures, as they are held by the seed companies. Therefore, protocols
for pelleting the relatively small numbers of pelleted seed for research are needed.
During several trials on seed pelleting, Methocel™ and gypsum was identified as
suitable pelleting materials. The vortex mixer was identified as the best equipment for
pelleting using a one-by-one adding method, which was feasible for pelleting both tiny-
seeds and small-quantities seeds. The seeds pelleted showed a uniform and well-
rounded appearance. However, when applying the same 2-D X-ray for seed separation,
the seed projections were hard to be extracted for further analysis, because of the poor
differentiation between seeds and pellets.
This research explored the potential of using 2-D X-ray to separate naked non-target
seed from naked target seeds by seed shape parameters. The outcomes confirmed that the mean values of shape parameters can separate contaminants from target seeds,
however at the extreme ends of the range seed parameters overlap will limit the value
of the shape parameters. Pelleting seeds under laboratory conditions can also be realized
by using vortex mixer as equipment and using Methocel™ and gypsum as pelleting
materials. Nonetheless, 2-D X-ray was not a reliable tool to detect pelleted seeds, since
it is hard to separate seed projections from pellets with images only from a top view.
3-D X-ray could potentially be applied in future research because of its higher
resolution than 2-D X-ray. In addition, 3-D X-ray images enable analysts to analyze
seeds from different angles other than one fixed angle, which makes the analysis free
from image overlap problems. Although research on 3-D X-ray for seed separation is
at its beginning, it is potentially useful for pelleted seed analysis.
The following Figures were removed for copyright reasons, but may be accessed via their respective source: Figs 16 & 17 (=Landis & Keane, 2010, Figs 1 & 3) and Figs 18-24 (=Blott & Pye, 2008, Figs 1, 8, 9, 14, 15, 16, 17).