Page 12
Accuracy of seed counting using the average area
method
The average area method is similar to that which uses weighing to count seeds, in
which the total weight of all seeds is divided by the average weight of a few seeds..
A key difference is that it does not matter if the weighed seeds overlap, whereas a
systematic under-reporting error will be added to the average area method.
Given the above, WinDIAS uses a statistic developed by Steiner and Mackay (2001)
for estimating the variance of seed counting accuracy based on weighing.
The standard deviation
σ
r
of the count is derived
1
as follows:
𝜎
𝑟
2
= n
𝛾
2
{1 +
𝑛
𝑝
}
Where
𝛾
=
𝜎
𝑎
σ = standard deviation of area of the individually measured seeds
a = average area of the individually measured seeds
n = estimated total number of seeds = Total Area/a.
p = number of individual seeds identified and measured
Conveyor Operations
WinDIAS can automatically grab and measure the area of leaves fed individually
through the Conveyor, saving both individual readings and running totals.
Leaves too long to be grabbed all at once can be measured using the Long Leaf
mode.
Conveyor operations only work with the video camera supplied.
1
Steiner and MacKay “Scale Counting” IIQP Research Report RR-01-07. University of
Waterloo,Canada 2001
