Seed Treatment Assessment
Seed treatment’s application is to optimize seed performance, cost-effectiveness and yield of the crop.
Seed treatment examples are:
1
Seed Dressing
2
Coating
3
Pelleting
4
Priming
5
Upgrading
6
Seed Disinfection
7
Seed Pathogen Detection
Seed Treatment with the VideometerLab
Seed imaging by the VideometerLab is used to increase and validate the efficacy and homogeneity.
On-line or at-line solutions for optimizations of seed imaging are also possible.
To the right we show a petri dish with coated rice seeds and below we will discuss in detail how to analyze the seed coating process.
Seed Coating
One of the examples of seed treatment is seed coating process. It applies a thin water permeable polymer based coating fim onto the seeds. In addition, the process is optimized to obtain a specified amount of polymer per unit area (the loading) with high homogeneity, and low dust-off and flake-off.
Moreover, cosmetic requirements like color, opacity, and surface finish must also be met. Coverage is the percentage of seed area with a loading above a given threshold, and coverage will decrease with higher heterogeneity, and higher dust-off/flake-off.
VideometerLab provides a fast and non-destructive way to simultaneously measure optical loading, coverage, heterogeneity, and color on a seed sample.
As an illustrative example on sample measurement and coating process optimization we create a recipe for the coating on rice and map the loading characteristics for the coating array shown in the figure below.
The coating array consists of 25 treatments of rice with the same polymer. Only the amount of coating liquid and the concentration of polymer in the coating liquid are various. The total amount of polymer used is shown in the table on the left. In the coating array we try to span the relevant range of loading and loading homogeneity.
How to Design a Seed Coating Recipe for a Given Polymer and a Given Crop
Before you can create a recipe you need to prepare and image at least two samples for seed treatment:
• A control sample of untreated seeds
• A reference sample of “ideally” coated seeds
These samples are shown.
Then, from the two images a machine learning algorithm will learn a loading mapping (image on the left).
This loading map shows the loading in every pixel of the image. Average loading and heterogeneity (Coefficient of variation, CV) can now be calculated for each sample.
Results of the seed coating recipe are shown below. The loading map, average loading, and loading heterogeneity obtained directly from the recipe is shown on the top left. In addition, bottom left figure shows the relation between optical loading and amount of polymer. Hence, this can be used as a calibration curve between the two.
Bottom right figure shows the loading heterogeneity in the coating array. We see that increasing amount of coating liquid decreases heterogeneity. Moreover, is it also visible that the optimal concentration for low heterogeneity varies with the amount of coating liquid. 50% concentration is the worst at 10 ml of coating liquid but the best at 50 ml.
Seed Pelleting
Seed pelleting adds inert material to seeds in order to increase plantability. Important measurements on seed pellets performed with VideometerLab are:
• Size distribution
• Color variation
• White dots
• Shape distribution