Meet the Factor: Midge - A Small Defect with a Big Impact
Grain grading isn’t just about what you can see at a glance. It’s about understanding what each kernel tells you. One of the more subtle but important quality factors in cereals like wheat is midge damage. While easy to overlook, midge can influence grading, processing quality, and the marketability of a load.
At Ground Truth Ag, our AI models recognize midge damage with precision, using thousands of expert annotated samples. Our benchtop system combines machine vision and near-infrared spectroscopy (MV/NIRS) to provide fast, consistent detection across every sample.
What Is Midge Damage?
Midge damage occurs when the orange wheat blossom midge lays eggs on the wheat head, and larvae feed on the developing kernels. The result? Kernels that are shrivelled, grooved, or have signs of chewing and deformation.
Even when present in small amounts, midge can lead to grading deductions, especially in high-quality classes like Canadian Western Red Spring (CWRS) wheat, where appearance and integrity matter.
Photo source: saskatchewan.ca
Where We Detect Midge Today
We currently detect and assess midge damage in:
CWRS Wheat – Midge impacts test weight, appearance, and suitability for milling.
Other wheat classes – Detection support is expanding as our models grow.
By training our system on a wide variety of real-world samples, we are improving the model’s ability to pick up early-stage or light midge that might otherwise go unnoticed.
Why Midge Matters
For Farmers & Sellers – Midge damage can mean the difference between a top-grade premium and a downgrade. Early, accurate detection helps sellers plan blending strategies or separate bins before delivery.
For Buyers & Processors – Midge-affected kernels can compromise flour output and appearance. Knowing the level of midge present supports better sorting and processing decisions.
For Grading Consistency – Visual midge detection can vary significantly between graders. Automating this step improves objectivity and repeatability; no matter who is operating the system.
How We Detect Midge
Our benchtop system captures high-resolution images of the entire sample, unlike traditional grading, which might only inspect a handful of kernels. From there, our AI evaluates shape, texture, and damage markers to flag suspected midge.
This method:
Reduces the subjectivity of human inspection
Identifies even subtle signs of midge
Provides standardized results across locations and operators
When combined with other grading factors, midge detection feeds into a full picture of grain quality; helping both sellers and buyers make confident, data-backed decisions.
Want to see midge detection in action? Book a demo to learn how our AI-powered benchtop works.