Canadian orchard AI — codling moth smart traps (BC) and tree-fruitlet computer vision (Ontario)

NA-Canada (British Columbia, Ontario, Alberta)

Content

Canadian orchard AI in mid-2026 is anchored by two CAAIN-funded projects plus the existing Vivid Machines deployments surfaced in Scale AI and Digital units. Spans BC, Ontario, and Alberta.

CropVue Technologies (Surrey, BC) — codling moth smart trap AI

CropVue Technologies Inc. and the Okanagan-Kootenay Sterile Insect Release Board (OKSIR) have partnered to develop an AI-driven pest monitoring solution for codling moth (Cydia pomonella) in apple and pear orchards. Codling moth is a destructive insect pest that causes significant fruit damage and economic loss in Canada and worldwide.

Currently the OKSIR program suppresses this pest by releasing sterile moths, but field staff must manually inspect sticky traps to distinguish sterile (non-breeding) moths from wild (fertile) moths — a tedious, costly process.

The CropVue smart trap system integrates rugged IoT-enabled camera traps with a machine learning model trained to recognize subtle visual differences or applied markers on the moths’ bodies. The camera trap captures an image of captured moths, and the AI model classifies each moth as sterile or wild pest.

Outcomes.

• Eliminate most manual trap checks.
• Reduce labour costs.
• Provide more frequent and accurate pest population data to growers and pest control managers.
• Field personnel alerted to problem areas faster, allowing for timely interventions (targeted additional sterile releases or other controls) and preventing crop damage.
• Anticipated outcomes: validated AI model for insect classification, set of functional camera trap prototypes, proof-of-concept demonstration in commercial orchard environment.

Amy Jancewicz (President, CropVue Technologies).

Vivid Machines (Toronto, ON) — tree-fruitlet computer vision

Vivid Machines (Toronto-headquartered; “physical AI for fruit farming”). Vision-based orchard scanning that records fruitlet placement on every tree. Real-time crop load estimation. Featured at VivaTech 2026. Active project with Quinton Gibson Ontario apple farm. Partner: Tall Grass Ventures (Calgary AB).

The Grower Feb 2026 confirms “visioning system capable of scanning each tree and accurately recording fruitlet placement on every tree in an orchard.”

Cross-cutting — Vivid Machines in Scale AI + Digital

Vivid Machines also appears in Scale AI and Digital Technology Supercluster projects:

• Scale AI: Transforming Fruit Production; Apples to All: Growing an AI-Enabled Agricultural Data Platform.
• Digital: Same projects.

This is the cross-cluster funding pattern documented in Scan 1 (5+ leads in both Scale AI and Digital). Vivid Machines is a Canadian federal cluster funding stack anchor for fruit/orchard AI.

What this unit is doing in the taxonomy

This is the on-farm production — orchard cell unit. Anchors:

• The CAAIN portfolio unit (units/caain-portfolio-canada.md).
• The Scale AI Agriculture unit (units/scale-ai-agriculture-canada.md).
• The Digital Supercluster Agriculture unit (units/digital-supercluster-agriculture-canada.md).
• The BC horticulture AI cluster (see Scan 3).

Why it matters for talks

1. Canadian orchard AI covers codling moth IPM (CropVue + OKSIR) and tree-fruitlet CV (Vivid Machines). The full value chain.
2. The OKSIR + CropVue partnership is a Canadian structural pattern — sterile-insect-release industry program + AI vendor. The AI tool reduces manual trap checks from labour-intensive to automated.
3. Vivid Machines is the cross-cluster funding anchor — appears in CAAIN, Scale AI, and Digital. The cross-cluster stacking is a Canadian structural feature.
4. The Surrey BC + Toronto ON + Calgary AB geographic spread is the orchard AI cluster.
5. The fruitlet placement per tree is a per-tree precision agriculture pattern — distinct from broadacre field-row AI.

Critical context

• CropVue’s CAAIN project is small ($310K total) but anchored on the OKSIR program (industry-programmatic deployment context).
• Vivid Machines’ deployment scale (number of orchards, acreage) is not publicly disclosed. The Quinton Gibson Ontario apple farm is the named anchor.
• The codling moth sterile-vs-wild classification requires the AI to detect subtle visual differences or applied markers. The marker-based approach is more accurate but requires physical marking of sterile moths.
• The tree-fruitlet placement per tree CV is computationally intensive (scanning every tree in an orchard). Operational throughput depends on orchard size and camera hardware.