AI for Coffee Variety Recognition

In this recorded talk, we unpack the practical “AI stack” for coffee: smartphone phenotyping, computer vision, deep learning, and farm data workflows that help coffee breeders, agronomists, and growers move from intuition to measurable, repeatable, scalable monitoring — from plot to tree to bean. You’ll hear a mobile-first perspective shaped by collaboration with coffee R&D partners (including national coffee research & development teams and university collaborators), and focused on what actually works on real farms: uneven light, fast scouting, limited connectivity, and the need to link images to context (tree ID, plot, stage, notes, actions). What AI can support in coffee (end-to-end overview): ✅ Coffee tree inventory + plot mapping (link each tree to photos + attributes) ✅ Coffee variety identification + phenotype tracking across growth stages ✅ Flower counting → berry/cherry development → ripeness estimation → harvest timing ✅ Quality control concepts for coffee beans (image-based inspection & grading) ✅ Integrated Pest Management (IPM) scouting using a tree-level map + history ✅ Better workflows than “just Excel” (connected records, photos, measurements) ✅ Data collection protocols (avoiding blur, angle issues, and noisy labels) ✅ Offline/online operation for farms with weak internet (edge AI / on-device inference) ✅ AI literacy, trust, and data ownership — who owns the images and the insights? If you’re building agritech, doing coffee breeding research, running a farm/co-op, or supporting extension services, this video gives you a clear picture of the components, constraints, and design choices behind real-world AI for coffee production. 🕒 Timestamps (brief chapters) 0:00 Welcome + why AI matters for coffee 0:55 What “AI” really means in coffee breeding & production 2:35 From leaf/plant measurements to richer phenotyping data 4:05 Plot setup, tree inventory & photo-linked records 5:40 Coffee variety differentiation across stages (flowers / no flowers) 8:35 Scaling imagery: smartphone + tablet + drone workflows 9:15 Tree-level maps for scouting & IPM 10:20 Yield components: flowers → berries → harvest expectations 13:50 Ripeness stages & harvest-time decision support 15:50 Practical constraints: camera quality, blur & field realities 18:00 Coffee quality + bean inspection concepts 19:45 Why spreadsheets break (and what connected workflows add) 21:40 Data collection protocols + offline mode (edge AI) 23:25 AI literacy, trust & ownership 26:00 Computer vision + deep learning in daily farm operations 30:10 Closing remarks + collaborations (incl. Cavite State University mention) 💬 Question for you: Would you use AI tools like this on a coffee farm — YES or NO? And which task matters most: variety ID, flower/berry counting, ripeness, pest scouting, or quality grading? 📌 Related topics: AI in Agriculture, precision agriculture, digital farming, farm digitisation, smartphone-first computer vision & practical edge AI deployment. 🔎 SEO keywords (search terms) AI for coffee, coffee farming AI, coffee AI, coffee breeding AI, coffee research, coffee R&D, coffee agronomy, coffee extension, coffee phenotyping, plant phenotyping, smartphone phenotyping, mobile phenotyping app, computer vision agriculture, deep learning agriculture, machine learning crops, object detection, image segmentation, data collection protocol, dataset building, image annotation, data labeling, model training, on-device inference, offline AI, edge AI agriculture, Android AI app, coffee tree inventory, tree mapping, plot mapping, geotagging, GIS agriculture, farm digitisation, decision support system, integrated pest management IPM, field scouting, coffee flower counting, flower detection, coffee berry counting, coffee cherry counting, ripeness detection, maturity stages, harvest timing, yield estimation, yield forecasting, quality control, coffee bean inspection, grading, post-harvest, traceability, supply chain, sustainability, climate-smart coffee, drone mapping, UAV agriculture, remote sensing, NDVI, multispectral imaging, hyperspectral imaging. 👍 If this was useful, like, subscribe, and share it with a coffee grower / breeder / agronomy team. 🧩 Mini checklist for getting started: Define the task (variety ID, counting, ripeness, QC). 2) Collect consistent images. 3) Label/verify. 4) Pilot on a small plot. 5) Measure errors and improve the protocol. 6) Think early about connectivity, privacy & data ownership. #drmarynakuzmenko #coffeefarming #coffeescience DR. MARYNA KUZMENKO Website: https://marynakuzmenko.com Linkedin:   / kuzmenkomaryna   Instagram:   / dr.maryna.kuzmenko   TikTok:   / dr.maryna.kuzmenko   X: https://x.com/mary_kuzmenko Facebook:   / dr.maryna.kuzmenko   __ 🎁👉 FREE online beginner-level course "AI in Agriculture: Practical Introductory Course" on Udemy: https://www.udemy.com/course/ai-in-ag...