Research-Driven Agriculture & AgTech
Dr. Yavuz Selim Sahin
AI-assisted crop protection, plant nutrition, and controlled-environment production
I provide agricultural consultancy focused on smart agriculture systems, supporting soil-based and soilless operations in plant nutrition, irrigation, climate control, and day-to-day production management.
Smart4Drip
VPDew
Climap
GrowLED
SmartMapHow I Work
I support this consultancy approach with software and data-driven tools. Smart4Drip is one of the applications I developed to make plant nutrition decisions more manageable through analysis, recipe optimization, and field data.
My consultancy is backed by active research in AI-driven precision agriculture and biological crop protection. With 20+ peer-reviewed publications spanning real-time pest detection with deep learning, crop disease monitoring through thermal imaging, biological control with entomopathogenic agents, and controlled-environment cultivation, I bridge the gap between scientific innovation and practical field solutions.
Consultancy Scope
- Soilless cultivation and controlled-environment operations
- Plant nutrition recipes and fertigation planning
- Climate, irrigation, and production optimization
- Field-data-based decision support and operational follow-up
Selected Research
- AI-based pest detection & segmentation (YOLOv8, deep learning)
- Thermal imaging for early crop disease diagnosis
- Biological control with entomopathogenic nematodes
- Controlled-environment cultivation & mushroom production
20+ peer-reviewed publications — view all on Google Scholar
Digital Tools
Decision-Support Applications
Smart4Drip
Plant nutrition planning and fertilizer recipe optimization
Smart4Drip is a decision-support application I developed to support my consultancy work by making plant nutrition management more analytical, measurable, and efficient.
The application combines water, crop, and production data to generate nutrition recipes according to crop variety and phenological stage. Users define the fertilizers they already have, and the system calculates suitable combinations and doses to standardize decision-making.
VPDew
Vapor pressure deficit, dew point, and humidity deficit guidance
VPDew is a climate-decision calculator I built to convert greenhouse temperature and relative humidity into actionable humidity deficit, dew point, and vapor pressure deficit targets.
The tool combines preset crop profiles with quick manual inputs so growers can move between density-based and kPa-based targets without interrupting daily operations. It is designed to be fast on both desktop and mobile.
Climap
Archive weather data on a map with charts and CSV export
Climap is a map-based archive climate explorer I built to turn Open-Meteo historical weather data into interactive charts and downloadable CSV exports for any location worldwide.
The tool calculates derived metrics — VPD, dew point, GDD, DLI, ET0, chill hours, frost days, and photoperiod — from raw station data. Charts include trend lines, shaded bands, and exportable high-resolution images with annotations. It works offline as a PWA and supports Turkish and English.
GrowLED
PAR/PPFD, RadSum, and DLI metrics for LED and sunlight scenarios
GrowLED is a light-decision calculator I built to combine PAR/PPFD, duration, and photon energy into actionable light metrics — RadSum, instantaneous irradiance, and DLI — for both LED and sunlight growing scenarios.
The tool lets growers and researchers toggle between grow-LED and sunlight contexts with configurable PAR energy share and photon energy inputs, making it fast to compare instantaneous intensity and cumulative radiation. It is designed as a PWA that runs offline on desktop and mobile.
SmartMap
Sentinel-2 NDVI/NDMI/NDRE for any user-drawn field, clipped to the polygon
SmartMap is a satellite plant-health and drought viewer I built on top of the Copernicus Sentinel-2 archive, surfacing NDVI, NDMI, NDRE, and true-color layers for any user-drawn field.
Designed for in-field agronomic decisions: draw a rectangle or polygon, pick a 2–4 week window, and the system automatically fetches the least-cloudy Sentinel-2 L2A scene. The image is clipped to the polygon shape and downloadable as PNG for reports. The backend rotates between multiple Sentinel Hub accounts to absorb monthly quota gracefully. Bilingual TR/EN UI.