Crop vision
Computer vision on phone, fixed-camera and drone images to detect disease, pests, nutrient stress and growth stage early.
Agricultural Services · InCrop
AI that grows more food with less water and less energy.
InCrop is LIDAVI’s artificial-intelligence practice for agriculture. We turn sensor, camera and weather data into decisions — when to irrigate, what is diseased, how much light a crop really needs — for open fields, greenhouses and AI-run vertical farms in desert and very cold countries.
~71%
of global freshwater withdrawals go to agriculture (FAO AQUASTAT, 2022)
~92% less
water per kg lettuce: hydroponic vs. field, desert-site study [1]
99% → 31%
crop-disease AI accuracy in the lab vs. on unseen field images [6] — why we validate on your farm
8
AI capabilities, delivered on one data platform with farmer-facing apps
AI capabilities
Eight ways InCrop puts AI to work on the farm
Each capability is a working module, not a slide: it ingests real farm data, produces a decision or alert, and is measured against a KPI you care about.
Precision irrigation
Soil-moisture sensing, weather forecasts and evapotranspiration models decide when and how much to irrigate — per zone, not per farm.
AI climate control
Model-predictive control of light, temperature, humidity and CO₂ in greenhouses and vertical farms to hit yield targets at the lowest kWh per kg.
Yield & harvest forecasting
Machine-learning forecasts of yield and harvest date from sensor history, images and weather, so sales and logistics can plan ahead.
Nutrient optimisation
Fertigation recipes adjusted from EC, pH and plant response data, reducing fertiliser use and nutrient run-off.
Energy optimisation
Tariff-aware scheduling of lights, pumps and HVAC; heat-recovery control; continuous kWh-per-kg tracking with anomaly alerts.
Farmer advisory assistant
A multilingual, voice-capable assistant that answers agronomy questions grounded in your farm’s own data and vetted local guidance.
Demand & supply-chain AI
Demand forecasting and harvest planning that match production to orders, cutting waste of perishable produce.
Where we apply AI
From open fields to fully indoor farms
Open-field farms
Smartphone crop scouting, soil and weather stations, zone-based irrigation scheduling and an advisory assistant that works in the farmer’s language — designed for patchy connectivity.
- Crop vision & advisory
- Precision irrigation
- Yield forecasting
Greenhouses & polyhouses
Retrofit sensing and AI climate control onto existing structures, with fertigation optimisation and disease early-warning from fixed cameras.
- AI climate control
- Nutrient optimisation
- Disease early warning
AI-run vertical farms
Fully controlled indoor farms for places where outdoor farming is hardest: hot deserts and very cold countries. AI runs the climate, light and water loop.
- Closed-loop water
- kWh-per-kg optimisation
- Automated operations
Platform architecture
Sensor to decision, edge to cloud
InCrop is built by the same teams that deliver LIDAVI’s IoT, cloud and app projects. Time-critical decisions run on edge devices at the farm so they keep working when the internet does not; training, analytics and fleet management run in the cloud.
- SenseSoil, climate, water-quality and energy sensors; fixed cameras, phones and drones.
- Decide at the edgeOn-site gateways run vision and control models with millisecond response and offline operation.
- Learn in the cloudData lake, model training and retraining on your farm’s data, monitoring for model drift.
- Act through appsiOS, Android and web dashboards for alerts, approvals, crop plans and KPI reports.
Flagship programme
AI-run vertical farming for desert and very cold countries
Two opposite climates, one problem: fresh produce is imported, expensive and short-lived. Fully enclosed farms decouple growing from the weather — and AI keeps their two biggest costs, water and energy, under control.
Hot desert climates
Gulf states, wider Middle East & North Africa, arid regions of India, Australia and the Americas
- Scarce fresh water; much of it desalinated, which costs energy
- Extreme heat and solar load push cooling demand up for most of the year
- Dust and sand ingress threaten equipment and food safety
- High dependence on imported vegetables and herbs
Very cold climates
Nordic countries, Iceland, Canada, Alaska, Siberia, high-altitude regions
- Short or no outdoor growing season
- Little or no natural daylight in winter months
- Heating is the dominant energy load for any enclosed growing
- Long supply chains: produce arrives after days in transit
Climate playbooks
One platform, engineered for your climate
Opaque, insulated envelope
No glazing: the desert sun stays outside. A sealed, highly insulated building removes solar heat gain so cooling only handles heat from lights and plants.
AI-scheduled solar and cooling
Forecasts of solar output, outdoor temperature and tariffs decide when to run lights and chillers, including night-time “lights-on” to use cooler air.
Every drop recovered
Plant transpiration is condensed on the dehumidification coils and returned to the nutrient loop — the key reason indoor farms use so little water.
AI water-quality control
Desalinated, municipal or treated grey water is polished by RO; models watch EC, pH and pathogens and adjust dosing and UV treatment automatically.
Thermal storage
Chilled-water storage charged at the cheapest or cleanest hours, discharged when the farm needs cooling most.
Dust-proof operations
Multi-stage filtration, air-locks and positive pressure, with predictive filter-maintenance alerts from pressure-drop data.
Turn LED heat into building heat
In the cold, grow-light heat is an asset. Heat pumps recover it to warm the farm, offices or neighbouring buildings; AI balances light and heat demand.
Plug into clean local energy
Design around hydro, geothermal or district heat where available — Iceland, for example, generates almost all of its electricity from renewables [5].
Super-insulated, airtight shell
Cold-climate envelope, thermal-bridge-free detailing and heat-recovery ventilation keep heating demand low even at deep sub-zero temperatures.
Polar night is irrelevant
Fully artificial lighting gives identical crop cycles in December and June. AI-tuned light recipes, not latitude, set the yield.
Hybrid where daylight helps
For sites with useful summer daylight, our energy model compares a greenhouse with supplemental LEDs against a closed farm and recommends the lower-energy option.
Remote operation
Snow-load rated structures, freeze-protected water systems and remote AI monitoring for isolated communities and camps.
Water conservation plan
Keep water in the loop — and let AI find every leak
In open fields, much irrigation water is lost to evaporation, drainage and run-off. In a sealed farm, transpired water stays in the room, so we capture it and reuse it. Outdoors, AI irrigates only what each zone needs.
- Recirculating nutrient solutionZero-discharge hydroponics; drain-to-waste only by exception and treated before release.
- Condensate recoveryCooling-coil condensate is collected, filtered, UV-treated and returned to the reservoir.
- AI irrigation & dosingSoil-moisture, EC, pH and weather data drive zone-level irrigation and nutrient decisions.
- Anomaly detection on water balanceModels compare expected and metered flows per zone and flag leaks or blocked emitters within minutes.
Energy reduction plan
Honest about energy — and using AI to cut it
Electricity is the main trade-off of indoor farming: the desert-site study behind the water savings above found hydroponic production used far more energy per kg than field growing [1]. That is why every InCrop design starts with an energy model and runs on AI optimisation.
- Photon-efficient LEDsFixtures chosen on tested efficacy (µmol/J); research puts current LED limits at 3.4–4.1 µmol/J [7].
- AI light recipesDaily light integral, photoperiod and spectrum tuned per crop stage — dimming instead of over-lighting.
- Heat recoveryHeat pumps move LED and equipment heat to where it is useful instead of fighting it.
- Load shiftingForecast-driven scheduling moves flexible loads to cheap, low-carbon or solar-rich hours.
- Digital twin & sub-meteringPer-zone meters feed a live kWh/kg dashboard and a simulation used to test changes before applying them.
White papers
Our thinking, in depth
Papers from the InCrop team. Each separates established evidence from industry claims and from our own design targets.
White paper 01
Water
Water-Smart Vertical Farming for Arid Regions
How closed-loop hydroponics, condensate recovery and AI-driven dosing can produce fresh greens where water is scarcest — and what it costs in energy.
- Version 1.0
- ~8 min read
- PDF available
White paper 02
Energy
Cutting kWh per Kilogram: AI Energy Optimisation for Indoor Farms
Lighting, HVAC and heat recovery in hot and cold climates, and how forecasting, model-predictive control and digital twins reduce energy per kg.
- Version 1.0
- ~7 min read
- PDF available
White paper 03
Crop AI
From Pixels to Decisions: Practical AI for Crop Health and Farm Operations
Why lab accuracy rarely survives the field, and a deployment approach for crop vision, forecasting and farmer advisory AI that does.
- Version 1.0
- ~6 min read
- PDF available
What the evidence says
Established facts, industry claims and our targets — clearly labelled
| Statement | Figure | Status | Source |
|---|---|---|---|
| Share of global freshwater withdrawals used by agriculture | ~71% (2022) | Established | FAO AQUASTAT [2] |
| Water use, lettuce, Yuma (Arizona desert): hydroponic greenhouse vs. conventional field | 20 vs 250 L/kg | Peer-reviewed | Barbosa et al., 2015 [1] |
| Yield, same comparison | 41 vs 3.9 kg/m²/yr | Peer-reviewed | Barbosa et al., 2015 [1] |
| Energy, same comparison — the trade-off | ~82× higher | Peer-reviewed | Barbosa et al., 2015 [1] |
| Greenhouses are currently more energy-efficient than closed plant factories; closed systems fit best where water or land is scarce | Qualitative | Peer-reviewed | Graamans et al., 2018 [3] |
| Deep-learning disease classifier: accuracy on held-out lab images vs. images from other sources | 99.35% vs 31.4% | Peer-reviewed | Mohanty et al., 2016 [6] |
| Efficacy limit of LED fixtures with current technology | 3.4–4.1 µmol/J | Peer-reviewed | Kusuma et al., 2020 [7] |
| “Vertical farms use up to 98% less water than traditional agriculture” | up to 98% | Industry claim | World Economic Forum, via trade press [4]; crop-dependent |
| Iceland electricity from renewable sources (hydro + geothermal) | ~100% | Established | National statistics, summarised in [5] |
| InCrop: water, energy and AI-model accuracy tracked live and reported to the client monthly | per contract | Our target | Set during feasibility, verified on site |
Results vary with crop, site, energy prices and system design. We quote project-specific figures only after a feasibility study and on-site model validation — never from marketing averages.
- Barbosa G.L. et al. (2015). Comparison of Land, Water, and Energy Requirements of Lettuce Grown Using Hydroponic vs. Conventional Agricultural Methods. Int. J. Environ. Res. Public Health 12(6), 6879–6891.
- FAO AQUASTAT — agricultural water as a share of total withdrawals.
- Graamans L. et al. (2018). Plant factories versus greenhouses: Comparison of resource use efficiency. Agricultural Systems 160, 31–43.
- Greenhouse Canada, Aug 2023, reporting a World Economic Forum study.
- Energy in Iceland.
- Mohanty S.P., Hughes D.P., Salathé M. (2016). Using Deep Learning for Image-Based Plant Disease Detection. Frontiers in Plant Science 7.
- Kusuma P., Pattison P.M., Bugbee B. (2020). From physics to fixtures to food: current and potential LED efficacy. Horticulture Research 7.
Responsible AI
AI the grower can trust
Your data stays yours
Farm data is owned by the farm. We use it to train your models only with written consent, and never sell it.
Validated on your farm
Every model is tested on data from your own site and crops before it is allowed to act, and re-checked as conditions change.
Human in the loop
Recommendations come with a reason and a confidence level; high-impact actions need a grower’s approval.
Safe fallbacks
If a model or connection fails, control returns to proven rule-based settings and the team is alerted.
Offline-first
Edge devices and apps keep working without connectivity and sync when it returns.
In the farmer’s language
Apps and advisory designed for local languages, voice and low-literacy use where needed.
What to grow indoors
Crops chosen for economics, not hype
Indoor vertical farming pays off for fast-cycling, high-value, perishable crops. Staple grains are not a fit today — for those, our AI works in the open field instead.
Leafy greens
Lettuce, spinach, kale, pak choi, rocket.
Strong fitCulinary herbs
Basil, coriander, mint, parsley.
Strong fitMicrogreens
Short cycles, premium retail and HoReCa.
Strong fitSeedlings & transplants
Disease-free starts for greenhouses and fields.
Strong fitStrawberries
Promising, but pollination and energy need careful design.
Case-by-caseTomatoes & cucumbers
Usually better in an AI-controlled greenhouse.
Case-by-caseFodder
Hydroponic fodder for livestock in arid regions.
Case-by-caseGrains & staples
Not indoors — use field AI: irrigation, scouting, forecasting.
Field AI insteadKPIs on your dashboard
What we measure, from day one
L / kgWater per kg of saleable produce
kWh / kgElectricity per kg
g / molBiomass per mole of light
kg / m² / yrYield per growing area
Model accuracyOn your own farm’s data, tracked over time
How we deliver
From data audit to AI that runs the farm
- Phase 1
Discover
Site, climate, water and energy audit; data and connectivity assessment; use-case prioritisation and business case.
- Phase 2
Pilot
Sensors, cameras and one or two AI use-cases on a pilot zone or container farm; models validated on your data.
- Phase 3
Build & integrate
Scale the platform: edge gateways, cloud, apps and — for vertical farms — racks, lighting and HVAC.
- Phase 4
Operate & improve
Training, monitoring, model retraining and continuous tuning to keep cutting water and energy per kg.
Engagement models
Choose how you want to work with us
AI consulting & design
Use-case discovery, data strategy, specifications and tender support.
AI platform subscription
Our sensing, edge and app platform, configured for your farm.
Turnkey AI vertical farm
We design, build and commission; you own and operate — or we operate and transfer.
Retrofit & digitise
Add sensing, AI control and dashboards to an existing farm or greenhouse.
Start with a conversation
Let’s scope an InCrop AI pilot
Tell us your location, crops, farm type and the data you already collect. We will come back with the AI use-cases worth piloting first and the questions that matter.
Call +91 98945 46777