Smart Al-Enabled Rapid Feed and Silage Quality Testing System for Dairy Farmers
Ministry of Fisheries, Animal Husbandry & Dairying · Agriculture, FoodTech & Rural Development · Software
Pick a defensible subset — visible spoilage, mould and silage pH — and state clearly which analytes you are not attempting and why, because promising aflatoxin detection from a phone photograph is the kind of claim that ends a pitch.
What it actually is
Cattle feed quality directly drives milk yield, but lab testing is expensive and out of reach for rural farmers, so adulterated or mouldy feed goes undetected. The ask is a rapid, portable, affordable way to assess feed and silage quality on the farm and turn the result into advice. The description lists a very long menu of things it would like measured.
What to build
A phone-based feed assessment tool paired with a low-cost sensing module, where a farmer photographs a feed or silage sample and the system classifies visible quality problems — fungal growth, mould coverage, foreign matter, sand and silica contamination, and spoilage indicators in silage — while a handheld near-infrared module estimates the bulk nutritional parameters the description names, with a pH strip reading for silage fermentation quality, all producing an instant multilingual advisory telling the farmer whether to feed, blend or discard, working offline and syncing to a cooperative dashboard when connectivity returns.
Smallest thing that wins the room
Photograph two silage samples side by side, one sound and one visibly mouldy, and show the app grading them differently with the spoilage regions highlighted and a plain-language feeding advisory in the farmer's language.
How crowded this one gets
A guess, projected from the 2025 statements — the last year where both the submission counts and the winners were published.
Quieter than 52% of the 226 · #110 of 226 by expected field
A normal-sized field. Your idea has to be good, not miraculous.
Why: central ministry statements sat below the average.
This is a guess, not a fact
Nobody has published 2026’s numbers yet. This is an analysed estimate from last year’s pattern, so please do not take it as the truth — check the live counter on the SIH portal before you decide anything. The range covers the middle half of likely outcomes, so one statement in two lands outside it. Entry closes at 500 ideas per statement, so no range goes past that — a statement that reaches the cap fills and shuts rather than drawing an unlimited crowd. The model reads only three things a team can see before choosing — software or hardware, the theme, and what kind of body posted it — and those explain about a quarter of the variation in last year’s field sizes (R² 0.25 on held-out statements). Trust the band more than the number, and the ordering more than either. It cannot see how good your idea is, which is the part that actually decides it.
The scores
The number is the shorthand. The line under it is the reason.
Acceptance potential
2/5A kitchen-sink description that guarantees shallowness: the analytes that matter most for feed value and safety are exactly the ones no affordable sensor can measure, so a team will demo mould detection from photographs while the statement asked for aflatoxin quantification, and an animal nutrition judge will notice the gap immediately.
Feasibility
2/5The description asks for nine feed analytes and six silage parameters, and most of them — crude protein, fibre, energy value, mineral deficiency, urea adulteration and especially aflatoxin — cannot be measured by a camera at all, requiring near-infrared spectroscopy or immunoassay hardware that is neither low-cost nor available to a student team.
Innovation scope
3/5The description enumerates the analytes, the candidate technologies and the expected features so thoroughly that the design space is mostly filled in, leaving room mainly in which subset you choose to attempt honestly.
Clarity
3/5Everything is listed but nothing is prioritised — the analyte menu is offered as things the system 'may detect', so a team cannot tell which measurements are actually required and which are aspirational.
Effort
MassiveTaken literally this is a spectroscopy instrument, a computer vision model, a biosensor path, an offline multilingual app and a cloud dashboard, which is several products rather than one.
Demo-ability
MediumThe visual mould and contamination grading demos well from a photograph, but the nutritional and toxin numbers — the parameters a dairy judge actually cares about — cannot be shown working.
In its favour
- Green flag: Visible spoilage, mould coverage and gross foreign matter genuinely are detectable from a photograph, so there is an honest, buildable core inside the oversized ask
- Green flag: You can create your own labelled dataset cheaply by photographing feed and silage samples at varying spoilage stages
- Green flag: The offline and multilingual requirements are explicitly stated, so implementing them answers the rural-access question by specification rather than improvisation
- Green flag: Silage pH by colorimetric strip with reference-scale correction is a real measurement you can do with nothing but the phone camera
Against it
- Red flag: Aflatoxin cannot be detected by a camera under any circumstances — it requires immunoassay or chromatography, and any app claiming to detect it from a photograph is making a food-safety claim it cannot support
- Red flag: Crude protein, fibre and energy value need near-infrared spectroscopy, and consumer NIR modules are both costly and calibrated for food rather than fodder, so accuracy on feed will be poor
- Red flag: The description lists fifteen measurable parameters as optional suggestions, which invites teams to promise everything and deliver photograph-based grading
- Red flag: Urea adulteration is a chemical test, and getting it wrong means telling a farmer that adulterated feed is safe
What you will be writing
- CNN mould and contamination classification (EfficientNet-Lite)
- Consumer NIR module (SCiO-class) for bulk composition
- OpenCV colour and texture feature extraction
- pH strip colorimetric reading with reference correction
- TFLite offline inference
- Flutter multilingual app with cooperative sync
- Animal nutrition
- Rapid field diagnostics
- Agricultural quality assessment
Prior art to read before you start
portable feed quality assessment · visual spoilage and contamination detection · farmer advisory from sample analysis
Analysed by Claude Opus. Every score above is a judgment call with its reasoning attached — kindly cross-check this against the official statement on the SIH portal before your team commits to it.