# Crude Fibre - Feed Inputs

> Crude fibre measurement for agricultural feed inputs - early feasibility work underway to assess calibration dataset requirements and partner appetite.

- Canonical URL: https://spectrify.ai/research-pipeline/crude-fibre-feed-inputs

- Sector: Sector 01 · Food & Agriculture
- Status: Early Exploration
- Tags: NIR, Feed, Fibre, Livestock, Agriculture

Crude fibre is a key nutritional parameter for agricultural feed quality certification - used by feed mills, livestock farms, and regulatory bodies to verify that raw material inputs (rice straw, sugarcane bagasse, coconut pith, etc.) meet formulation specifications. Current measurement by Weende proximate analysis takes 6–8 hours. NIR-based crude fibre prediction is well-established for homogeneous matrices (e.g., single-grain feeds) but less characterised for the heterogeneous mixed-input matrices common in Sri Lankan and South Asian livestock farming.

## What We've Established

### Technical Feasibility Assessment

Published NIR calibrations for crude fibre in standardised feed matrices (wheat straw, hay) show R² > 0.90. Heterogeneous inputs with highly variable particle size and composition present greater calibration challenges and may require particle size normalisation or matrix-specific sub-models.

## Technical Approach

Feasibility review is ongoing. The proposed approach is to begin with one feed input type (likely rice straw or coconut pith, due to volume and availability) before expanding. Reference method is ISO 6865 crude fibre determination. Calibration dataset requirements are estimated at 150–300 samples for a single matrix type.

## Where We Stand

Early feasibility exploration. No partner commitments and no dataset collection. The primary objective at this stage is to identify one or two feed mills or livestock operations willing to serve as data collection and validation partners.

## Next Steps

- Define target feed input matrices in priority order
- Identify feed mill or research institute partners for sample access
- Conduct a small-scale spectral feasibility scan on representative samples before committing to full calibration

## Related Projects

- [Milk Quality](https://spectrify.ai/research-pipeline/milk-quality) (Proposal Sent): Rapid compositional analysis for raw milk - fat, protein, and adulteration screening to replace slow wet-chemistry tests at collection centres.
- [Rubber - Dry Rubber Content](https://spectrify.ai/research-pipeline/rubber) (Awaiting Response): Dry rubber content (DRC) measurement at point of purchase - enabling transparent, data-driven pricing between smallholders and processors.
- [Black Pepper & Spices](https://spectrify.ai/research-pipeline/black-pepper-spices) (Early Exploration): Spectral signature mapping for black pepper and mixed spices - dataset acquisition phase. Adulteration detection and grade classification are primary targets.

[Research Pipeline](https://spectrify.ai/research-pipeline)