Optical sorting (color sorting) uses high-speed cameras, AI imaging algorithms, and precision air ejectors to separate defective seeds, foreign grains, and unwanted contaminants that size, gravity, or magnetic cleaning equipment cannot remove. It delivers a uniformly clean, consistent raw material stream, which directly elevates the maximum achievable protein purity in downstream dry fractionation.
1. Core Working Principle of Optical Sorting for Pulses
- Monolayer feeding: Clean pulses fall in a thin, single-file curtain to ensure every seed is individually scanned with no overlapping particles.
- Multi-spectrum camera scanning: High-resolution RGB + near-infrared (NIR) cameras capture color, brightness, surface texture, and internal chemical signatures of each kernel.
- AI real-time analysis: The system compares each seed against reference templates for healthy, undamaged pulses. It flags rejects based on discoloration, blemishes, shape irregularities, or chemical composition.
- Pneumatic ejection: Micro high-pressure air jets blast defective/foreign particles into a waste chute instantly; intact clean pulses continue to downstream processing.
- Recirculation loop (two-pass sorting): The ejected waste stream is re-sorted to recover accidentally discarded good pulses and minimize raw material yield loss.
2. Key Contaminants Optical Sorting Eliminates (Other Cleaning Machines Cannot Remove These)
2.1 Moldy, stained, damaged pulses
- Contaminants: Black/brown mold discoloration, insect-bored kernels, rotten, water-damaged seeds, heat-stained pulses from harvesting/storage.
- Why these harm protein purity: Moldy seeds accumulate tannins, phenols, bitter off-flavors, and mycotoxins. When milled, they introduce dark pigmented fiber and polyphenols that remain trapped in the final protein concentrate, lowering sensory quality and diluting protein percentage. These contaminants match the size and density of good pulses, so sieves and destoners cannot separate them.
- Optical sorting impact: Removes >99% discolored defective seeds upfront, producing pale, neutral-tasting protein powder with minimal residual phenolic impurities.
2.2 Cross-contaminant foreign grains & weed seeds
- Contaminants: Wheat, corn, barley, oilseeds, dark weed seeds mixed with target pulses (pea/faba bean/lentil).
- Why these harm protein purity: Foreign grains carry high starch, different fiber profiles, and inconsistent protein content. Mixed foreign seed flour creates a third interfering particle population during air classification, disrupting the critical bimodal protein/starch particle distribution and raising residual carbohydrate levels in the finished concentrate.
- Optical sorting impact: Identifies shape and color differences to strip all cross-species grain contamination before dehulling and milling.
2.3 Partial hull fragments & tough fibrous seed pieces
Thin hull slivers and broken seed coat pieces often pass through gravity destoners and vibrating screens due to matching density with small cotyledons. Optical cameras detect dark hull color against pale cotyledons and eject hull fragments, reducing total fibrous load entering the pin mill.
2.4 Unripe, low-protein underdeveloped seeds
Immature pulses have thinner cotyledons, lower native protein content, and higher carbohydrate levels. NIR spectroscopy built into optical sorters detects internal chemical composition to reject low-protein immature seeds, raising the baseline crude protein of the feedstock before fractionation.
3. Direct Mechanisms That Boost Final Protein Concentrate Purity
3.1 Reduces baseline fiber and non-protein impurities in feed
Hull fragments, weed seeds, and damaged grain are high in insoluble fiber and carbohydrates—primary diluents of protein content. Optical sorting eliminates these materials early, lifting the baseline protein of clean cotyledons from ~22–24% up to 25–28% before milling. This reduces the separation burden on pin milling and air classification, enabling the plant to reach 58–65% high-purity protein concentrates instead of being capped at 42–50%.
3.2 Eliminates irregular particles that disrupt milling PSD control
Moldy, cracked, and immature seeds have fragile, uneven cellular tissue. When sent to the pin mill, they over-grind easily, generating excessive fine starch micro-fines (10–20 μm). These starch fines overlap the size range of protein aggregates and slip past dynamic classifier wheels, permanently increasing residual starch in the final protein product.
Optical sorting removes irregular seeds, creating uniform feedstock for the pin mill. The mill runs at calibrated mild rotor speeds to produce the ideal D50=13–25 μm flour with minimal broken starch, preserving the critical size gap between protein bodies and intact starch granules for cleaner air classification.
3.3 Simplifies aerodynamic sorting for dynamic air classifiers
Without optical sorting, the flour contains three particle groups: free protein, intact starch, and fibrous foreign particle agglomerates. Fiber clumps distort airflow patterns inside the classifier chamber, causing unselective carryover of starch into the fine protein stream.
By removing fibrous contaminants pre-milling, optical sorting creates a clean bimodal particle mixture, allowing the classifier wheel to precisely tune the cut-point (10–22 μm) without interference. Residual starch in finished protein drops by 8–14%.
3.4 Minimizes closed-loop recycle volume and starch over-grinding
Mixed middling agglomerates (protein + starch + fiber) are continuously recycled back to the pin mill in standard dry fractionation lines. Every regrinding pass creates more fragmented starch micro-fines, worsening purity. Optical sorting cuts fiber-driven middling recycle mass by 30–40%, limiting repeated regrinding and permanent starch contamination of protein.
3.5 Improves food-grade quality and reduces anti-nutrient carryover
Moldy, discolored seeds carry high phytate, tannins, and mycotoxins. Optical sorting removes these materials, so the finished protein concentrate has lower astringency, lighter color, fewer anti-nutrients, and complies with food safety mycotoxin limits. This eliminates the need for extra polishing steps (like multi-stage air classification or electrostatic separation) to offset quality defects.
4. Quantifiable Performance Comparison (With vs Without Optical Sorting)
| Processing Metric | Raw pulses cleaned only by sieve/destoner/magnet | Full pre-cleaning + optical sorting |
|---|---|---|
| Baseline feed protein | 22–24% | 25–28% |
| Max single-stage air classification protein purity | 42–48% | 48–55% |
| Max two-stage polished protein purity | ≤56% | 58–65% |
| Residual fiber in final protein concentrate | 9–15% | 3–6% |
| Starch micro-fine contamination level | High | Low |
| Off-color / bitter flavor risk | Significant | Minimal |
5. Position of Optical Sorting in the Pre-Cleaning Flow
Raw pulse intake → Vibrating grading sieve + aspirator → Gravity destoner → Magnetic separator → Optical color/NIR sorter → Buffer silo → Tempering → Roller dehuller → Pin mill → Air classification fractionation
6. Summary of Optical Sorting’s Purity Enhancement Role
- Removes size/density-matched contaminants (moldy seeds, foreign grain, hull fragments) that mechanical cleaning equipment cannot separate.
- Raises the baseline crude protein of feedstock by eliminating low-quality, carbohydrate/fiber-rich defective material.
- Stabilizes uniform pin mill output particle size distribution, preventing over-generation of problematic starch micro-fines that contaminate protein concentrates.
- Eliminates fibrous interferences that distort air classifier airflow and sorting accuracy, tightening separation efficiency between protein and starch.
- Reduces anti-nutrients and sensory defects, delivering pale, neutral-tasting high-purity protein powder suitable for premium food applications.
Optical sorting is a high-value pre-treatment step that removes the root sources of protein dilution and contamination long before milling and fractionation, unlocking the full purity potential of dry pulse protein processing.
