Grün HDF-G Series AI-Driven Sorting System

Glass cullet recycling inherently presents a sorting challenge, and the situation becomes even more difficult when the material is damp: moisture on glass surfaces causes unpredictable light refraction, obscures color characteristics, and causes contaminants to adhere, making reliable identification far more challenging than in dry conditions. Mixed container glass, MRF glass, or flat glass cullet contains a variety of glass hues, opaque impurities, ceramics, stones, and heat-resistant glass, and consistent classification of these materials becomes even more difficult under damp conditions. For recyclers that need to process dry, damp, and mixed glass cullet, the ability to maintain high-precision sorting performance regardless of moisture content is the key to producing high-purity, furnace-ready cullet.

The Grün HDF-G Series is an intelligent machine-vision-based sorting system tailored for this exact scenario. It employs high-resolution RGB (HR-RGB) VIS imaging and proprietary AI deep learning algorithms to perform real-time scanning and surface feature analysis of dry container glass, MRF glass, or flat glass cullet with a particle size of 4–70 mm. Combined with high-performance air ejectors, the system efficiently sorts specific glass types from waste streams within milliseconds, whether the material is dry, damp, or mixed glass cullet.

Empowered by AI-driven surface feature recognition algorithms, the system can conduct deep learning training based on real production data, continuously optimizing the machine vision detection accuracy. This means the equipment’s performance is not fixed at the time of factory shipment; instead, it continuously adjusts and improves with the accumulation of production data, delivering stable and reliable sorting performance to help recyclers significantly improve cullet purity, remove impurities, consistently deliver recycled glass that meets the strict quality specifications of glass manufacturers, and command a premium in the market.

Technical Features

● High-Resolution RGB (HR-RGB) VIS Imaging

Clearly presents surface features of the material, providing reliable image input for AI algorithms. Efficiently identifies glass colors (clear, amber, green, etc.) and a wide range of contaminants (ceramics, stones, porcelain, metals, and plastics).

● Innovative Dual-Belt Sorting Structure

High-sensitivity sorting is performed on the upper level to establish extremely high purity of the main material stream; secondary re-inspection of upper-level rejects is conducted on the lower level, ensuring effective recovery of high-value fractions mixed within.

● Proprietary AI Deep Learning Algorithms

Trained against comprehensive identification and sorting standards, the system uses deep learning to replicate the surface-feature judgment logic of manual picking. The AI model can be continuously optimized with new production data, driving ongoing improvements in sorting accuracy.

● Wear-Resistant Belt Design

Employs a highly wear-resistant conveyor belt custom-built for glass cullet sorting conditions, stably accommodating feeding requirements of dry, damp, and mixed glass cullet. The belt surface is specially treated for abrasion and tear resistance, maintaining stable performance even under long-term friction and impact from glass material.

● High-Intensity LED Illumination

Maintains a constant imaging environment with clear, stable images unaffected by ambient light variations, ensuring consistent detection results across different periods and batches.

● High-Performance Air Ejector

Equipped with in-house, high-power, high-speed air ejectors, the system delivers high-precision material rejection and ensures long-term stable operation even under high-throughput sorting conditions.

Typical Application Scenarios

✦ Glass Recycling Plants

✦ Bottle-to-Bottle Recyclers

✦ Solar Glass Producers

Specifications