Grün VFG Series AI-Driven Sorting System

Dry glass cullet recycling faces a sorting challenge: mixed container glass, MRF glass, or flat glass cullet is mixed with glass of various hues, opaque impurities, ceramics, stones, and heat-resistant glass. These materials differ in color, brightness, and transparency, and accurate identification of colors and impurities often determines the final commercial value. For glass recyclers, the ability to efficiently classify glass by color and type, while detecting and removing non-glass impurities, is the key to producing high-purity, furnace-ready cullet.

The Grün VFG 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 target glass types from waste streams within milliseconds.

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).

● 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 Design

The chutes are manufactured from a special alloy and optimized for the high hardness and strong abrasiveness of glass material, effectively resisting glass-induced wear, reducing maintenance costs, and ensuring long-term stable operation under high-throughput conditions.

● 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.

● Multi-Specification Model Portfolio

Offers configurations with 4 and 6 chutes to meet varying throughput requirements.

Typical Application Scenarios

✦ Glass Recycling Plants

✦ Bottle-to-Bottle Recyclers

✦ Solar Glass Producers

Specifications