PHOENIX‑AI Double‑Layer Intelligent Sorter

Accurately identify discoloration and hard‑to‑detect mildew features on garlic cloves. High‑dimensional HDR cameras capture 2D images of garlic cloves to precisely locate defect positions. Deep‑learning AI recognition mode easily identifies imperfect, broken‑skin and skin‑retained garlic cloves. Double‑layer structural design realizes multi‑stage screening for garlic cloves.

High‑Dimensional HDR Camera: Superior Imaging · Enhanced Performance

Compact machine‑vision system integrating image acquisition and processing. Under motion capture mode, it accurately tracks flowing garlic cloves and captures subtle surface features, meeting stringent sorting standards for garlic cloves.

Distortion Correction Technology: Ultra‑Low Distortion · Higher Precision

Optical design and algorithm‑driven image distortion correction restore the true geometry of garlic cloves, avoiding misjudgment of particle size and surface blemishes.

AI Sorting Solution: Precise Algorithm · Powerful Computing Support

Mimics human “eye‑and‑brain” recognition logic. Resists interference from fragments and debris. It precisely extracts surface features of garlic cloves and compares learned data to identify skin‑retained, mildewed, yellowed and damaged cloves. Dedicated AI NPU chip delivers robust computing power and accelerated inference for high‑speed garlic clove inspection.

Keyetech Cloud Platform: Model Factory · Brain‑Like Intelligence

Self‑developed special detection model library for garlic cloves, with over 30 online‑available models to fit multi‑grade garlic clove sorting requirements.

Dynamic Ejector Valve System: Cope With Variable Throughput

Intelligently controlled multi‑channel parallel valves deliver high‑speed, high‑sensitivity microsecond‑response rejection of defective garlic cloves. Flexible actuation ensures high yield and high efficiency even under high‑capacity garlic clove processing conditions.