Inspection System

IMAGE LG DL

Empties Inspection with Deep Learning

Automated returnable-crate inspection with a Deep Learning software module for bottle-type classification.

Empties' Crate InspectionUp to 7,500 cph

Containers: Glass, Ref Pet, Crate
Line position: Pre unpacker

Send inquiry

Application

Intact Crates Before Unpacking

Ensure that only intact and unpackable returnable crates reach the unpacking unit. Add Deep Learning support for classifying recurring bottle-type characteristics in crate contents.

Line positionPre unpacker
ProductsGlass, Ref Pet, Crate
TasksEmpties' Crate Inspection

Operating principle

Camera Inspection with Deep Learning

High-resolution cameras, specialized optics and dedicated illumination modules inspect crate contents and visible damage. The Deep Learning software module learns recurring crate-content characteristics from recorded images and assigns them to bottle types.

PerformanceUp to 7,500 cph

Benefits

Automated unpackability checks help keep unsuitable crates away from the unpacking unit.
Early detection of damaged or foreign crates supports more stable downstream workflow.
Automating the inspection process can reduce manual inspection effort and related labour costs.
Continuous crate inspection helps reduce downtime and damage caused by non-unpackable empties.
The Deep Learning module supports easier teach-in and higher detection accuracy for recurring bottle-type characteristics.
Inspection data and statistics provide transparency on the inflow of empties into the process.
The application scope covers common returnable crates independent of compartment type, colour or logo.

Features

Suitable for common returnable crates with different compartment types, colours and logo conditions.
Unpackability check with interspace control for foreign object detection.
Detection of foreign objects above bottles, lying bottles, flipped bottles and broken bottles.
Bottle sorting by material, diameter, height, contour and colour.
Checks for neck labels, closure type and closure material.
Differentiation between one-way PET and reusable PET bottles.
Crate sorting by format, compartment type, conveyor flow direction, colour and logo.
Detection of crate damage and deformations.
Bottle counting for detection of missing bottles.
Detection of residual caps on bottles.
Integrated tracking, easy teach-in, advanced statistics and remote maintenance references.
Multi-line concept with cascadable and expandable system references.
Deep Learning software module for assigning learned crate-content characteristics to bottle types.