Bottle Label Inspection System

detect missing and damaged labels

bottle label reject systemHelps you deliver safe and well-marked products to customers without having to manually check every label.

How it Works

Our factory-tested and approved vision system uses a high-resolution camera, high-quality lens, and diffusion-axis lighting. Showcasing finely tuned deep learning performance, the system determines presence, orientation, position, and integrity of labels with a customizeable software to save multiple sets of tolerances for your own label needs. Authorized users can create distinct inspection recipes that are saveable so that the same recipe can be loaded for each time a label is run. A reject cylinder automatically pushes bottles from the line for all labels that do not meet specifications.

A Challenge Resolved

types of defect for label inspectionA major healthcare company had to implement costly and inefficient manual checks for scientific and medical serums. There was an unacceptable risk that bottles with missing, smudged, or damaged labels were unknowingly distributed. Artemis Vision’s machine vision system for label inspection automatically and reliably ensures that bottle labels are free of error.

We trained a neural network to recognize errors based on client samples, and then re-trained the neural network to repeatably detect smaller and smaller errors, to as little as 2 degrees of rotation and 1/10th of an inch in defect size.

defect detection label inspection system


CapabilitiesDetermines label presence, orientation, position, and integrity
Resolutioncustomizeable; base model uses 4096 x 3000
ConfigurabilitySave multiple sets of tolerances for your own label needs. Configurable for label and product size. System has been factory-tested for bottles from 30ml-4L.
ApplicationsConsumer goods and commodities labels, including for strictly regulated products such as medical substances, children’s products, food, cosmetics, and pesticides
FootprintIntegrates into your existing packaging line