HUMANLIKE INSPECTION

If a human can see your quality error so can Vision - except Vision doesn't get tired.

Vision technology

Labra AI VISION is a modern, deep learning enabled machine vision system for quality assurance, measurement and data collection. VISION uses neural networks to learn distinctive class features from example objects which it is shown. The software trains a neural network on your behalf by comparing example images which the user sets to different classes.

During training the software automatically finds features from the images which separate the classes. These features are then used by the neural network to classify your production. Features can be differences in the shape of the object, differences in color, differences in patterns, cracks or other surface defects.

The software finds these features automatically - the user can train and run a neural network model from the Graphic User Interface (GUI) without having to write a single line of code or specify classification rules. It is possible to start using the software with only a relatively small set of example data and increase accuracy over time with more examples.

Benefits

FAST DEPLOYMENT

Capture images of examples, train your model and start running Vision - all in under a day.

HIGH ACCURACY

The software is invariant to rotations of the object and recognizes both surface and shape defects.

CONTINUOUS LEARNING

Model accuracy increases over time with more example images.

Use cases

Identify and remove faulty produce with any distinguishing visual factor from your production line. You don't need to know what the features are or how to express them to a computer. Simply let Vision image your production and it will learn from the images it takes. If a human can see the error from an image, so can Vision.

Vision stores all images and information related to each product it encounters on the production line. This includes product IDs and dates to resolve customer complaints, which may occur years from now.

Similarly to identifying faulty produce, Vision can classify which product it is imaging. You can use this information later for automation or data collection purposes.

Vision can track the speed, location and size of an object. This information can be relayed to your automation system for the removal of the item.

Vision can highlight the area of the item where the potential error within the product is. This can be used to assist a human operator or used later to understand potential flaws in production.

Vision can measure the length, width and height of your produce including the sizes of components or holes.

Vision can read and store information from barcodes and QR-codes as well as check that the barcode or QR-code is in the proper place and readable.

Delivery includes

VISION SOFTWARE

The standard Labra AI Vision software running in a Linux environment.

SERVER COMPUTER

The standard Labra AI Vision server computer with a powerful GPU.

CAMERA & LIGHTING

Camera and lighting are chosen for your specific use case and surrounding.

INSTALLATION & TRAINING

One-day hardware and software installation as well as user training.

Deployment

  • USE CASE SURVEY

    Use case survey and live demo at your premises.

  • OFFER

    An offer for all hardware and software as well as a technical drawing of lighting and camera placement.

  • SYSTEM INSTALLATION AND TRAINING

    Installation of Labra AI Vision and training of users.

  • 3-MONTH TRIAL PERIOD

    Three month trial period with an option to continue use or cancel and receive deposit.

  • IN OPERATION

    Labra AI Vision inspecting the quality of your production while receiving regular updates and support on demand.

Order a use case survey >