Improving quality assurance and production throughput with our brand new visual inspection tool
May 23, 2021
Stan Callewaert

Improving quality assurance and production throughput with our brand new visual inspection tool

Computer Vision
Ai in Manufacturing

As manufacturers strive to improve quality assurance1 and production throughput2, they are also facing increased pressure to meet sustainability targets4 and manage rising energy costs3. To address these challenges, ML6 has developed an AI-powered visual inspection solution. This solution utilizes vision data (but could use other data as well) to enhance quality assurance and increase production throughput. It is composed of various components, with standardized code available for each component, allowing ML6 to quickly and effectively customize and integrate the solution to meet specific product and quality requirements. This approach enables ML6 to deliver projects faster and at a higher level of quality.

Visual Inspection Tool

This blogpost gives extra context to the above video. You will find extra information and visualisation about the concepts that are introduced in the video.


It consists of several components. For each of the components there is standardised code available, in that way ML6 can focus on customising and integrating the solution to the specific product and quality parameters. It thus enables ML6 to deliver these kinds of projects faster and at a higher quality.

Quality assurance and production throughput solution components

Goal of the solution

The goal of the solution is one of the following:

  1. Quality Inspection is real-time reporting and alerting about the quality of products on the production line to operators, process engineers, line managers or plant managers
  2. Defect Removal is about removing products from the production line that cause downtime or don’t meet the quality standards. 
  3. Process Steering/Intelligent Machines is about steering the machine to produce faster and at a higher quality.

No in-line quality control, quality control, defect removal and process steering difference

Our process begins by gathering images directly from the production line, which are necessary to train the AI algorithm in the solution. This is done using a camera setup integrated into the production line. However, deciding on this hardware setup can take time. To solve this, we have developed a flexible camera setup that can quickly collect images from your production line without any additional integration work. These images can then be used to make decisions about the hardware strategy and to give an idea of what an AI solution for your factory might look like. We can also come to your factory for a few hours and use this flexible setup to gather data for free, to demonstrate how the solution would work on your products.

We can come for a couple of hours to your factory to gather data with it this flexible setup for free to show how a solution will on your products.

Flexible camera setup from ML6


As an example of what is possible we are going to focus on a fictive example for a pancake manufacturing company. In this example, the actions that could be attached to the detections in the images could be:

  1. Reporting of the quality of the produced pancakes per day, per SKU, per line, …
  2. Alerting when too many defects are detected so that actions can be taken by an operator or process engineer.
  3. Rejecting partly baked pancakes to avoid downtime because they could stick to parts of the production line and cause blockages.
  4. Steering the pancake baking machine so that pancakes are baked as short as possible while still staying within quality boundaries. This both increases production throughput and quality.

Next to a regular camera that was used in the pancake example, ML6 also has experience with near-infrared, microscopic, hyperspectral, thermal, 3D, pressure and other kind of sensor setups. These cameras are a powerful tool to detect different kinds of defects, including those not visible to the human eye, such as the sugar level in a pancake.

If you're interested in learning more about how your company can begin using digital innovation for quality assurance and increasing production throughput, please reach out to Stan at the email

Related posts

Want to learn more?

Let’s have a chat.
Contact us
No items found.