Intelligent quality control of (matt) reflective surfaces

Automated inspection of damages and defects in the industrial production with the ALVISTO system


Our system for quality control of reflective surfaces adresses all challenges that are commonly associated with previous quality inspection systems. No more complex setups, hefty investments, or specific location requirements. ALVISTO is designed for flexibility and easy integration into your existing production processes. This means that manual, and therefore error-prone, visual inspections are not needed anymore. ALVISTO stands out from the competition by utilizing diverse image processing techniques and state-of-the-art artificial intelligence (AI) methods.

ALVISTO scans the inspected material and then analyzes the video material with the help of AI. Our fully automated system takes less than a minute to inspect the surface with high precision. Utilizing deep learning, we categorize surface defects into individually customizable defect classes, tailored to the specific needs of your industry. A more detailed description of the procedure can be found here.


Flexible design and easy integration

AI-based visual quality control is characterized by a simple, mobile and cost-efficient hardware setup that does not require any special safety precautions or environmental conditions. This allows us to offer you a simple integration in existing productions or even mobile usability — wherever you need.

The 100 percent test coverage and our real-time analysis make it possible to quickly detect quality defects, such as dust inclusions, in industrial production and take appropriate measures early on in the production process before further costs are incurred. ALVISTO is also ideal for inspecting and analyzing damages caused by hail or other impacts.

 

Individually customizable for your application

In addition to implementing ALVISTO for your individual application, we also offer to retrofit your existing image processing solution with an AI module. We are happy to advise you in optimizing your existing system.

© Fraunhofer IAIS

Want to get started?

Let’s plan a workshop and figure out how you can use ALVISTO best. We usually start with a joint pre-meeting, followed by a short feasibility study in which we test the compatibility of your parts with our technology. If the results are positive, we continue the workshop individually tailored to your needs and interests.

© Fraunhofer IAIS

ALVISTO is our system with which we scan your products, detect damage and production errors and then evaluate and categorize them. It is not limited to one type of product or industry but can be customized for many purposes. It is already being used successfully in many areas and facilitates and optimizes quality control. ALVISTO is characterized by the intelligent interaction of various image processing procedures and AI methods.

Our surface inspection system — your benefits at a glance


Low hardware and maintenance costs; cost reduction through efficient testing


Modular design; no special environmental conditions required and mobile use


100 percent test coverage during ongoing production


Artificial intelligence: autonomous training of various error characteristics regardless of the component size


No expert knowledge required for usage


Customization for different industries


Inline testing possible


Error characteristics can be individually adjusted


Use on highly and diffusely reflective surfaces

The AI-based visual quality control process involves three steps

Step 1: Capturing the object via video

Using deflectometry, the contact-free recording of reflective data, the object to be examined is passed under a scanner and recorded with a video camera.

Precise detection of the reflected light using the scanner works regardless of changing light conditions. It is therefore not necessary to specially seal off the system from external influences. This allows our technology to be used in halls or workshops regardless of the lighting.

Step 2: 2D reconstruction of the surface

The video is analyzed immediately and the surface of the object is reconstructed in two dimensions. We use methods from classical image processing, which are characterized by their short runtime and therefore enable real time processes.

The component geometry of the object does not need to be known for this.

 

Step 3: Detecting the damage with deep learning

The damage is detected and classified based on 2D reconstruction using artificial intelligence methods (deep learning).

Various defect features, such as scratches, paint defects or dents, right down to the smallest dust inclusions, can be trained on a self-learning basis.

The error characteristics can be adapted flexibly and extended to your individual quality criteria.

Application examples

Automotive/Transport

  • Quality control in autobody painting
  • Quality testing of add-on parts from suppliers
  • Detection of hail damage
  • Inspection of rental and leasing vehicles

 

Renewable energy

  • Error detection in solar panel production
  • Quality testing in the manufacture of wind turbine components
  • Inspection and maintenance in the service area for solar systems

Domestic appliances

  • Quality inspection of the surfaces of device housings
  • Quality control in ceramic hob production
  • Quality monitoring in the production of sheet metal blanks

Elektronics

  • Quality control in display production for automotive and smartphones
  • Quality assurance in the “Refurbished Products” area
  • Checking lens filters for integrity

Healthcare

  • Ensuring surface quality in vial production
  • Quality assurance of medical devices

Success Story

© Fraunhofer IAIS

Using ALVISTO to detect hail damage

ALVISTO has been used for the automated inspection of hail damage at various locations throughout Germany since the beginning of 2022. It detects the hail dents and divides them into size classes. A mobile, arc-shaped scanner is used for this purpose, which can be set up within 30 minutes.

There are five cameras with single-board computers in the arch. This means that the image material is processed completely offline; a connection to a cloud or anything similar is not necessary.

The advantages: The scanner can be flexibly transported to different locations and, if necessary, only used there for a short time (e.g. in halls with ambient light). ALVISTO evaluates the damage in a quarter of the time that it would have taken an expert to fulfill the same manual check. The results recorded are robust: the deviation is less than five percent when the same vehicle is scanned multiple times.

 

Image left: The hail damage on the car detected by ALVISTO is displayed on the scan.