With vision for pioneering weld seam inspection

AI Hardware Innovation

Reliable weld seams are important for piping systems in the semiconductor industry. To test this quality, SCS has developed an innovative vision inspection system for its customer Georg Fischer.

  • Problem definition

    In the semiconductor industry, reliable weld seams on pipes are extremely important. If a spot leaks, the factory has to be stopped, which has major financial consequences.

  • SCS solution

    SCS has developed suitable imaging processes and algorithms to analyse and verify the quality of weld seams.

  • Added value

    Based on SCS's work, the customer was able to further develop the project into a product. This won second place at the IVS Innovation Award 2019 and was launched in February 2020.

Project insights

The customer Georg Fischer offers a wide range of products that enable pipes to be welded together (Fig. 1). The pipes are welded by heating and pressing them together. Molten material emerges at the pressing point (cross-section Fig. 2). Possible solutions for the quality control of such weld seams were developed together with the customer. The solution ideas developed in the innovation workshop formed the basis for a feasibility study carried out by SCS.

Figure 1: Pipe welder (Source: Georg Fischer)

Figure 2: Weld seam (Source: Georg Fischer)

The aim of the study was to develop a method that could be used to determine different quality characteristics of a weld seam. For example, the weld seam thickness, pipe misalignment, etc. were to be detected.

Various methods and algorithms were developed and tested in the feasibility study. Ideas were developed, difficulties discussed and the best possible solutions worked out in the SCS Vision Lab. The test set-ups enabled a quick evaluation of the various approaches and their feasibility. For example, the test setup with a laser (Fig. 3) initially looked promising, but was quickly discarded thanks to a short iteration cycle. It turned out that the different optical properties of the pipe materials were not suitable for quality control with a laser.

Figure 4: The experimental setup with a laser.

For this reason, the current solution uses transmitted light to provide similar images for different materials. This setup made it possible to use inexpensive camera modules and lighting (Fig. 4). This solution is therefore simpler and delivers very good results.

The captured images were then analysed and evaluated using a classic vision algorithm. The vision algorithm analyses the various dimensions and determines the weld seam quality (Fig. 5).

Figure 4: Setup with backlight illumination.
Figure 5: Evaluation algorithm.

Related projects

Visual inspection of railway wagons

Is the block brake worn? How thick is the pantograph contact strip? Is an incorrectly fitted screw coupling hanging down? In future, SBB's "Visual ... More

Visual innovation through AI in ophthalmology

Optical coherence tomography (OCT) is a gentle laser microscope and is used by ophthalmologists to examine the retina for abnormalities. Trained AI ... More

Reliable AI for the measuring wheel set

SBB uses so-called measuring wheel sets to measure the dynamic contact forces between rail and wheel. Sensors on the wheels measure the forces. It is ... More

Smart Meter Toolkit

According to the Electricity Supply Ordinance, the grid operator must enable the end customer to receive real-time measurement data from the smart ... More

SRF media archive

The media archive for Swiss Radio and Television SRF enables journalists to access archive material dating back to the 1950s and send the image and ... More
Show all projects
Pascal Iselin Algorithms, Analytics & AI How can I help you?