Automatic quality control of welding seams

In an assembly step of a Swiss high-tech company, tiny metal pieces are contact welded together. To ensure a high-quality standard, each welding seam was manually analyzed with microscopes.

We developed algorithms that analyzes the high resolution images automatically to detect and evaluate the welding seams. The detection algorithm had to be flexible enough to produce accurate results for various material combinations and potential discoloring. The algorithms proved to be more precise than the manual evaluation.

Key Technologies/Terms

  • Image Processing
  • Algorithms
  • C#/.NET