SolVisionCase Study

AI Detection of Welding Defects

Welding Quality Control for Air Conditioning and Refrigeration Products

Automating welding bead inspection

When manufacturing air conditioners and refrigerators, components that make up the heat exchanger such as tubes, mirror plates, front and rear headers need to be welded. To verify the welding quality of these parts, an inspection process is required.

Irregular shaped welding defects

Welding factories are often high temperature, high heat facilities, and working conditions are hazardous. Welding defects also tend to be complex and irregular, and consequently manual inspection may not be the best solution for maximizing throughput or consistent production. 

AI defect detection with SolVision

Powered by AI, SolVision can automate welding inspection processes by learning the different shapes and features of weld beads from sample images, then training the AI to quickly detect whether a bead is defective based on its features e.g. is it too thin, thick, or missing. By recognizing defective beads, this effectively controls the quality of the welding process. 

Example

Complete weld bead

Welding Quality Control for Air Conditioning and Refrigeration Products

Missing Weld Bead

Welding Quality Control for Air Conditioning and Refrigeration Products

Incomplete welding

Welding Quality Control for Air Conditioning and Refrigeration Products

More Case
  • 快速精準辨識多種橡膠射出成型之瑕疵

    Inspecting Rubber Injection Molded Parts

    Using AI deep learning, Solomon SolVision can inspect injection molded parts by learning the characteristics of different defects from sample images to locate and identify them precisely.
  • male professional checking the condition of a fire extinguisher

    Fire Extinguisher Inspection Using AR + AI

    Discover how a company improved fire extinguisher inspections by implementing META-aivi, enabling real-time issue detection, streamlining processes, and reducing inspection time by 60%.
  • AI Inspection for Aluminum LEDs

    SolVision can locate and recognize the smallest defects and production anomalies on aluminum PCBAs through image processing.
  • Automotive Kitting and Assembly Line Automation

    Kitting and Assembly Line Automation

    Equipped with AccuPick, Solomon’s AI-based 3D bin picking software, and SolScan, a structured light scanner, our solution excels at accurately recognizing intricate objects.