perfume bottle and packaging

SolVisionCase Study

Glass Bottle Inspection Using AI

Case

Translucent Glass Bottle Quality Control

Translucent glass bottles, often created using gradient-colored techniques, feature a distinctive matte finish achieved through high-pressure air jets that apply mist particles to the glass surface. Known for their aesthetic appeal, these bottles are produced using two primary methods: internal sandblasting and external sandblasting. The gradient effect not only enhances the visual design but also adds a unique translucent quality to the glass.

Challenge

Defining Defects on Gradient-Colored Glass

Gradient-colored glass often exhibits uneven spots and imperfections resulting from the sandblasting process. These defects vary in color, size, and location, making it challenging to establish a clear definition or standardized criteria for what constitutes a defect. Consequently, traditional logic-based inspection methods are ineffective for accurately identifying and assessing these imperfections.

Solution

Precise Inspection with SolVision

SolVision leverages advanced image processing to automate the inspection of gradient-colored glass bottles, effectively identifying production flaws. Utilizing its Segmentation tool, the system learns to recognize defects through sample images. Once trained, the AI model can swiftly detect and categorize defects based on location, angle, and size distribution during production, ensuring precise and efficient quality control.

Glass Bottle Inspection

Defects on glass fragrance bottle are detected by AI
Defects on glass fragrance bottle are detected by AI

Fragrance Bottle Defect Detection

Outcome

Automated defect detection
Advanced image processing
Enhanced quality control
Related Posts
  • 快速精準辨識多種橡膠射出成型之瑕疵

    Injection Molded Rubber Inspection Using AI

    SolVision enhances injection molded rubber inspection using AI for precise defect detection and improved quality control.
  • Quality Control of Wafer Dicing

    Using SolVision’s Instance Segmentation tool, irregular lines and multi-drilling defects are labeled in sample images to train the AI model.
  • pile of shiny metal bolts

    Metal Bolt Defect Detection Using AI

    SolVision ensures a flawless 100% defect detection rate in just 49 milliseconds per bolt, addressing reflective surface challenges and eliminating false negatives.
  • Inspecting Safety Certification Marks

    Inspecting Safety Certification Marks. Solomon SolVision’s AI automated detection of defective certification marks .