Visual Inspection of Glass Bottles

Detecting defects on layered glass bottles with AI

Translucent glass bottles

Gradient colored glass is made by using high-pressure air jets to create mist particles on the glass surface. Also referred to as matte glass, there are two popular production methods: internal sandblasting and external sandblasting. The gradation style allows for many design styles, while the translucent effect adds to the aesthetics. 

Defining defects on gradient colored glass

Sandblasting creates uneven spots and imperfections that vary in color, size and location. As there is no clear definition of defects or their features, logic-based rules cannot be established to implement conventional inspection methods

Smarter inspections with AI

Through advanced image processing, SolVision can fully automate inspection processes to detect production flaws. The system’s Segmentation tool allows the features of defective glass bottles to be learned through sample images. During production, the trained AI model can quickly detect and label the location, angle, and size distribution of different defects on a scanned image.

AI Inspection

Scanned Image

Detection Result

Case of Defect Detection of Gradient Glass Bottle
Case of Defect Detection of Gradient Glass Bottle
Related Posts
  • OCR Date Reading on Aluminum Bottle Caps

    SolVision’s OCR boosts food, beverage, and pharma product quality control, overcoming bottle cap printing challenges on high-speed production lines with AI.
  • Pickup truck production line, workers are working

    Electric Vehicle Assembly Verification Using AR + AI

    Discover how META-aivi’s AR + AI capabilities enhanced electric vehicle assembly verification and quality control for one of the world’s leading EV manufacturers.
  • a group of square objects

    Detecting Chipping Defects in Wafer Dicing

    The location, size and shape of cracks vary each time, and traditional optical inspection cannot accurately identify such unpredictable defects.
  • 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.