AI Glass Bottle Defect Detection for Gradient-Colored Surfaces
The Case
Inspecting Gradient-Colored Glass Bottles for Surface Defects
Gradient-colored glass bottles used in cosmetic and fragrance packaging require consistent surface appearance across the finished container.
These bottles can be produced using internal or external sandblasting processes that create a matte, translucent finish. During production, the sandblasting process can also create uneven spots and other surface imperfections.
Surface defects must therefore be identified during quality inspection before bottles proceed to subsequent production or packaging stages.
The Challenge
Inspecting Irregular Defects on Gradient-Colored Glass
Surface imperfections caused by the sandblasting process can vary in color, size, shape, and location.
This variation makes it difficult to establish fixed rules that clearly distinguish acceptable surface characteristics from defects. Traditional logic-based inspection systems depend on predefined thresholds and inspection criteria, making them less suitable for defects that do not follow consistent visual patterns.
The gradient-colored finish also makes fixed inspection criteria difficult to apply because acceptable appearance varies across the bottle surface.
The Solution
AI-Based Glass Bottle Defect Detection with SolVision
SolVision uses AI-based image processing to automate the inspection of gradient-colored glass bottles.
SolVision’s Segmentation tool uses sample images to train an AI model to recognize production defects. Once trained, the model identifies defective regions based on learned visual patterns rather than relying only on predefined inspection rules.
This approach supports inspection of irregular defects that vary across different areas of the bottle surface.
Fragrance Bottle Surface Inspection
 |
 |
| Defect Detection |
During glass bottle inspection, SolVision detects uneven spots and other surface imperfections created during the sandblasting process.
The system segments defective regions and evaluates defect characteristics including location, angle, and size distribution during production inspection.