SolVisionCase Study
AI Inspection on Reflective Metal Surfaces
Automated inspection and classification of different defects Quality control of hardware product casing
Quality control of hardware product casing
Consumer electronic devices such as computers have hardware casings that need to be inspected before assembly to maintain quality consistency.
Reflective surfaces can be difficult to inspect
Subtle defects such as scratches on metal surfaces are hard to detect under normal lighting conditions, and barely visible to the human eye, making it challenging for such inspection tasks to be carried out manually.
Efficient inspection and classification with SolVision
Using SolVision’s Segmentation tool, the system can be trained to recognize different defects based on their unique features, then build a database that sorts them into customizable categories – in this case, ‘noticeable defects’, ‘minor defects’, and ‘micro defects’. AI deep learning recognizes ‘noticeable defects’ and disregards smaller, acceptable defects to improve inspection accuracy and speed, while also ensuring unsatisfactory products do not enter the assembly process.
Defect classification
Level 1 : Obvious Defects
Level 2 : Minor Defects
Level 3 : Micro Defects
