Project Background
In industrial manufacturing, sheet metal parts are widely used in machinery, automotive, electronics, and other industries, and their quality directly affects product performance. Defects in sheet metal parts include subtle depressions/protrusions, scratches, abnormal hole sizes, straightness issues, among others, making detection challenging. With a wide variety of workpieces and differing sizes and shapes, traditional visual inspection relies on manual labor and auxiliary tools, resulting in low efficiency and significant errors, which are difficult to meet the demands of modern production.
Customer Requirements
Precisely detect all types of defects, without overlooking even subtle flaws.
Be able to statistically analyze defect images from different batches to assist in process improvement.
Achieve an accuracy rate of 99.9% for inspection results to ensure product quality.
Maintain an inspection cycle of ≤40s/piece, balancing efficiency and accuracy.
Be capable of stable operation 7*24h to adapt to automated production.
Allow for quick model switching with minimal manual intervention.
Technical Challenges
The complexity of workpieces makes configuring the lighting environment difficult, affecting image acquisition.
High detection accuracy requirements necessitate optimization of multiple aspects.
The inspection cycle is short,requiring high system processing capabilities.
Image acquisition and positioning for large-sized workpieces are challenging, necessitating high design standards for equipment.
