SST Surface OS provides an unmatched performance boost to existing surface inspection systems based on proprietary supervised and unsupervised deep learning models for image recognition. The ready-to-install SST Surface OS unlocks the true potential of existing hardware systems without requiring the installation of any new equipment. The system is running live at a number of hot and cold steel processing lines.
SST Surface OS includes the most accurate surface defect classifiers for slabs, hot rolled coils, pickled coils, cold rolled coils, and galvanized coils. The underlying deep learning architectures have been designed and optimized by Smart Steel Technologies specifically for each steel manufacturing process.
At the same time, SST Surface OS centralizes all surface defect data from hot rolling, pickling, cold rolling, and galvanizing lines. It provides a one-of-a-kind cross-process perspective on the manifestation of defects. Upstream defects are easily distinguished from downstream defects, greatly reducing the effort required to find the root cause behind systemically persistent defects. Similarly, it is now easy to check which upstream defects really propagate downstream, and which do not.
SST Surface OS permanently reduces maintenance cost for existing surface inspection systems. For this purpose, the software package includes tools specifically designed to reduce manual efforts.