AI-powered corrosion detection advances industrial safety
IISc researchers developed an unsupervised machine learning system to detect under-deposit corrosion accurately through optical microscopy. By analyzing porosity and deposit thickness, it achieves 73% accuracy without labeled data. This method offers scalable, objective corrosion assessment, improving predictive maintenance and reducing costly failures across industries globally.