Summary:

A cement manufacturer relied on manual inspection for surface defects in production. The process was time-consuming, inconsistent, and prone to human error, leading to quality issues and customer complaints.

What We Delivered:

• Surface Anomaly Detection using high-resolution imaging and pattern recognition

• Automated Decision Support with AI-recommended remedial procedures

• Process Automation Pipeline for noise reduction, edge detection, and feature enhancement

• Defect Classification System with historical trend analysis

  
  

Business Outcomes:

• 65% reduced inspection time

• 96% defect detection accuracy (compared to 78% with manual inspection)

• Enabled preventive maintenance through early defect detection

• 25% reduced unplanned downtime

  
  

Tech Stack:

• OpenCV

• PyTorch

• Azure Machine Learning

• Azure IoT Edge