The client is a Tier 1 auto components manufacturing company supplying products to leading OEMs

Problem Statement

  • Reducing dependence on manual quality inspection and resulting errors
  • Eliminating manual repetitive tasks which result in inconsistencies over a period of time
  • Avoiding the human subjectivity during inspection

Business Impact

  • Our solution classified components with 99.9% accuracy for the most prominent defects

Solution Given

  • Image classification algorithm based on component images
  • Automated machine learning to learn characteristics of defective and good components
  • Trials to test accuracy of algorithm with set of components
  • Cloud and edge based solution integrated with IP camera for image classification