Intellectual Property India Publishes Patent Application for An Automated Weld Defect Detection And Classification System

MUMBAI, India, Sept. 7 -- Intellectual Property India has published a patent application (202621089916 A) filed by Pdpm Indian Institute Of Information Technology, Design And Manufacturing, Jabalpur on July 23, 2026, for An Automated Weld Defect Detection And Classification System.

Inventors include Kumar, Ishank; Patel, Himanshu; Sangam, Ayush; and Raja, Avinash Ravi.

The application for the patent was published on September 04, 2026, under issue no. 36/2026.

Abstract: An automated weld defect detection and classification system including an image capturing module, a processing unit, a feature extractor, a classification module. Operating at 45 to 60 frames per second (FPS), the processing unit processes the captured frames of an active welding zone using a YOLO-based deep learning architecture optimized via transfer learning with pre-trained ResNet, VGG, or EfficientNet weights and detects, classifies, and localizes multiple weld defect classes including cracks, porosity, lack of fusion, undercuts, slag inclusions, and burn-throughs within a single frame at an accuracy of 85% to 95%.

Disclaimer: Curated by HT Syndication.