Application of Convolutional Neural Networks in Defect Detection System for Powder Metallurgy Small Gears

Abstract

This study focuses on the development of a precise defect detection system for powder metallurgy small gears using convolutional neural network (CNN) technology. Through extensive data collection and analysis of images, the CNN model was trained and optimized to accurately identify common defects such as porous characteristics, cracks, and surface irregularities. The executed results validated the system's accuracy and sensitivity in analyzing the porous nature of powder metallurgy sintered parts. This work successfully addresses challenges in gear inspection and establishes a foundation for an automated quality inspection system. These achievements aim to elevate production efficiency and quality standards in the powder metallurgy industry while providing valuable insights for future advancements in inspection technologies.