Use AI to detect defect pin in electrical connector

Yong, Tian Ze (2025) Use AI to detect defect pin in electrical connector. Final Year Project, UTAR.

Preview

PDF Download (2187Kb) | Preview

Abstract

This project aims to develop an intelligent inspection model capable of detecting defects in small electrical connector pins, which are critical components in many electronic systems. The work is structured into two primary components: data preparation and model development. In the data preparation phase, a custom dataset will be generated, featuring images of electrical connectors with three common types of pin defects: missing, shifted, and rotated pins. High-quality image data is essential for accurate model training and reliable detection outcomes. The model development phase leverages the YOLOv8 object detection algorithm, selected for its balance of speed and accuracy in real-time applications. Image processing techniques are employed to enhance dataset quality, and the dataset is annotated manually to ensure precision in model training. Performance evaluation will be conducted using several key metrics—accuracy, recall, precision, and F1 score—to assess the model's capability in identifying defective pins effectively. This project ultimately seeks to offer a practical and automated solution for improving quality control in electrical connector manufacturing processes, reducing the need for manual inspection and minimizing human error.

Actions (login required)