Pipeline weld detection robot based on YOLOv5-Lite

Abstract

This paper proposes a real-time detection robot for external pipeline welds based on YOLOv5-Lite. The robot consists of multiple parts, including the car body, motion system, image acquisition and transmission system, control system and weld inspection system. Through a new mechanical structure, the robot can make forward and rotating actions on the outer wall of the pipeline, and transmit the acquired image to the weld defect detection model through ESP32-CAM to inspect the quality of the pipeline weld in real time. This paper is designed to realize the intelligence, efficiency and high accuracy of modern detection.