Design of a Small-Type Wheeled Pipeline Robot Driven by Monocular Vision

Based on the foundation of existing mature pipeline robots, a small-type pipeline robot is developed by using wheel locomotion mechanism in this paper, which is applicable to small pipelines with an inner diameter of about 120mm. The support mechanism of the proposed pipeline robot is acted by springs, similar to that of the wall-pressing pipeline robots. The waterproof probe capturing video in CCTV format is arranged on the proposed robot and then the image is converted into USB format through a converter, which is found after a lot of effort, allowing the image to be transmitted to the computer for real-time processing. In the later stage, deep learning methods capable of detecting imaging characteristics of landmarks of pipe or depth of field such as YOLOv8 are planned to be used to model the interior of the pipeline, to achieve the goal of driving the pipeline robot with vision.