Real-Time Pose Estimation of Lifting Components in High-Voltage Transmission Tower Construction Using Machine Vision

In the construction of high-voltage transmission towers, accurately determining the position and pose of lifted components at high altitudes is crucial for ensuring safety and precision. Traditional methods such as radio communication and sensor-based positioning have limitations, including signal interference and difficulties in complex environments. To address these challenges, this study proposes a novel machine vision-based approach for real-time pose estimation during lifting operations. The method involves attaching feature labels to lifting components and using a camera to capture images. A unique feature point extraction technique is employed to obtain the pixel coordinates of the target, and the Perspective-n-Point (PnP) algorithm is used to convert these coordinates into spatial pose data in three-dimensional space. The proposed method demonstrates a low error rate in pose measurement and achieves real-time performance, processing over 8 images per second on high-performance platforms. Experimental results show that the approach effectively improves the accuracy and real-time performance of lifting operations, providing a promising solution for complex construction environments.