Performance Evaluation of Low Frame Rate RGB-D SLAM in Piping EnvironmentsPerformance Evaluation of Low Frame Rate RGB-D SLAM in Piping Environments

Abstract

Many factories and power stations lack piping working drawings, thus leading to the demand for 3D models. To address this issue, we are motivated to develop a piping measurement device deployed SLAM algorithm with an RGB-D camera and a motor. One method for reducing the computation cost of SLAM is low frame rate (LFR) processing. To evaluate the impact on SLAM performance with LFR input, we conducted a set of experiments with indoor high-altitude mounted pipes. The results indicate that the feature-based RGB-D SLAM is not affected significantly by LFR input in experimental piping scenes.