Investigating the accuracy of boat propeller blade components with reverse engineering approach using photogrammetry method

Abstract

This research aims to investigate the accuracy of 3D scanner by conducting a comparative analysis of AI-based photogrammetry method and Agisoft Metashape software, a more affordable and accessible approach compared to expensive commercial solutions. The proposed method uses photogrammetry and CAD (Computer-Aided Design) to measure propeller blade geometry through a three-stage process. In the first stage, the images are taken with a digital camera with calibrated and encoded targets. Additionally, reconstruct the propeller blades in Geomagic Studio. Taking photos of objects only takes a short time and is professional. Subsequently, in the second stage, the 2D images are processed using Agisoft Metashape software and Artificial Intelligence (AI) algorithms to generate an accurate 3-dimensional (3D) model. The third and final stage employs CAD programs to measure various radii of each blade in the 3D virtual model. To validate the accuracy of the measurements, the results are compared with those obtained through machine CMM measurements of each area on the propeller blade. The results show that this method successfully automates the entire reverse engineering process at low cost, lower shooting time, and producing 3D models with high accuracy. The validation shows that the 3D results using Agisoft Metashape have a smaller error (%) value compared to the 3D results using AI. Ultimately, Agisoft Metashape yields higher accuracy results compared to AI, but the results suggest that AI can serve as an alternative for the reverse engineering process in this modern era.