digilib@itb.ac.id +62 812 2508 8800

Abstrak - Rizky Dassa
Terbatas  Irwan Sofiyan
» Gedung UPT Perpustakaan

Digital Image Correlation (DIC) is an advanced technique utilized in the field of material mechanics to measure deformations by systematically comparing images taken pre-deformation and post-deformation. DIC can meticulously ascertain the spatial variations in displacements through its algorithm. DIC offers a non-contact approach to measure deformations across an entire surface, contrasting with contact methods like LVDTs which provide localized deformation measurements at specific contact points. This non-intrusive method facilitates a comprehensive measurement, thus ensuring detailed and accurate readings in a short time. In the past few years, the Faculty of Mechanical and Aerospace Engineering of Institut Teknologi Bandung has successfully created and developed a 2D DIC software which equipped with SURF (Speeded-up Robust Features), a feature detection and matching algorithm and an iterative algorithm, RANSAC (Random Sample Consensus), which is utilized for identifying outliers from a set of data points. However, real-world deformations are in a three-dimensional realm and measuring deformation in three-dimensions is crucial because they provide insights into hidden deformations that can significantly affect material and structural integrity. Neglecting these displacements could lead to incomplete analysis and potential oversight of critical stress points or failure mechanisms. This project aims to develop a Stereo DIC system utilizing SURF and RANSAC algorithms that has been used in the 2D DIC. Two cameras utilized to capture object deformation images from different angles. These images are reconstructed to pinpoint a 3D cloud point from both cameras which then used to map the displacement between images. The developed stereo DIC have successfully measure an out-of-plane deformation with a small error percentage.