SOFTWARE DEVELOPMENT OF TWO-DIMENSIONAL DIGITAL IMAGE CORRELATION USING SPEEDED UP ROBUST FEATURES (SURF) ALGORITHM

• Main Author: Wiryanto, Wiweka
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/65358
• Institution: Institut Teknologi Bandung

Deformation measurement of an object is one of the critical things in today's mechanical engineering world. By knowing how much deformation that occurs, we can understand how close a component to failure. Unfortunately, the current deformation measurement methods are still not practical to measure the overall deformation of an objects. These problems can be solved by using the digital image correlation (DIC). DIC works by tracking the displacement of some points in the image before deformation and after the deformation occurs and then measuring the displacement distance of each point. From the displacement of each point, the DIC software will be able to analyze it into deformations and stresses due to loading on the measured objects. Previously, FTMD ITB had succeeded in developing DIC software for small deformation measurements using template matching correlation parameters such as normalized cross correlation. However, this correlation method is less optimal when applied to larger deformations because it tends to be less accurate and becomes a large computational load. Therefore, in this undergraduate thesis, the problem is tried to be solved by using a feature matching algorithm called Speeded-Up Robust Features (SURF) along with Random Sampling results tend to be up to 97% more accurate and up to 89% faster in terms of computational load than the previous algorithm.