Video Processing Application for Stress-Strain Diagram

This MATLAB application in intended to generate the stress-strain diagram by processing the video of the experiment captured by smart phone

The tension test of rebar is an indispensable part of Reinforced Cement Concrete (RCC) structural analysis. Traditionally, in this experiment, the instantaneous deformation with time is measured manually with an extensometer and elongation scale while a Universal Testing Machine (UTM) applies the load to the test specimen. However, this manual approach requires a considerable amount of human intervention while taking the reading as well as being prone to inaccuracies. To enhance the testing method, digital UTM with built-in software to generate stress-strain diagrams is used worldwide which in turn requires a lot more financial and energy-hungry effort compared to the traditional method. The key objective of this MATLAB application is to generate the stress-strain diagram by processing the video of the experiment captured by smart phone with negligible additional cost and required energy efficiency. Thus, integration of this computer-vision-based MATLAB application in Rebar Tension Testing can be considered as a sustainable choice for the fact that it can be used as long as the apparatus is functioning with efficient energy consumption complying with the 12th Standard Development Goal – Responsible Consumption and Production. Accordingly, this innovative algorithm may drive Structural Engineering experiments one step further towards sustainability. This application utilizes digital image processing-based algorithm to detect color-marked endpoints of gage-length of rebar while tensile load is being applied by processing the video of the experiment and thus instantaneous strain vs time data is obtained without using the extensometer which is very costly. It is combined with the Load vs time data obtained from video leveraging Optical Image Recognition (OCR) in MATLAB. Key techniques include RGB thresholding, subpixel refinement, and KLT point tracking for robust and accurate analysis.