Experimental Study on the Interaction Between Image and Static Vehicle on the Flood Level Measurement

Abstract

Flooding has become a widespread, increasingly severe problem in Malaysia in the past few years and causes problems for calculating the water level because flash floods occur quickly, and the position is uncertain. The aim of this study is to optimize the similarity equation between an image and a static vehicle. To solve this problems, this study propose and conducts experiment by using 3D printed vehicle model and mobile image. Before conducting the experiment, design and manufactured the model using a 3D printer. Prepare a clear aquarium tank filled with water, then place the vehicle model into the tank. Incrementally adjust the water levels to simulate different flood stages. Use a mobile camera to capture images of the water surface and the vehicle model at each water level, ensuring consistent camera positioning for accurate comparison. Then, measure and record the water levels using a ruler, as well as the corresponding images or video frames for each level. After conducting the experiment, this study has collected all the data for optimization. Model 1 has an average accuracy of 98.28%, model 2 has 98.16%, model 3 has 98.19% and model 4 has 98.52% after optimized the similarity equation and the average accuracy increases. This finding could greatly contribute to future flood observing systems.