Abstract
Abstract. Floods are the most frequent type of natural disaster that cause loss of life and damages to personal property and eventually affect the economic state of the country. Researchers around the world have been made significant efforts in dealing with the flood issue. Computer vision is one of the common approaches being employed which include the use of image segmentation techniques for image understanding and image analysis. The technique has been used in various fields including in flood disaster applications. This paper explores the use of a hybrid segmentation technique in detecting water regions from surveillance images and introduces a flood index calculation to study water level fluctuations. The flood index was evaluated by comparing the result with water level measured by sensor on-site. The experimental results demonstrated that the flood index reflects the trend of water levels of the river. Thus, the proposed technique can be used in detecting water regions and monitoring the water level fluctuation of the river.
Highlights
Flood frequency and flood impacts have worsened with the extreme changes in climate events, affecting more than 1.5 billion people worldwide from 2000 to 2019 (UNDRR, 2020)
A common approach is by using computer vision, where the images are captured and processed using image processing techniques
Image segmentation plays an important role in image processing and computer vision applications
Summary
Flood frequency and flood impacts have worsened with the extreme changes in climate events, affecting more than 1.5 billion people worldwide from 2000 to 2019 (UNDRR, 2020). Image segmentation is the process of partitioning an image into multiple regions, often based on similar characteristics of the pixels in the image to obtain useful information from surveillance images. It has been applied in the fields of medical imaging, automated driving as well as in water management practices such as in flood monitoring applications (Feng et al, 2019; Jena et al, 2018; Kim & Kim, 2003; Ko & Kwak, 2012; Popescu et al, 2015; Yuan et al, 2017). Previous researchers have developed various image segmentation techniques including thresholding, boundary-based, region-based, hybrid technique, and many more
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
