International Research Journal of Engineering and Technology (IRJET) Volume: 11 Issue: 03 | Mar 2024
www.irjet.net
e-ISSN: 2395-0056
p-ISSN: 2395-0072
PROJECTION OF THE EXTENT OF INUNDATION CORRESPONDING TOTHE FORCASTS OF FLOOD LEVELS IN A RIVER KAMALESH R 1, RITHICK A2, SUBRAJA T 3 1
Student, Dept. of Computer Science and Business System, Bannari Amman Institute of Technology, Tamil Nadu, India 2Student, Dept. of Computer Science and Business System, Bannari Amman Institute of Technology, Tamil Nadu, India 3Associate Professor Dept. of Computer Science and Business System, Bannari Amman Institute of Technology, Tamil Nadu, India ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Remote sensing has become an effective tool
inundation maps is useful for risk analysis, damage assessment, evacuation planning, and emergency response. However, the lack of high-resolution topographic data, the unpredictability of hydrological inputs, and the computing expense of executing two- dimensional (2D) hydrodynamic models—which who were the victim of fraudulent practices of spammers who send emails pretending to be from reputed companies with the aim to gain sensitive information like passwords, credit card numbers etc., These has resulted in mimic the movement of water over the terrain—often pose challenges to inundation mapping. For inundation mapping, machine learning (ML) techniques have become a viable addition to or replacement for conventional hydrodynamic models in recent years. ML approaches don't need explicit physical assumptions or equations; instead, they can learn complex nonlinear relationships between input and output variables from data. Additionally, ML approaches can manage noisy, missing, and uncertain data and offer quick, scalable solutions. In this paper, we offer a novel work flow for flood mapping and introduce a spatiotemporal context learning (STCL) method to address these problems. The primary goal is to precisely and automatically draw the boundary of the water's surface and It was trained using the permanent pixels and a range of spectral properties. By utilizing all available spatiotemporal and spectral data, the suggested method enhances the capacity to map flooded areas.
for mapping inundation because of its ability to cover large areas both spatially and temporally. In the investigation of floods using remote sensing, numerous techniques have proven successful. In general, supervised techniques yield more precise results than unsupervised ones. Its results are subjective and challenging to achieve mechanically due to human participation, which is crucial for disaster response. Our work presents a novel strategy that combines the Modest AdaBoost classifier with the spatiotemporal context learning method with the goal of accurately and automatically extracting flooding. The confidence value of each pixel, or the likelihood that a pixel will remain intact, was first determinedby building the context model using the photos. We show that our method can provide accurate and timely inundation maps that are easily applicable to various river basins and flood scenarios, all with little data and processing needs. We also discuss the limitations and challenges of our approach and make some suggestions for additional study.
Key Words: machine learning, flood mapping, inundation
mapping, optical sensors, Modest AdaBoost, HJ-1A/B CCD, andGF-4 PMS
1. INTRODUCTION
2.
Among the most destructive natural catastrophes, floods result in enormous losses in terms of people, property, and infrastructure. Flood forecasting has to be accurate and fast in order to effectively manage and mitigate disasters. Yet, predicting floods is a difficult undertaking that calls for accurate data, powerful computers, and intricate models that can accurately depict the hydrological and hydraulic processes that lead to the formation and spread of floods. Estimating the area and volume of water that covers the land surface during a flood event, or the breadth and depth of inundation, is a crucial part of flood forecasting. Information from
© 2024, IRJET
|
Impact Factor value: 8.226
LITERATURE SURVEY
Chang Li-Chiu and Chang Fi-John, 2019 suggested a different strategy. constructing a platform for the intelligent integration of hydro informatics for regional flood warning systems. This paper initially provides an overview of the major benefits of the machine learning techniques for flood forecasts that were suggested in this special issue. Then, using cutting-edge machine learning, visualization, and system development techniques, it creates an intelligent hydro informatics integration platform (IHIP) to create a user-friendly web interface system that enhances online forecasting and flood risk management. To successfully
|
ISO 9001:2008 Certified Journal
|
Page 811