Scholarlink

<p>Water has been a crucial element for the sustenance of civilization throughout history and civilizations have sprung up around a body of water in one form or another. It becomes imperative to address the pressing issue of water shortage and the shrinking size of urban water bodies, which is particularly relevant in Indian cities like Bangalore. The effective management and preservation of these invaluable resources depend on the development of accurate and automated tools to monitor them. The proposed framework introduces a novel approach, combining a level set-based segmentation algorithm with double exponential smoothing to monitor water bodies using multispectral satellite images. In-depth review of nine lakes within Bangalore was carried out using a Landsat time series data set spanning 1987 to 2020. The resulting forecasting model, employing a univariate smoothing methodology, showcased exceptional performance metrics. Notably, it yielded an average error of 0.072 and exhibited a robust correlation coefficient of 0.94 when cross-referenced with proven results. The proposed framework holds great potential for practical implementation in the domain of long-term water body analysis, effectively catering to the requirements of administrative and decision-making entities. Moreover, the adaptability of this framework for the incorporation of additional external factors, as well as its potential to analyze seasonal dynamics, offers exciting avenues for further exploration. The dataset of delineated lake images prepared in this study presents an opportunity for the advancement of image-to-image regression networks, enabling the prediction of both area and shape variations for lakes, thereby enhancing predictive accuracy and insights.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A generic framework for forecasting lake surface area dynamics using level set segmentation and double exponential smoothing

  • 期刊论文

  • 发表日期：

    2025年7月19日

查看全文
  • T. V. Bijeesh,
  • K. N. Narasimhamurthy,
  • P. K. Sathish,
  • B. J. Bejoy,
  • Michael Moses Thiruthuvanthan

摘要

Water has been a crucial element for the sustenance of civilization throughout history and civilizations have sprung up around a body of water in one form or another. It becomes imperative to address the pressing issue of water shortage and the shrinking size of urban water bodies, which is particularly relevant in Indian cities like Bangalore. The effective management and preservation of these invaluable resources depend on the development of accurate and automated tools to monitor them. The proposed framework introduces a novel approach, combining a level set-based segmentation algorithm with double exponential smoothing to monitor water bodies using multispectral satellite images. In-depth review of nine lakes within Bangalore was carried out using a Landsat time series data set spanning 1987 to 2020. The resulting forecasting model, employing a univariate smoothing methodology, showcased exceptional performance metrics. Notably, it yielded an average error of 0.072 and exhibited a robust correlation coefficient of 0.94 when cross-referenced with proven results. The proposed framework holds great potential for practical implementation in the domain of long-term water body analysis, effectively catering to the requirements of administrative and decision-making entities. Moreover, the adaptability of this framework for the incorporation of additional external factors, as well as its potential to analyze seasonal dynamics, offers exciting avenues for further exploration. The dataset of delineated lake images prepared in this study presents an opportunity for the advancement of image-to-image regression networks, enabling the prediction of both area and shape variations for lakes, thereby enhancing predictive accuracy and insights.