This project aims to track changes in water bodies using satellite imagery over a four-year period. The inspiration for this project came from an article discussing the fact that some of the world's largest lakes are drying up. In order to track these changes, the project utilizes satellite imagery from the Sentinel Hub and Resisc45 datasets. In the modeling phase, the project employs two different neural network architectures, U-net and Residual U-net, to segment the water bodies in the satellite images and accurately track changes over time. The use of these advanced machine learning techniques allows for a more precise and efficient analysis of the water bodies in question. Overall, the goal of this project is to use satellite imagery and deep learning to better understand and monitor changes in water bodies, with the ultimate goal of informing decision-making and management strategies to preserve these important resources.
Water scarcity is becoming a critical issue globally, with climate change exacerbating the problem. Many regions are experiencing a decline in water availability due to increased demand from population growth and industrial expansion, as well as changes in precipitation patterns and rising temperatures. This is leading to lakes and other water bodies drying up, which not only affects local communities and industries dependent on these resources, but also has wider implications for biodiversity, ecosystem services, and water security. The development of the satellite-based information industry has revolutionized the way we can study and monitor our planet. The use of satellite imagery, remote sensing, and other technologies allows for the collection of vast amounts of data on various aspects of the earth's surface, including water bodies. This data can be used to track changes over time, providing a comprehensive picture of the state of water resources and the factors influencing them. Additionally, the use of machine learning and other advanced analytics techniques can help to extract valuable insights from this data, which can inform decision-making and management strategies to preserve water resources.
In order to tackle this project, we proposed several solutions. The first solution is to use the latest technology in satellite-based information to track changes in water bodies over a four-year period. We utilized satellite imagery from the Sentinel Hub dataset as well as a Resisc45 dataset that we labeled manually using an online tool named makesens.ai to gather a comprehensive picture of water bodies and their changes. Additionally, we employed machine learning techniques such as U-net and Residual U-net to segment water bodies and accurately track changes over time. By using this advanced technology, we can ensure that the data we collect is accurate and up-to-date. Another solution we proposed is to make the information and insights generated by the project accessible and useful to stakeholders, such as water managers and policymakers. To achieve this, we used visualization tools and analytical methods that can be easily understood by non-technical users. By providing stakeholders with the information, they need to make informed decisions, we can help to preserve water resources and address the critical issue of water scarcity.