Research project

Rivers have been the cradle of civilization. But often meteorological phenomena and unchecked human activities cause floods that are difficult to predict and can cause immense damages. Governments around the world focus on developing and deploying sophisticated models to predict floods. But these models require expensive scientific infrastructure and gauges to collect data from rivers. Hence only few percent of rivers across the world are equipped with such infrastructure. But there is an alternative indicator which is excellent early warning indicator for flood risk, which is riverbank width. Various studies made by ‘Federal Office for Water and Geology, Switzerland’ have resulted in guidelines for ideal bank-width based for given river width. But still the key question of how to reliably yet cost effectively monitor riverbank-width and identify zones for governments to take timely preventive actions and reduce impact of floods.

This project aims to address this problem by using an AI model that leverages satellite-based

images of the rivers to accurately measure its width, analyze its existing bank-width, compare it with theoretical ideal bank-width and identify the risk zones where existing bank-width are smaller than ideal. This model being cost effective and scalable, can help governments to monitor flood risk zones to undertake timely interventions like clearing the bank-zone, increasing river’s water carrying capacity via dredging or creating flood-detention before narrowing channels