OUR PROJECT

Flooding is the most frequent and devastating natural disaster, causing over $3.5 billion in damages annually in the U.S. alone. With the frequency of flooding projected to triple by 2050 due to climate change, the need for reliable flood monitoring has never been more urgent. Unfortunately, current calibration methods for flood detection systems are inefficient, error-prone, and heavily reliant on costly equipment and specialized expertise, limiting their effectiveness and accessibility.

Our project addresses this challenge by developing a browser-based tool that simplifies the calibration process for smart cameras used in flood monitoring systems. Partnering with our client, Dr. Eck Doerry, we aim to harness the power of computer vision to streamline the generation of calibration data. Users upload an image of the monitoring area, customize detection parameters, and submit it for automatic marker detection and labeling. They can then interact with the results, designate a "zero point," and generate precise calibration data. This data, including marker distances and spatial measurements, is exported in a standardized JSON format for easy integration into broader flood detection systems.

By enabling real-time and accurate calibration without the need for expensive equipment or extensive training, our project lays the groundwork for effective and accessible flood monitoring. With this tool, communities can deploy calibrated flood detection cameras faster and more reliably, providing crucial data to warn vulnerable areas of impending danger. The result: lives saved, infrastructure protected, and millions of dollars in damages potentially averted.