Disaster Risk Monitoring - entropy evaluation

Github Repo: Disaster Risk Monitoring - Entropy Evaluation

We utilized the Sentinel Constellation’s Biosphere monitoring information, through NIR and SWIR monitoring of the Earth, to track entropic anomalies. Since the NIR and SWIR information concerns plant nutrient and hydration levels, which form integral part of the natural water cycle, we can track disruptions in the watercycle before they are apparent to the naked eye or even disparate measurements 

Floods and droughts are becoming more frequent and more costly, but the people responsible for responding to them still face two major problems: they often detect risk too late, and they are overwhelmed with raw environmental data that is difficult to act on.

Our project addresses that gap.

We built a risk monitoring as a service platform that turns satellite anomalies into actionable field intelligence  for earlier flood and drought warning.

Demo setup

Our system monitors satellite data, detects unusual environmental changes, and assigns alert status to specific areas by coordinates.  

When an anomaly is detected, local sensors can then be deployed to validate the risk on the ground.

Here’s a quick look at the workflow.

After demo / technology

Technically, our prototype analyzes historical satellite data linked to flood and drought regions and detects anomalies through entropy-based analysis of NIR and SWIR signals.

But the important point is not the math — it’s the outcome.

Instead of giving users more raw data, the system highlights where risk may be developing and helps focus attention on the areas that matter most.

We also built a working dashboard that shows the alert status of a given area by coordinates.

 Business / value proposition

The real value of this project is that it improves disaster monitoring in a practical way.

Today, monitoring systems often force a tradeoff: satellites provide broad coverage, but limited local verification. Ground systems provide detail, but they are expensive and sparse. Our approach combines both in one workflow.

First, satellite monitoring gives scalable early detection across large regions. Then, local off-the-shelf sensors can be deployed only where anomalies appear, which makes response more targeted and cost-efficient.

This is especially valuable in regions with sparse infrastructure, where continuous dense sensing is not realistic.

For customers like governments, insurers, environmental monitoring agencies, and infrastructure operators, that means earlier warning for flood and drought risk, fewer false alarms, and better prioritization of resources.

 Why now

Why now? Because climate change is increasing the frequency and severity of these events, while lower-cost sensors and better remote sensing tools make this kind of system more feasible than ever.

What we built in this hackathon is not just a dashboard and not just a detection model.

It is the foundation of a practical climate-tech solution: a scalable risk monitoring platform that helps decision-makers move from overwhelming data to timely action.

In one sentence:We turn anomaly detection into actionable field intelligence.

 With climate change increasing the urgency of flood and drought resilience, we believe this is a practical and commercially promising climate-tech solution.

Project Summary technical

Software: 

CDSE, OpenEO - APIs

OpenEO Python Client, NumPY, Rasterio, OpenCV, FastAPI, Pydantic, Uvicorn - SDKs

Hardware:

ESP32CAM, Raspberry Pi Zero W, LilYGo TTGO T-Beam 

Data: 

Copernicus, Sentinel 2L2A band B04, B08, NDVI Score, Shannon Entropy analysis, temporal variance

WHO WE ARE? 

Eliana Gabriela Postolache – leads the team by ensuring strong coordination, clear direction, and effective problem-solving, while also contributing to the development and management of CI/CD pipelines.

Anastasia Ciobanu – heads the software development, managing the technical implementation and system infrastructure.

Ștefan Alin Speteanu – is responsible for the hardware side, focusing on embedded systems and overall system integration.

Selin Birinci – supports the hardware development by contributing to system design, data analysis, and technical research.

Mihai Iulia Florentina – leads the creative direction, shaping the project’s presentation through storytelling, design, and pitch development.