NEWMOS_slide.png

Discover the NEWMOS framework in our technical report.

NEWMOS Nurjana Earth and Wildfire Monitoring Observation and Simulation

The wildfires are one of the most dangerous threats of our times, which constantly harm our environment.

Nurjana Technologies designed a new solution called NEWMOS.

 

This project allows to timely detect the wildfires hotspots through satellite images.

Main Features or the systems are:

  • Strategic planning of the activities for the risk prevention of wild and forest fires

  • Tactical phase characterized via monitoring of ongoing wild fires as well as the fighting activities management

  • Estimation of damages and post event analysis

NEWMOS permits an automatic fire surveillance of an entire region or nation in near real-time. The NEWMOS algorithm uses geostationary satellites (MSG) which guaranteed a good observation frequency of 5 minutes.

The algorithm is based on a polynomial multivariable model. The real innovation introduced by Nurjana is an improved machine-learning model that estimates radiance in the region under surveillance.

Comparison between real Brightness Temperature and the models.

From the left: Satellite image, Machine Learning Model representation, Polynomial Model representation

Polynomial Model represents the state of the art of radiance forecasting methods, but with the machine learning there is an improve of performances of more than 100%. This is due to the fact that machine learning model is more accurate and similar to the real data respect to the polynomial one.

The Machine Learning approach is based on a feed-forward and fully-connected Neural Network. It is been trained using SEVIRI Instrument images, using a week of data. This is a new approach in this kind of algorithm to detect wildfires.

Follow us on

  • Nurjana Technologies su Linked In
  • Nurjana Technologies Facebook
  • Nurjana Technologies su Instagram
  • Nurjana Technologies Youtube Channel

©2020 - Nurjana Technologies S.r.l

Via M.Betti 27/29, 09067, Elmas (CA), Italy VAT: 03450730928 

 

Cookies  |  Privacy Policy  |  Credits

regione_por.jpg