AI-POWERED HARMFUL ALGAL BLOOMS FORECASTING

WHAT IS HABTRAIL?

HABTRAIL is a deep learning model developed specifically for tackling harmful algal blooms (HABs) on a larger scale in ocean and coastal areas at a lower cost as compared to conventional in-situ methods. It comprises two data services. The first one is capable of detecting and monitoring the expansion of HABs with water quality pigments derived from Sentinel-2/-3 as an alert system. The second one offers another data service through a mobile app, which is capable of identifying HABs in near to real-time by uploading a “suspicious picture” of the water. Since HABs severely impact aquaculture, fisheries and tourism,they constitute a significant threat to the public health. HABTRAILs mobile app therefore provides the public with a free tool to quickly detect HABs with high accuracy before engaging in any marine activity.

ADVANTAGES

Systematic monitoring of large or small areas

Machine learning approach

High predictive accuracy

Well optimized hyperparameter scheme

3D web-based GIS

Easy integration API

Configurable alerts by user and area of interest

Subscription-based service - pay per use

CASE STUDY

HABTRAIL as a health risk predictive model and an early warning system - Preliminary Results

Each year, blooms and Harmful algal blooms seize the headlines across the globe putting a hold on recreation, causing the migration of coastal animals, resulting in a detrimental effect on economies and having a deadly threat to human life. HABTRAIL is a health risk predictive model and an early warning system built on a deep learning algorithm to monitor and predict HAB with high accuracy and confidence intervals. HABTRAIL uses water quality assessment pigments including Turbidity, Temperature, CDOM, Salinity and reflectance in a multispectral scheme to create an early warming predictive model by statistically examining the quantitative relation between Chl_a and HABs cell counts to define a quantitative threshold based on a dependency association between the two variables.

Deep Learning

FCNN-CNN

Architecture

HABTRAIL TEAM

ISSAH SULEIMAN

Data Scientist

ANA MARTINS

Scientific Advisor

ANDRÉ TOSTE

Software Engineer

MIGUEL CORREIA

Business Development

AWARDS

Copernicus Masters - Portugal Space Atlantic Challenge

“HABTRAIL has been selected the winner of the Portugal Space Atlantic Challenge because it is an idea that aims to tackle a very complex and pressing problem of our society using advanced Machine Learning technologies applied on both Satellite Imagery and in-situ data. The team behind the idea has expertise in all relevant domains ranging from the technical and scientific side to business. Furthermore, the idea of trying to detect HABs is very relevant to many Atlantic regions and has great potential to scale up as more sensors become available in the future to be able to monitor different species in different areas across the globe.”

Joan Alabart, Industrial Relations & Projects Officer, Portugal Space

Copernicus Masters Hall of Fame - HABTRAIL

CONTACT US

EYECON GROUP

Parque da Ciência e Tecnologia da Ilha Terceira
9700-702 Terra-Chã
PORTUGAL

info@eyecon-group.com

*Privacy Policy and Protection of Personal Data